HIV-RT and E. coli RNase H digestions were the same as in DeStefano et al. (1993) except using 40 m~ KCI. RESULTS. The RNase HZ-specific Cleauage Is ...
THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 269, No. 41, Issue of October 14, pp. 25922-25927, 1994 Printed in U.S.A.
Structure-specific Cleavage of theRNA Primer from Okazaki Fragments by Calf Thymus RNaseHI* (Received for publication, June 21, 1994 and in revised form, August 7, 1994)
Lin HuangS, Yong Kim&John J. TurchiO, and RobertA. Bambaralnll From the $Department of Biochemistry and Wancer Center, University of Rochester, School of Medicine and Dentistry, Rochester, New York 14642 and §Department of Biochemistry, Wright State University, Dayton, Ohio 45435
Cleavage specificity of RNase HI was examined on tors. In contrast, forms of RNase HI1 have lower molecular model Okazaki fragments, to determine the likely roleweights, of higher isoelectric points, and cannotutilize Mn2+as a this nuclease in lagging strand DNA replication. Each cofactor (Vonwirth et al., 1989). substratewaspreparedbyannealing ashort RNA Eder and Walder (1991) isolated an 89-kilodalton form of primer,madebytranscription in vitro, toasingleRNaseHI from K562 humanerythroleukemia cells. They stranded syntheticDNA template, and subsequentlyex- showed that the enzyme would hydrolyze the 5’ phosphodiester tending the primer by DNA polymerization. The calf of a single ribonucleotide in a DNA-RNA-DNMDNA heteroduthymusRNase HI makes a structure-specific endonu- plex. Interestingly, they suggested that the cellular function cleolytic cleavage in the RNA primer, releasing it intact, was to participate in removal of RNA residues incorrectly inand leaving a mono-ribonucleotide at 5‘ the terminus of corporated into genomic DNA. They did not favor a role in the RNA-DNA junction.Thisspecificcleavage,one initiator RNA removal because the enzyme could not cleave the nucleotide upstream of the RNA-DNA junction, is RNA junction ribonucleotide. primersequence-andlength-independent.Cleavage RNA primer is not extended with An early study showing that activityof RNase HI increases specificity is lost if the DNA synthesis inbovine lymphocytessuggests DNA, or if the substrate a has nick at the RNA-DNAjunc-in parallel with tion. In addition, the cleavage at a single site requires the involvement of the enzyme in DNA replication (Busen et al., M e . Cleavage in the presence of Mn2+ is less specific. 1977). Intriguing observationson the possible role of RNase H in DNA lagging strand replication have been reported by DiNeither human immunodeficiency virus reverse transcriptase norEscherichia coli RNases H perform sucha Francesco and Lehman(1985). They showed that theRNase H structure-specific cleavage before an RNA-DNA junc- isolated from Drosophila melanogaster embryos is capable of tion. Our work indicates that calfRNase HI is designed removing initiator RNA primers from2 to 9 nucleotides in The RNaseH also interacted withDrosophila polymerto recognize Okazaki fragments. It has the specificity length. to remove their initiatorRNA segments, except for one ri- ase-primase, stimulatingits ability to carry out DNA synthesis. bonucleotide, by a single endonucleolytic cleavage in Another observation (Funnel et al.,19861, made with the E. coli vivo. o n C replicationsystem, suggestedthat the removal of initiator RNA needed the 5’ to 3’ exonuclease activity of DNA polymerase I, with the assistance of RNase H. Moreover, analysis of A critical step in lagging strand synthesis during DNA rep- mutants of RNase H indicated a role in initiatorRNA removal, lication is removal of the RNA primers usedfor initiating each but also suggested that the action of RNase H could be substiOkazaki fragment. Ribonuclease H, which specifically cleaves tuted by other enzymes in its absence (Ogawa and Okazaki, the RNA strand from an RNMDNA hybrid, has been proposed 1984). to degrade RNA primers from nascent DNA (Keller, 19721, From results of reactions designed to reconstitute steps in but there is very little direct experimental evidence for this mouse cell DNA replication in uitro, Goulian et al. (19901 suggested that RNase HI participates in removal of RNA primers reaction. involvement of Since Stein and Hausen (1969) first identified the enzyme, prior to joining of Okazaki fragments. The SV401 DNArepseveral different forms of RNase H have been purified and RNase HIin Okazaki fragment maturation in characterized from a number of organisms, including bacteri- lication in uitro has also been suggested (Ishimi et al., 1988; ophage T4 (Hollingsworth and Nossal, 19911, yeast (Kanvanet Waga et al., 1994; Waga and Stillman 1994). However, in both al., 19831, Drosophila (DiFrancesco and Lehman, 19851, Krebs systems, the exact function of RNase HI remains to be elucicells (Cathala et al., 1979), chick embryo (Kitahara et al., 19821, dated. We have recently described the maturation of Okazaki rat liver (Roewekamp and Sekeris, 19741, calf thymus (Busen, fragments in vitro, using purified enzymes from calf thymus 1980; Hagemeier and Grosse, 1989;Rong and Carl, 19901, and (Turchi et al., 1994). The reactions involved were removal of generhuman erythroleukemia cells (Eder and Walder, 1991). Eu- initiator RNA, synthesis from an upstream fragment to karyotic RNase H can be classified into two groups depending ate a nick, and thenligation. On the substrate used, the RNase on size, charge and cation requirements. Forms of RNase HI HI cleaved off the initiatorRNA, leaving one ribonucleotide 5‘ isoelectric of the RNA-DNA junction. The calf 5’ to 3’ exonuclease was have molecular mass between68 and 90 kilodaltons, points of about 6.4, and can use bothMnZ+and Mg2‘ as cofac- then able t o remove the remaining ribonucleotide, which was resistant to degradationby RNase HI. This role for the 5’ to 3‘ * This researchwas supported by National Institutes of Health Grant exonuclease is consistent with ourprevious results suggesting GM24441, and byCancer Center Core Grant CA11198.Thecosts of that it is involved in lagging strand processing U“urchi and publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “uduertisement” The abbreviations used are: SV40, simian virus 40; HIV, human in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 11 To whom correspondence should be addressed. “el.: 716-275-3269; immunodeficiency virus 1; RT, reverse transcriptase;W ,avian myeloblastosis virus. Fax: 716-271-2683.
25922
Substrate Specificity
Calf RNase HI
of
A
DNA Bambara, 1993), and may serve the samefunction as the 5' to 3' exonuclease activityof E. coli DNA polymerase I(Murante et (HI al., 1994). In the current report,we perform a thorough assessment of the cleavage specificity of calf RNase HI for substrates with the structure of Okazaki fragments.We discuss how results of this work further define the proposed role of RNase HI in the completion of the lagging strand replication in uiuo.
25923 RNA
38 DNA 75
5'
EXPERIMENTALPROCEDURES Materials-The75-nucleotidelong synthetic oligonucleotidedescribed in Fig. l was kindly supplied by Bristol-Myers Squibb (Seattle, WA). Deoxyribonucleotides were purchased from Pharmacia Biotech Inc. Ribonucleotides, E. coli RNase H and RNase inhibitor were purchased from Boehringer Mannheim. [a-32PlUTPand [a-32PlCTP(800 Ci/mmol, each) were from DuPont NEN. T3 RNA polymerase and Sequenase (version 2.0) were fromU.S. Biochemical Corp. pBS+plasmid was from Stratagene Cloning Systems (La Jolla, CA). HIV-RT was a generous gift from the Genetics Institute (Cambridge, MA). All of the other reagents were from Sigma. RNase HI Purification-Calf thymus RNase HI was purified by the procedure reported by Eder and Walder (1991).Details of the purification have been describedin Turchi etal. (1994).The final pool of RNase HI had a specific activity of 30,000 units/mg as measured on a poly(L3H1rA)-(dT)substrate according to Eder and Walder (1991). Preparation of RNA Primers-The three different length RNA primers were made by run off transcription from a T3 promoter using linearized pBS+ plasmid as template. The pBS+ plasmid was digested by HindIII, PstI or Sal1 to make templates for transcription of the 13-, 21-, and 31-nucleotide long RNAs, respectively. The RNA primers were internally labeled by inclusion of 100 pCieach of [a-32P]UTPand [CY-~~PICTP in the reaction. The final reaction components were:40 m~ Tris-HCI (pH 8.0), 6 m~ MgCl,, 10 n m dithiothreitol, 2 m~ spermidine, 2 unitdpl RNase inhibitor, 1uniupl T3 RNA polymerase, 1mM ATP and GTP, 20 p~ CTP and UTP, and 30 ng/pl DNA template. The reaction mixtures (50 pl) were incubated a t 37 "C for 2 h. The RNAprimers were purified by 20%polyacrylamide, 7 M urea gel electrophoresis (Sambrook et al., 1989). Primer Extension- The labeled and purified RNA primers were annealed to the 75-mer template and extended using Sequenase (version 2.0) and dNTPs. The reactions contained 50 m~ Tris-HC1 (pH 8.0), 10 m~ Mg acetate, 50 mM NaCI, 1m~ dithiothreitol, 1m~ of each dNTP, 0.4 uniffpl Sequenase (version 2.0), 20 pmol of ssDNA template. The reaction mixtures (30 pl) were incubated at room temperature (22 "C) for1.5 h. The annealed and extended products were further purified by 15% native polyacrylamide gel electrophoresis. RNase HIAssay-The calf RNase HI digestions were performedin 20 mM HEPES (pH 7.0), 50 rn NaCl, 1 mM dithiothreitol, 2 m~ MgCI,, 0.001%Nonidet P-40 in avolume of 10 pl. Reactions were incubated at 37 "C for 15 min, and stopped by the addition of 98% formamide and 10 m~ EDTA. baction products were heated at 90 "C for 5 min, and then separated by 15% polyacrylamide, 7 M urea gel electrophoresis. Products were visualized by autoradiography using double Dupont Cronex Lightning Plus intensifylng screens a t -70 "C. The buffer conditions for HIV-RT and E. coli RNase H digestions were the same as in DeStefano et al. (1993) except using 40 m~ KCI. RESULTS
