Junction ribonuclease: an activity in Okazaki fragment processing.

Abstract

The initiator RNAs of mammalian Okazaki fragments are thought to be removed by RNase HI and the 5'-3' flap endonuclease (FEN1). Earlier evidence indicated that the cleavage site of RNase HI is 5' of the last ribonucleotide at the RNA-DNA junction on an Okazaki substrate. In current work, highly purified calf RNase HI makes this exact cleavage in Okazaki fragments containing mismatches that distort the hybrid structure of the heteroduplex. Furthermore, even fully unannealed Okazaki fragments were cleaved. Clearly, the enzyme recognizes the transition from RNA to DNA on a single-stranded substrate and not the RNA/DNA heteroduplex structure. We have named this junction RNase activity. This activity exactly comigrates with RNase HI activity during purification strongly suggesting that both activities reside in the same enzyme. After junction cleavage, FEN1 removes the remaining ribonucleotide. Because FEN1 prefers a substrate with a single-stranded 5'-flap structure, the single-stranded activity of junction RNase suggests that Okazaki fragments are displaced to form a 5'-tail prior to cleavage by both nucleases.

Full text

Page 1

Proc. Natl. Acad. Sci. USA
Vol. 95, pp. 2244–2249, March 1998
Biochemistry
Junction ribonuclease: An activity in Okazaki fragment processing
RICHARD S. MURANTE, LEIGH A. HENRICKSEN, AND ROBERT A. BAMBARA*
Department of Biochemistry and Biophysics, and Cancer Center, Box 712, University of Rochester School of Medicine and Dentistry, 601 Elmwood Avenue,
Rochester, NY 14642
Communicated by I. Robert Lehman, Stanford University School of Medicine, Stanford, CA, December 29, 1997 (received for review
September 15, 1997)
ABSTRACT
fragments are thought to be removed by RNase HI and the
5?-3? flap endonuclease (FEN1). Earlier evidence indicated
that the cleavage site of RNase HI is 5? of the last ribonucle-
otide at the RNA-DNA junction on an Okazaki substrate. In
current work, highly purified calf RNase HI makes this exact
cleavage in Okazaki fragments containing mismatches that
distortthehybridstructureoftheheteroduplex.Furthermore,
even fully unannealed Okazaki fragments were cleaved.
Clearly, the enzyme recognizes the transition from RNA to
DNA on a single-stranded substrate and not the RNA?DNA
heteroduplex structure. We have named this junction RNase
activity. This activity exactly comigrates with RNase HI
activity during purification strongly suggesting that both
activities reside in the same enzyme. After junction cleavage,
FEN1 removes the remaining ribonucleotide. Because FEN1
prefers a substrate with a single-stranded 5?-flap structure,
the single-stranded activity of junction RNase suggests that
Okazaki fragments are displaced to form a 5?-tail prior to
cleavage by both nucleases.
The initiator RNAs of mammalian Okazaki
TwoclassesofRNaseH,RNaseHIandII,havebeenidentified
in a variety of eukaryotes, purified and characterized (1–8).
Class I enzymes have a native molecular mass of ?68–90 kDa
and are able to use both Mn2?and Mg2?as cofactors. In
contrast,classIIRNaseHisapproximately30kDaandcanuse
only Mg2?as a cofactor. Both forms of RNase H randomly
degrade RNA in long RNA?DNA hybrid structures (2, 9).
RNase HI is involved in the removal of the initiator RNA
primers of Okazaki fragments, an essential step during chro-
mosomal DNA replication of the lagging strand (10). These
RNAs, 7–14 nucleotides in length, are generated by the
primase activity of DNA polymerase ? (pol ?) and prime DNA
synthesis of both the leading and lagging strands at the DNA
replication fork. The pol activity of this same enzyme then
extends each primer with deoxynucleotides. Pol ? is thought to
be replaced by pol ? that continues synthesis to the next
downstream fragment. There is considerable evidence that the
RNA is removed by the combined action of RNase HI and a
5? to 3? exo?endonuclease (11–13). This latter enzyme has