The RNase HZ-specific Cleauage Is RNA Sequence- and Length-independent-RNase HI cleavage specificity was measured on RNAinitiated DNA segments annealed toa DNA template. Three substrates were testedwhich in the RNA segments were different in length and had different RNA and DNA sequences around the RNA-DNA junction. Fig. 1 describes the substrates used in this study in detail. Those designated H , P , and S have RNA segments of 13, 21, and 31 nucleotides long, respectively. Their sequences, and that of adjacent DNA are shown in Fig. 1B. These RNA primers were internally labeled by synthesis with [a-32P]UTPand [~u-~~PICTP. Consequently, most RNA-containing cleavageproducts shouldhavebeen labeled and thereby detectable. Fig. 2A shows the digestion of substrate (H), having a 13-nucleotide-long RNA primer, with increasing
B CTCTAGAGGATCCCCGGGTA-3' FIG.1. Structure of the substrates used in this study. A, the lengths of segments in this andsubsequent figures are given in nucleotides. The saw tooth and straight lines represent RNA and DNA segments, respectively. The nucleotide sequence of substrate R81 can be determined from information in Turchi et al. (1994).B , the sequences of the RNA initiated DNA segments in the substratesare shown. The bold letters ( H , P and S ) designate the 3' ends of the 13-,21-, and 31nucleotide long RNA primers, respectively.
amounts of calf RNase HI. One digestion productwas 12nucleotides long, indicated by an arrow. The calf RNase HI used in this study had no activity against double and single strand So, the uniqueproduct presumably was DNA (data not shown). the first 12 nucleotides of the RNA primer. The gel was basically free of any other oligomers shorter than the12-mer. This experiment indicates that thecalf RNase HI was actingendonucleolytically on initiator RNA. The 13th nucleotideshould stillhave beenlabeledsince RNase HI digestion leaves a 5' phosphoryl group at the cleavage site. This remainingU residue is attached to the38-nucleotide long downstream DNA segment. The resultant 39nucleotide long product can also be seen in lanes3 and 4 . A double band was evident at the expected position of the 39-nucleotide long segment. This doublet was most likely the result of inexact elongationof the DNA segment by Sequenase. Apparently Sequenase added an additional, non-template directed nucleotide, overhanging the5' end of the templateDNA, as has previously been observed (Molecular Biology Reagents/ Protocol, U. S. Biochemical Corp.).Interestingly, use of the inexactly elongated substrate demonstrated that the RNase HI cleaved off a 12 nucleotide long RNA segment regardlessof the length of the DNA. Fig. 2B shows the same setof reactions but using substrates ( P ) and (SI (Fig. lA). This allowed comparison of three substrates having different RNA primer lengths anddifferent sequences around the RNA-DNAjunction (seeFig. 1B). The major RNase HI digestion products were 20 (Fig. 2 B , lanes 6, 7, and 8) or 30- (lanes 10, 11, and 12) nucleotide long fragments, indicated by arrows for substrates ( P ) and (SI, respectively. These experiments demonstrated that calf RNase HI cleaves initiator RNA by making a single cut one nucleotide before the
Substrate Specificity of Calf RNase HI
25924 A Substrate:
H RNase HI: - A
Substrate: RNapHI:
-
P
5 6 7 8
-
P Substrate: H Tone: 4 4
S d
91011 12
I 2 3 4 5 6
S
7 8 91011 I2 131415161718
M
M
*
-51
FIG.3. Time course of digestion of substrates (H), (P), and ( S ) by calf RNase HI. Products were analyzed after the following digestion times: lunes I , 7, 13: 0 min; lunes 2 , 8 , 14: 1 min; lunes 3, 9, 15: 5 min; lunes 4,10, 16: 10 min; lunes 5, 11, 17: 20 min; lunes 6, 12, 18: 30 min. The arrows at the left indicate the specific cleavage products, which are 12.20, and 30 nucleotides long from substrates (H), ( P ) ,and (S), respectively. The amount of the enzyme used in this experiment was 0.15 unit for each of lunes 1-18.