been called flap endonuclease (FEN1) (14) in mammals and
is known as RAD two homolog (RTH1) and RAD27 in
Saccharomycescerevisiae(15).RemovaloftheRNAisrequired
before DNA synthesis and ligation can complete the replica-
tion process.
The exact functions of RNase HI in DNA replication have
been the subject of several studies. Drosophila RNase HI was
found to stimulate synthesis by pol ? (4). In addition, RNases
H from yeast and calf thymus (3, 6) associate with and
stimulate their cognate pol ?. These observations suggest that
RNaseHIandpol?formacomplexinvivocapableofsynthesis
and removal of RNA primers. Reconstitution assays with both
mousecellandSV40replicationproteinsrequireRNaseHIfor
the maturation of lagging strands in vitro (12, 13).
Although cleavage of long RNA?DNA hybrids is random,
cleavage of certain substrates is clearly structure specific. Eder
et al. (16) found that when a series of ribonucleotides were
followed by a 3? DNA segment and annealed to DNA to form
a duplex, human RNase HI preferentially cleaved to leave a
DNA segment with a 5? monoribonucleotide. In reactions
reconstituting Okazaki fragment processing, RNase HI from
calf thymus similarly cleaved the RNA-DNA strand 5? to the
junctional ribonucleotide. At the enzyme concentrations used,
the RNA primer was released by a unique cleavage with no
further degradation (11). RNase HI is capable of this struc-
ture-specific cleavage regardless of the length of the RNA or
the sequence context of the RNA-DNA junction (17). There-
fore, RNase HI has been described as both a random and
structure-specific endonuclease.
Here we show that cleavage at the RNA-DNA junction does
not require the presence of any duplex structure. Highly
purified RNase HI preparations contain a junction RNase
activity that recognizes distinctly different aspects of substrate
structure than conventional RNase H activity. The nature of
the substrate, either an RNA-DNA junction or an RNA?DNA
hybrid, determines that cleavage mechanism is employed.
EXPERIMENTAL PROCEDURES
Materials. Synthetic DNA oligonucleotides were synthe-
sized by Genosys Biotechnologies (The Woodlands, TX).
[?-32P]-radiolabeled CTP and UTP (800 Ci?mmol; 1 Ci ? 37
GBq) and [?-32P] ATP (3000 Ci?mmol) were purchased from
Dupont?NEN. Poly[8H]adenylic acid (?15 Ci?mmol) was
obtained from Amersham Life Sciences. Ribonucleotides and
deoxyribonucleotides were purchased from Pharmacia LKB.
Sequenase (version 2.0), T3 RNA polymerase, and T4 polynu-
cleotide kinase were from Amersham Life Sciences. Restric-
tion endonucleases were purchased from New England Bio-
labs. The plasmid, pBS?, was obtained from Stratagene. All
other reagents were from Sigma.
Purification of RNase HI. Two separate approaches were
undertakentopurifyRNaseHIbothstartingwith100goffetal
calf tissue. The first method, as described by Busen (18), was
followed through hydroxylapatite chromatography with the
exception that the (NH4)SO4precipitation step was omitted.
All columns were equilibrated with buffer A (30 mM Tris?HCl,
pH 7.8?0.1% 2-mercaptoethanol?0.1 mM phenylmethylsulfo-
nyl fluoride?15% glycerol) prior to use. Pooled fractions of
peak RNase HI activity from the hydroxylapatite column were
dialyzed into buffer A and applied to a 40 ml Affi-gel Blue
column previously equilibrated with buffer A. After the col-
umn was washed with buffer A containing 250 mM MgCl2,
RNase HI activity was eluted with buffer A containing 1.0 M
The publication costs of this article were defrayed in part by page charge
payment. This article must therefore be hereby marked ‘‘advertisement’’ in
accordance with 18 U.S.C. §1734 solely to indicate this fact.
© 1998 by The National Academy of Sciences 0027-8424?98?952244-6$2.00?0
PNAS is available online at http:??www.pnas.org.
Abbreviations: FEN1, flap endonuclease 1?RAD two homolog 1; pol,
DNA polymerase; nt, nucleotide.
*To whom reprint requests should be addressed.
2244