FIG.2. Calf RNase HI made asingle cut one nucleotide before the RNA-DNAjunction on the three substrates tested. A, enzyme titration of calf RNase HI on substrate (H). Increasing concentrationsof enzyme are represented by the triangle above the figure. The arrow indicates the RNase HI-specific digestion product. The amountsof enzyme used were: lune I , no enzyme; lune 2, 0.15 unit; lune 3, 0.3 unit; and lune 4, 0.6 unit. Lane M shows mobilities of the 13-, 21-, and 31-nucleotide long RNA transcripts made in vitro. These were applied a s molecular weight markers. Lune R shows fractionation of a 5' "'Plabeled 140 long unrelated RNA segment that had been subjected to limited alkaline hydrolysis. These two marker ladders were used in this and the subsequent figures. B , enzyme titration of calf RNase HI on substrates ( P , lunes , 5 4 ) and (S, lunes 9-12). The amounts of the enzyme used to generate the products inlunes 5-8 and lunes 9-12 are parallel to those used in lunes 1 4 . In lune 9, there is a 31-mer band present as a contaminant. The band can be observed in this and subsequent experiments. It is easily recognized and discounted, and has no effect on the production of specific cleavage products (see Fig. 3).
R31
Slrbnmte: RNase HI:
N
-A - A M R
1 2 3 4
5 6 7 8
31-
20-
.
10-
RNA-DNA junction. Thecleavage was not RNA-DNA sequence or RNA length directed. When digestion was carried out over a time course withlower enzyme concentration (Fig. 3), the specific cleavage products were consistently observed. This result indicates that these products did not exist transiently.They predominated throughout the reaction. The Substrate Structure Requirements for Calf RNase HI Recognition-To test whether RNase HI directs its specific cleavage on substrates ( H I , (P)and ( S ) by recognizing the RNA-DNA junction, we subjected substrates (R31)and ( N )to cleavage (see Fig. lA 1. The substrate (R31)has only an internally labeled 31-nucleotide RNA primer annealed to the template. The substrate ( N ) has a 20-nucleotide-long unlabeled DNA annealed downstreamof the 31-mer RNA, separated by a nick at the RNA-DNA junction. Calf RNase HI digested the RNA strand of these two substrates as shown in Fig. 4. However, there was no specificity of cleavage observed. Furthermore, the presence of the unattached DNA segment did not alter the distribution of the enzyme digestion pattern (lanes 1 4 versus lanes 5-8).The stable annealingof the 20-mer DNA to substrate ( N ) was confirmed by its decreased mobility in native gel elecirophoresis (data not shown). The absence of mononucleotide products indicated that this nonspecific cleavage of substrates (R31)and ( N )was the result of calf RNase HI endonucleolytic activity, rather than anylow level exonuclease contamination in the preparation (Fig. 4). RNase HI digests theRNA strand of an RNA/DNA hybrid en-
1-
FIG.4. Calf RNase HI lost cleavage specificity when the substrate lacked a covalent bond at the RNA-DNAjunction. Lane M shows a 31-nucleotideRNA primer used a s a molecular weight marker. Lane R shows an RNAladder prepareda s described in the legend to Fig. 2. The numbers a t t h eleft indicate the sizeof components inthe ladder. The RNase HI amounts used were:lunes I and 5, no enzyme; lunes 2 and 6, 0.15 unit; lunes 3 and 7, 0.3 unit; lunes 4 and 8,0.6 unit.
donucleolytically as reported by other investigators using calf thymus and other RNases HI (Cathala et al., 1979; Busen, 1980; Kanvan et al., 1983; Hagemeier and Grosse, 1989). We also examined digestion of a long segment of RNA annealed to a DNA template (Fig. 5). This substrate(R81,see Fig. lA)was digested by calf RNase HI, and the extent of digestion increased with increasing concentration of the enzyme. However, there was no evidence of specificity for the positions of cleavage. We note that products shorter than6 to 8 nucleotides are not shown. This result demonstrates thatcalf theRNase HI
Substrate Specificity Calf of Substme: RNaseHI:
R81
A
- A
".
1 2
RNase HI MnZ+(mM):
~
~
- --
MgZ* (rnM): 0 2 5 10 15
3 45
- 81
25925 B
0 0.5 2 4 6
.-
.- -
Mn2* (rnM): o
o
0.1 0 . ~ 0 . ~
Mg2* (mM): 0 5 5 5 5 1112 13 14 (5
1 2 3 4 5 6 7 8 910
- 51
.51
- 31 - 20 - 21
- 13
F
-
FIG.5. Calf RNase HI specific cleavage requires an Okazaki fragment-type substrate. The numbersat the right side of the figure show the mobility of a mixture of RNA transcripts made in uitro, and used here as markers. The amounts of enzyme used were: lune 1, no enzyme; lune 2, 0.08 unit; lune 3, 0.15 unit; lune 4, 0.3 unit; lune 5,0.6 unit.
FIG.6. Cation requirement forcalf RNase HI specific cleavage. In both Panels A and R , the substrate( P )was used.The numhers at the right side of the figure indicate the original fragment (51-mer), the (31-mer)after RNase cleaved RNA(2O-mer) and the remaining segment HI cleavage. The concentrationsof the Mg2' and Mn2+are listedon the top of the figure. The RNaseHI used to make reaction products in each of lunes 1-15 was 0.3 unit.
sible role of Mn2+will be discussed. digests randomly, when it acts on the standard RNase H subComparison of Calf RNase HI with HN-RT and E. coli strate. RNases H-HIV minus strand synthesisis primed by the celOverall, the above results show that cleavage specificity re- lular lysine t-RNA (Skalka andGoff, 1993). The HIV-RT intrinquires an Okazaki fragment-like structure, in which an RNA sic RNase H removes this tRNA primer by making a specific segment is attached covalently to a DNA segment at thejunc- cleavage and leaving one ribonucleotide on the 5'-end of the newly synthesized minus strand DNA (Furfine and Reardon, tion point. The Cation Requirement for Calf RNase HI Specific 1991; Pullen et al., 1992). Synthesis of the HIV plus strand M e or Mn2+ DNA is thought tobe initiated on the minus strandDNA temCleavage-Since RNase HI isknown to use either (Biisen, 1980)as cofactors, we tested thecation requirement for plate from a 19-nucleotide long segment of RNA designated the cleavage specificity using substrate (PI. We observed specific polypurine tract (Skalka and Goff, 1993). In reactions that cleavage over a range of M e concentrations up to15 mM (Fig. model events in theHIV plus strand synthesisreaction, Huber 6 A , lanes 1-5). Interestingly, the cleavage activity increased as and Richardson (1990) showed that the intrinsicRNase H acvia the M e concentration was raised to10 mM, but declined at 15 tivity of HIV-RT releases thepolypurine tract primer intact mM. The intracellular concentrations of M e are 5 to 10 mM a cleavage at theRNA-DNA junction. To determine whether the cleavage specificity that we ob(Linder, 1991).This experiment demonstrates that the physiological concentration of M e is required for calf RNase HI to served with calf RNase HI isexhibited by all RNase H enzymes, perform structure-specific cleavage most efficiently. we exposed our model Okazaki fragments to digestion by HIVWe also observed increasing cleavage as theMn2+concentra- RT. Fig. 7 shows the digestion pattern of HIV-RT on substrates tion was raised to 6 mM (Fig. 6 A , lanes 6-10).However, the (H), (P)and (SI. Lanes 1 to 4 show results with substrate (HI. overall enzymatic activity was less when Mn2+was the sole There were no digestion products with this substrate, which cation. Furthermore, radioactivity lost from the starting sub- has an RNA primer 13 nucleotides long (lanes 14). However, strate wasconverted to thespecific product at a lower efficiency HIV-RT produced the same length cleavage products on subthan in the presence of M e . Although the specific product was strates ( P ) and ( S ) (lanes 5-8 and 9-12, respectively). The the most evident of all products, i t was accompanied by many length of RNA cleaved off was 16-17 nucleotides. This obserother products, all present a t low concentration. This indicates vation is consistent with previous work published from this that digestion was much less specific. laboratory (DeStefano et al., 1993), which demonstrated that We then carried out an experiment to determine whetherHIV-RT the initial cleavages occurred a t a fixed distance of 15-21 effect of Mn2+is dominant when both M e and Mn2+are pre- nucleotides from the 5' terminus of the RNA. Evidently, with sent. Physiological concentrations of Mn2+are 0.1-1% that of the 13-nucleotide long RNA, the preferred cleavage site is past the RNA-DNA junction. Sincethe DNA could not be cleaved, no M e (Linder, 1991). In an attempt to approximate cellular concentrations of these cations, we held the M e at 5 mM and cleavage occurred at all (Fig. 7, lanes 14). titrated theMn2+ concentrationfrom 0 to 0.5 mM. Fig. 6B shows We have also tested the specificity of E. coli RNase H on that cleavage specificity was maintained throughout this ex- substrates (HI,(P)and (5'). E.coli RNase H degradedthe RNA periment. However, the overall cleavage activity was reduced primers in a random fashion (Fig. 8). Multiple cleavage sites as the level of Mn2+was increased. This result indicates that were observed when the longest of the threeRNA primers was cleavage specificity is likely to be manifested in vivo. The pos- employed (substrate ( S ) ,lanes 9 to 12).