Page 2

MgCl2. Peak fractions were combined and dialyzed into buffer
B (buffer A with 50 mM KCl) and applied to a 6 ml Affi-gel
Blue column. The column was washed sequentially with buffer
A containing 0.1, 0.3, 0.5, and 1.0 M KCl. The peak activity
eluted with 300 mM KCl and was dialyzed against buffer B.
The dialysate was loaded onto a 1 ml Mono-Q column and
eluted with a linear salt gradient of 50–500 mM KCl. RNase
HI eluted at ?200 mM KCl and stored at 4°C. The specific
activity of the peak Mono-Q fraction was 134,000 U?mg. One
unit of RNase HI produces 1 nmol of trichloroacetic acid
soluble ribonucleotide monomer from [3H]poly(rA)?
oligo(dT) in 15 min at 37°C.
The second method for purification of RNase HI was
adaptedfromKawanetal.(3).Afteracalfthymushomogenate
was prepared (11), DNA was removed by polyethylene glycol
precipitation (3). Affinity chromatography over native DNA-
cellulose and hydroxylapatite chromatography also were per-
formed according to this procedure (3). Fractions containing
the peak RNase H activity from the hydroxylapatite column
were dialyzed into buffer A (pH 7.8) with 10 mM NaCl and
applied to a 1 ml Q-Sepharose column. Activity was eluted
with a 10 ml linear salt gradient from 10 to 500 mM NaCl. The
active fractions were pooled, dialyzed against buffer A (pH
8.3) with 10 mM NaCl, and loaded onto a 1 ml SP-Sepharose
column. A 10 to 500 mM linear salt gradient was used to elute
RNase HI with a similar specific activity as the above prepa-
ration.
Glycerol Gradient Sedimentation. Approximately 20 ?g of
RNase HI from the peak Mono-Q fraction was diluted to 15%
glycerol in final volume of 50 ?l and applied to a 3.95 ml
glycerol gradient (15–35%) in buffer A with 100 mM NaCl.
Protein standards consisting of catalase (232 kDa), BSA (68
kDa), and RNase (13.7 kDa) were applied to parallel gradi-
ents. Density gradient sedimentation was performed by cen-
trifugation in a Beckman SW60.1 rotor for 72 hr at 45,000 rpm.
Gradients were fractionated from the top by careful removal
of 100 ?l aliquots from the meniscus.
Preparation of Okazaki Fragment Substrates. Internally
labeled RNA primers were prepared by T3 RNA pol directed
run-off transcription on pBS(?) plasmids digested with Hin-
dIII (19) in the presence of 625 pmol [?-32P]CTP and
[?-32P]UTP, 1 mM ATP and GTP. For 5?-end-radiolabeled
primers, the RNA13ntwas incubated with T4 polynucleotide
kinase in the presence of [?-32P]ATP as described (30).
Following separation on a denaturing 20% polyacrylamide gel,
RNA primers were purified and annealed to the DNA tem-
plate T1 (5?-TACCCGGGGATCCTCTAGAGTCGACCTG-
CAGGCATGCAAGCTTTTGTTCCCTTTAGTGAGGGT-
TAATTTCGAGCT-3?). Then, the RNA primers were ex-
tended with 10 units of SequenaseTMand 1.3 mM of each of
the four dNTPs. These RNA-DNA primers (5?-GGGAACA-
AAAGCUTGCATGCCTGCAGGTCGACTCTAGAGGA-
TCCCCGGGTA-3?) were isolated after separation through a