25926
Substrate Specificity Subslgle:
H
-
P
-
S
-
HIV-RT d d M R 1 2 3 4 5 6 7 8 9101112
31
RNase of Calf
HI Stbabae: E.mt MR
H P S - 4- 4 - 4 12 3 4 5 6 7 8 9101112
31 ~
21-
13.
FIG.7. Cleavage of substrates ( H ) , (P), and ( S ) using HIV-RT. The lanes M and R are equivalent to those described in the legend to Fig. 2. The amounts of HIV-RT used were: lanes 1,5,9: no enzyme; lanes 2, 6, 10: 1 unit; lanes 3, 7, 11: 2.5 units; lanes 4 , 8, 12: 5 units.
21
13
FIG.8. Cleavage of substrates ( H ) , (P), and ( S ) using E. coli RNase H. The lanes M and R are equivalentto those described in the legend to Fig.2. The amountsof E. coli RNase H used were: lanes 1,5, 9: no enzyme; lanes 2, 6, 10: 0.125 unit; lanes 3, 7, 11: 0.25 unit; lanes 4,8, 12: 0.5 unit.
nucleotide requires a second nuclease, the calf 5' to 3' exonuclease (Turchi et al., 1994). Although RNase HI cleaves most RNA hybridized to DNA by a nearly random endonucleolytic mechanism (Cathala et al., 1979; Busen 1980; Kanvan et al., 1983; Hagemeier and Grosse, 1989), Okazaki fragment structure appeared be tocleaved with greater specificity (Turchi et al., 1994). Our current results DISCUSSION suggest that cleavage specificity requires an RNA initiated Investigating mouse cell DNA replication reactions in vitro, DNA segment annealed toa DNA template. Variations in this Goulian et al. (1990) demonstrated the involvement of RNase Okazaki-type structure result in loss of specificity (Figs.4 HI in removal of initiator RNA from DNA segments. In their and 5). The distinct RNAlDNA conformation around the RNA-DNA study, RNaseHI alone removed about 80% of the RNA in primers madeby DNA polymerase dprimase. In thepresence of the junction may serve as a recognition point for RNase HI binding. mouse 5' to 3' exonuclease, the remaining primer RNA was One possible reason for the recognition of RNA-DNAjunctions completely eliminated. We later showed that calf RNase HI is helical structure. While RNA-DNA hybrids form into an A could remove all but the last ribonucleotide of RNA initiated helix, DNA double strands make aBhelix(Kornberg and DNA(Turchi et al., 1994). The calf 5' to 3' exonuclease was then Baker, 1992). It is likely that theRNA-DNA junction region of able to cleave off that last nucleotide. an Okazaki fragment annealeda DNA to template hasa unique Analyzing SV40 DNA replication in vitro,Ishimi et al. (1988) intermediate helical form, which might direct the enzyme recobserved an approximately 2-fold stimulation by RNase HI of ognition and the specific cleavage. A recent study of HIV-RT the generationof covalently closed circular (form I) DNA. Waga crystal structure showed that the RT heterodimer, bound to a et al. (1994) and Waga and Stillman (1994) demonstrated that DNA primer-template, hadregions of A-DNAand B-DNA sepa5' to 3' exonuclease and DNA ligase I, with other replication rated by a significant bend (Kohlstaedt et al., 1992). proteins, are essential components for production of form I It iscurious that the uniquecleavage site for calf RNase HI SV40 DNA in vitro. Although they did not detect a n effect of is one nucleotide upstream of the RNA-DNA junction. It seems added RNase HI, they pointed out that their DNA ligase I that cleavage right at the junction point would be the most preparation contained RNase HI that could have participated efficient way to prepare Okazaki fragmentsfor ligation. Interestingly, there is evidence suggesting cleavage at the junction in maturation of the Okazaki fragments. In the current report,we have examined the cleavage spec- in other systems. T4 RNase H releases the RNA pentamer ificity of calf RNase HI on model Okazaki fragments. These aremade by the T4 gene 61 and 41 