denaturing 15% polyacrylamide gel and annealed to the
appropriate template oligomer as specified in the figure leg-
ends. The RNA portion of RNA-DNA primer sequence is
underlined and in italics.
DNA sequences used to create mismatched Okazaki frag-
ments were: DNA mismatch (5?-TACCCGGGGATCCTCT-
AGAGTCGACCTGCAGGCATGCTAGCTTTTGTTCCC-
TTTAGTGAGGGTT-3?),
(5?-TACCCGGGGATCCTCTAGAGTCGACCTGCAGGC-
ATCCAAGCTTTTGTTCCCTTTAGTGAGGGTT-3?), and
downstream RNA mismatch (5?-TACCCGGGGATCCTCT-
AGAGTCGACCTGCAGGCATGGAAGCTTTTGTTCCC-
TTTAGTGAGGG-3?). The underlined nucleotide indicates
the site of mismatch.
RNase HI Assays. Assays using Okazaki fragment substrates
contained 60 mM bis?Tris (pH 7.0), 10 mM MgCl2, 5 mM
2-mercaptoethanol,5%glycerol,100mg?mlBSA,and?2fmol
upstreamRNAmismatch
of substrate per 12 ?l. Reactions were initiated by the addition
of 1 ?l of enzyme preparation that was diluted as noted in the
figure legends, incubated at 37°C for times indicated and
terminated with 12 ?l formamide loading buffer [90% form-
amide(vol?vol)?10mMEDTA?0.01%withbromophenolblue
and xylene cyanole]. Cleavage products were separated by
electrophoresis through a denaturing 12% polyacrylamide gel
(20) and analyzed using a Molecular Dynamics Phosphor-
Imager and IMAGEQUANT software (Molecular Dynamics).
RNase H assays with [3H]poly(rA)?oligo(dT) were per-
formed as described by Eder et al. (5) at 37°C for 10 min.
RESULTS
We initially began to explore the substrate specificity of RNase
HI to define its role in initiator RNA removal. We wanted to
verify that RNase HI, known to be a nearly random endonu-
clease on RNA?DNA hybrids, is responsible for the RNA-
DNA junction specific cleavage observed with model double-
FIG. 1.
endonucleolytic activity. Aliquots (1 ?l) of each fraction from the final
column of the first preparation (Mono-Q) were diluted 1:4000 and
tested for both RNase HI structure-specific and random cleavage
activity. (A) To generate a model Okazaki substrate, an internally
radiolabeled (32P)RNA13ntwas extended with DNA38ntand annealed
to a DNA template. Aliquots of the diluted fractions were assayed at
37°C for 10 min. Cleavage results in a specific RNA12ntfragment and
an oligonucleotide39nt containing the remaining ribonucleotide and
deoxynucleotides (indicated by arrows). (B) Random RNase HI
cleavage was assayed on a [3H]poly(rA)?oligo(dT) substrate. RNase
HI activity degrades the substrate releasing small oligonucleotides into
the acid soluble fraction that were quantitated by scintillation count-
ing. The percent of acid soluble radioactivity (?) was determined for
each fraction. Additionally, the percent of the Okazaki fragment
substrate converted to the RNA12ntproduct in A was determined and
also plotted (F).
RNase HI contains both a structure-specific and random
Biochemistry: Murante et al.Proc. Natl. Acad. Sci. USA 95 (1998)2245

Page 3

strandedOkazakisubstrates.Therefore,wepurifiedRNaseHI
to a greater extent than we reported (11). Throughout puri-
fication over ion and affinity resins, we monitored both
structure-specific and random endonucleolytic activity. Struc-
ture-specific cleavage was examined using a model Okazaki
fragment (Fig. 1). An internally radiolabeled oligonucleotide
containing a region of 13 nucleotides of RNA (RNA13nt)
immediately followed by 38 nucleotides of DNA (DNA38nt)
was annealed to a DNA template. After the substrate was
incubated with an aliquot of each protein fraction, the reaction
products were analyzed using a Molecular Dynamics Phos-
phorImager. Elution of the structure-specific cleavage activity
from the final column, Mono-Q, is shown in Fig. 1A. Cleavage
results in the formation of an RNA12ntand a 39 nucleotide
segment corresponding to the remaining ribonucleotide and
DNA. Smaller fragments represent additional nonspecific
cleavage by RNase HI. Random endonucleolytic activity was
measured on a standard RNase HI substrate [3H]poly(rA)?
oligo(dT) and detected as acid soluble radioactivity. The
amount of random cleavage from the Mono-Q fractions was
quantitated by scintillation counting, and the data plotted as
the percentage of3H counts found in the acid soluble fraction
(Fig. 1B). Additionally, the data from the structure-specific
assay is presented as the percent of substrate converted to the
RNA12ntproduct (Fig. 1B). Clearly both structure-specific and
random endonucleolytic activities copurified. This provided
substantial evidence that RNase HI is responsible for specific
cleavage at RNA-DNA junctions.
pol ??primase lacks the ability to perform 3?-5? exonucleo-
lytic proofreading. Consequently, it frequently misincorpo-
rates both ribo- and deoxyribonucleotides during priming and
DNA synthesis (21, 22). This creates mismatched nucleotides
in the 5?-end region of Okazaki fragments. We anticipated that
a mismatch at or near the RNA-DNA junction would alter the
structure of the RNA?DNA heteroduplex and prevent RNase
HI from recognizing and cleaving the substrate. Failure of
RNase HI to initiate removal of RNA primers would prevent
completion of DNA replication.
Substrates were generated to examine the ability of the
RNaseHIpreparationtocleavetheRNA-DNAjunctioninthe
presence of a mismatched base pair. A 5?-radiolabeled oligo-
nucleotide of RNA13nt-DNA38nt was annealed to a DNA
template. The template strand was designed to generate a
specific mismatch with the RNA-DNA strand. Each substrate
wasincubatedwiththepeakfractionofRNaseHIactivityfrom
the Mono-Q column. During the reaction, aliquots were taken
at selected time points and analyzed as described in the
experimental procedures. Following their separation by
PAGE, products were quantitated by a Molecular Dynamics
PhosphorImager. When the site of mismatch was the de-
oxynucleotide at the RNA-DNA junction, RNase HI effi-
ciently cleaved the substrate (Fig. 2, lanes 1–7) to generate the
12 nucleotide RNA product. While this substrate had distorted
helical structure, the actual cleavage site, 5? of the last RNA,
was still able to form an RNA?DNA hybrid structure. There-
fore,wetestedcleavageonasubstratewiththemismatchatthe
ribonucleotide just upstream (Fig. 2, lanes 17–23) and just
downstream(lanes9–15)ofthecleavagesite.Again,RNaseHI
specifically and efficiently cleaved 5? of the junctional ribonu-
FIG. 2.
oligonucleotide was annealed to a DNA template designed to create a specific mismatch. Above the appropriate lanes, substrate structure
representations are given. Gray lines indicate RNA and solid lines depict DNA. A sawtooth segment marks the position of the mismatch relative
to the template strand. The peak fraction of Mono-Q diluted 1:4,000 was assayed with the indicated substrates as described. A time course of
endonucleolytic cleavage reveals that RNase HI efficiently cleaves in spite of a mismatch at the first deoxynucleotide (lanes 1–7) or a mismatch
attheribonucleotidedownstream(lanes9–15)orupstream(lanes17–23)ofthesiteofcleavage.Forallsubstrates,anRNA12ntfragmentwasreleased
from the substrate (arrow). The size marker (M) containing an RNA13nttranscript is also shown (lanes 8, 16, 24). Reaction times (min) are as
indicated.
Mismatches near the Okazaki fragment RNA-DNA junction do not lower cleavage efficiency. A 5?-radiolabeled (32P)RNA13nt?DNA38nt
2246Biochemistry: Murante et al.Proc. Natl. Acad. Sci. USA 95 (1998)