primasehelicase completely RNA initiated DNA segments annealed to templateDNA. The from the 5'-end of DNA, making products from dimer to penenzyme recognizes the RNA-DNA junction. It makes a single tamer length (Hollingsworth and Nossal, 1991). However, the endonucleolytic cleavage, one nucleotide upstream of the RNA- authors could not exclude the action of a combination of 5' to 3' DNA junction. It cleaves off most of the RNA primer as an exonuclease with RNase H on their RNA primer removal, beintact segment. It leaves the 5' phosphorylated junction ribo- cause T4 RNase H also displays a 5' to 3' exonuclease activity. nucleotide, attached to the DNA portion of the Okazaki frag- The RNase H of HIV-RT could completely remove the polypument. Cleavage specificity is independent ofRNA primer rine tractRNA after extension with DNA (Huber andRichardlength or the sequence around thecleavage site. Thisspecificity son, 1991). The major RNA product was the fully intact polyis not sharedby all other RNases H. Removal of the last ribo- purine tract, and no ribonucleotides were left at the 5' end of Since neither HIV-RT nor E. coli RNase Hperform structurespecific cleavage on initiator RNAof the substrates we employed, we conclude that not all RNases H are designed to recognize, and specifically cleave, the initiatorRNA of Okazaki fragments.
Substrate Specificityof Calf RNaseHI
25927
the DNA. Hydrolysis of the phosphodiester bond at the RNA- primase/polymerase a through physical interaction, in yeast DNA junction by AMV-RT RNase H has also been observed and Drosophila, respectively. Masutani et. al. (1990) had iden(Furfine and Reardon, 1991). However, HIV-RT leaves one ri- tified a DNA primase stimulatoryfactor from mouse FM3A cell bonucleotide at the 5'-end of the DNA when it digests away the having a n RNase H activity. These observations suggest that tRNA lysine primer (Furfine and Reardon, 1991; Pullen et al., RNase HI has a multifunctional role in initiator RNA process1992). The sameenzyme has yeta different cleavage preference ing that mustbe examined in more detail before its participawhen exposed to anRNA segment annealed to DNA (DeStefano tion in DNA replication can be completely understood. et al., 1993) or our model Okazaki fragments (Fig. 7). It produces endonucleolytic cleavages 15-21 ribonucleotides from 5'REFERENCES end of the RNA. The position of cleavage is independentof the RNA length, although influenced somewhat by RNA sequence Biisen, W. (1980) J. Biol. Chem. 255,9434-9443 Biisen, W., Peters, J. H., and Hausen, P. (1977)Eur. J. Biochem. 74, 203-208 (DeStefano et al., 1993). In spite of the apparent capacity of Cathala, G., Rech, J., Huet, J., and Jeanteur, P. (1979)J . Biol. Chem. 254, 7353other RNases H to cleave the junctionribonucleotide, we have 7361 seen no evidence whatsoever that calf RNase HI can actat that DeStefano, J. J., Mallaber, L. M., Fay, P. J., and Bambara, R. A. (1993) Nucleic Acids Res. 21, 43304338 site. DiFrancesco, R. A,, and Lehman, I. R. (1985) J. Bid. Chem. 260, 14764-14770 Using two-dimensional NMR techniques and distance geom- Eder, P. S., and Walder, J. A. (1991) J. Biol. Chem. 266,6472-6479 etry calculations, Nakamura et al. (1991) were ableto construct Funnel, B. E., Baker, T. A,, and Komberg,A. (1986)J . Bid. Chem. 261,5616-5624 E. S., and Reardon, J. E. (1991)Biochemistry 30, 7041-7046 a DNA/RNAhybrid complex model with E.coli RNase H. As we Furfine, Goulian, M., Richards, S. H., Heard, C. J., and Bigsby, B. M. (1990)J . Biol. Chem. observed, E. coli RNase H does not display junction-specific 265, 18461-18471 cleavage on Okazaki fragment model substrates (Fig. 8). This Hagemeier, A,, and Grosse, F. (1989)Eur. J . Biochem. 185, 621428 suggests that determinationof three dimensional structure of Hollingsworth, H. C., and Nossal, N. G. (1991) J . Bid. Chem. 266, 1888-1897 Huber, H. E., and Richardson, C. C. (1991) J . Biol. Chem. 265, 10565- 10573 the mammalian RNase HI complex with DNA, and a compar- Ishimi, Y., Claude, A,, Bullock, P., and Hurwitz, J . (1988) J. Biol. Chem. 263, 19723-19733 ison with E. coli RNase H, might reveal information on the mechanism of junction recognition. The absence of cleavage Karwan, R., Blutsch, H., and Wintersberger,U. (1983)Biochemistry22,5500-5507 Keller, W. (1972) Proc. Nut1 Acad. Sci. U.S. A. 81, 3993-3997 specificity in the bacterialenzyme also suggests that there are Kitahara, N., Sawai, Y., and Tsukada, K. (1982) J. Biochem. (Tokyo)92,855- 864 fundamental differences in the steps involved in removal of Kohlstaedt, L. A,, Wang, J., Friedman, J. M., Rice, P. A,, and Steitz, T. A. (1992) Science 256, 1783-1790 initiator RNA between prokaryotes and eukaryotes. A,, and Baker,T.A. (1992) DNA Replication,W. H. Freeman & Co., New We have shown that either Mg2' or Mn2+can serve as cofac- Komberg, York tors for cleavage of Okazaki fragmentmodel substrates by calf Linder, M. C. (1991) Nutritional Biochemistry and Metabolism with Clinical Applications, Elsevier Science Publishing Company, Inc., New York RNase HI. However, the enzyme displays highest activity in Masutani, C., Enomoto, T., Suzuki, M., Hanaoka, F., and Ui, M. (1990) J . Biol. the presence of Mg2' (Fig. 6A). Additionally, while nearly all Chem. 265, 10210-10216 cleavage in the presence of Mg2' is specific, activity in the Murante, R. S.,Huang, L., Turchi, J . J., and Bambara,R. A. (1994) J . Biol. Chem. 269, 1191-1196 presence of Mn2+results ina large component of random cleavNakamura, H., Oda, Y., Iwai, S., Inoue, H., Ohtsuka, E., Kanaya, S., Kimura, S., age products. In the presence of both metal ions, the ability of Katsuda, C., Katayanagi, K., Morikawa, K., Miyashiro, H., and Ikehara, M. (1991) Proc. Natl. Acad. Sei. U, S. A . 88, 11535-11539 Mg2'to promote specific cleavage appears to predominate(Fig. 6B). Biisen (1980) also observed different behavior of the en- Ogawa, T., and Okazaki, T.(1984) Mol. Gen. Genet. 193,321-237 Pullen, K. A,, Ishimoto, L. K., and Champoux, J. J. (1992) J . Virol. 66, 367-373 zyme depending on whether it was acting in the presence of Roewekamp, W., and Sekeris, C. (1974)Eur. J. Biochem. 43, 405-413 Mg2' or Mn2+. Heproposed that the enzyme exists in two dif- Rong, Y. W., and Carl, P. L. (1990) Biochemistry 29,383-389 J., Fritsch, E. F., and Maniatis, T. (1989) in Molecular Cloning: A ferent conformations depending on the type of divalent cation Sambrook, Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY activation. We propose that the Mg2' dependent form is de- Skalka, A. M., and Goff, S. P. (1993) Reverse Panscriptuse, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY signed to perform specific cleavage during Okazaki fragment Stein, H., and Hausen,P. (1969) Science 166,393-395 processing. The role of the Mn2+ activatedform is less clear. It Turchi, J. J., and Bambara, R. A. (1993) J . Biol. Chem. 268, 15136-15141 may be involved in the degradation of RNA/DNA hybrids in the Turchi, J. J., Huang, L., Murante, R. S., Kim, Y., and Bambara, R. A. (1994)Proc. Natl. Acad. Sci. U.S. A., in press cell, possibly products of transcription. Vonwirth, H., Frank, P., and Biisen, W. (1989)Eur: J. Biochem. 184, 321-329 Both Kanvan et al. (1983) and DiFrancesco and Lehman Waga, S., and Stillman, B. (1994) Nature 369, 207-212 (1985) demonstrated a stimulating effect of RNase H on DNA Waga, S., Bauer, G., and Stillman, B. (1994)J . Biol. Chem. 269, 10923-10934