Page 4

cleotide. These data show that the nuclease does not require
complete hybridization at the cleavage site.
It was surprising that structure-specific cleavage occurred in
the presence of a mismatch because the random nuclease
activity of RNase HI depends on an RNA?DNA hybrid
structure. In these structures, RNA?DNA hybrids form A-type
helices, whereas DNA?DNA duplexes form B-type helices.
Presumably, the junction in RNA-DNA?DNA substrate
adopts an intermediate structure (23). We had anticipated that
this unique structure could signal junction-specific cleavage.
However, the mismatched substrates are still cleaved effi-
ciently, even though they must have considerable distortion in
the junction structure. Results of the mismatch experiment
show that recognition of the junction cleavage site by RNase
HI does not require the helical structure present at a normal
RNA-DNA junction in a heteroduplex. While surprising, this
conclusion is consistent with the observation that RNase HI
cleaves 5? of a single ribonucleotide embedded in a DNA-
RNA-DNA?DNA heteroduplex (16, 24), anticipated to have a
helical structure different from that of a duplex Okazaki
fragment substrate. This observation also explains why initi-
ator RNA mismatches do not disrupt mammalian DNA rep-
lication.
We next considered that the surrounding duplex of either
RNA?DNA or DNA?DNA may be needed for RNase HI
cleavage. To examine the significance of the RNA?DNA
heteroduplex upstream of the site of cleavage, we utilized
model Okazaki substrates lacking any RNA?DNA heterodu-
plex. A 5?-radiolabeled oligonucleotide of (32P)RNA13nt-
DNA38ntwas annealed to a DNA template complementary to
only the DNA sequence. Over time, RNase HI activity re-
leased an RNA12nt(Fig. 3A). This shows that specific cleavage
can occur in a substrate in which RNA?DNA hybrid cannot
form. We also tested RNase HI activity on the RNA13nt-
DNA38ntoligonucleotide in the absence of any DNA template.
Remarkably, a specific RNA12ntsegment was again released
(Fig. 3B) Specific cleavage was also seen with two single-
stranded Okazaki fragments having different sequences sur-
rounding the junction (not shown). Furthermore, in compar-
ison to a standard Okazaki substrate, the activity of RNase HI
with the unannealed substrates was similar to that of a fully
annealed substrate (Fig. 3C). The highly purified RNase HI
displayed a cleavage activity that is independent of any het-
eroduplex structure.
The ability of RNase HI to cleave a single-stranded junction
substrate was surprising given that activity on single-stranded
RNA had not been previously reported. Though our enzyme
preparation did not measurably degrade any of a single-
stranded [3H]poly(rA) substrate, the junction-specific cleavage
of single-stranded RNA-DNA substrates might still represent
a contaminating activity in the RNase HI preparation. There-
fore, we sought to verify that junction activity on both double-
and single-stranded Okazaki fragment substrates and RNA?
DNA hybrids were contained within RNase HI.
Mono-Q fractions with the highest RNase H and Okazaki
fragment junction cleavage activity were subjected to glycerol
gradient sedimentation. Fractions from the glycerol gradient
were analyzed on an 8–14% SDS-polyacrylamide gel stained
with silver (25) (Fig. 4A). Assays using either the standard
RNase H [3H]poly(dA)?oligo(dT) substrate or double- and
FIG.3.
representations, and reaction times (min) are given. Gray lines indicate RNA and solid lines depict DNA. Reactions were performed as described
in Fig. 2. (A) A 5?-radiolabeled oligonucleotide of (32P)RNA13nt?DNA38ntwas annealed to a DNA template complementary to only the DNA
sequence(5?-TACCCGGGGATCCTCTAGAGTCGACCTGCAGGCATGCA-3?).Overtime,RNaseHIactivityreleasedanRNA12ntproduct.(B)
RNase HI activity also was examined on the RNA13nt?DNA38ntoligonucleotide in the absence of any DNA template. Again, a specific RNA12nt
segment was released. (C) RNase HI activity on a fully annealed substrate generated the RNA12ntproduct at the same rate as in A and B. The
size marker (M) containing an RNA13nttranscript is indicated.
Junction-specificcleavagebyRNaseHIisindependentofanyRNA?DNAheteroduplex.Abovetheappropriatelanes,substratestructure
Biochemistry: Murante et al.Proc. Natl. Acad. Sci. USA 95 (1998)2247

Page 5

single-stranded Okazaki fragment substrates were performed
(Fig. 4B). Results demonstrate that the peak of enzymatic
activity on all three substrates cosediment. Cosedimentation is
so exact that the ratio of activities is essentially the same for
every fraction over the peak. They are also coincident with
protein bands at 70, 30, 27, and 18 kDa. This represents nearly
homogeneous RNase HI where the lower molecular weight
species are degradation products of the full-length 70-kDa
protein (7). RNase HI was also purified by a second method
that involved different types of fractionation from the first
procedure. Junction-specific cleavage activity on an RNA-
DNA substrate again coeluted with the random RNase HI
activity on the [3H]poly(rA)?oligo(dT) substrate. Fractions
from the final SP-Sepharose column were analyzed by SDS?
PAGE and showed the same 70-, 30-, 27-, and 18-kDa protein
bands as obtained following glycerol sedimentation (data not
shown).TheseresultsprovidecompellingevidencethatRNase
HI contains multiple nucleolytic activities.
The susceptibility of RNase HI activity to increasing salt
concentration was measured on the [3H]poly(rA)?oligo(dT)
andjunctionspecificsubstrates.Overall,bothactivitiesshowed
similar sensitivity to titration with KCl (0–500 mM). Total
activity on the two different substrates was inhibited 50% at
?200 mM KCl. This result further supports the conclusion that
both activities reside in the same protein. On the RNA-DNA
substrate, total activity consisted of both the specific junction
cleavage and several additional nonspecific cleavages. Inhibi-
tion of the nonspecific cleavages generally occurred at a lower
salt concentration than at the specific junction site. In effect,
the specificity for the junction can be increased at appropriate
salt concentration.
DISCUSSION
The original model for initiator RNA removal from mamma-
lian Okazaki fragments predicts a specific cleavage of the
duplex RNA by RNase HI. This is followed by exonucleolytic
removal of the remaining junction ribonucleotide by the FEN1
nuclease. Both enzymes are required because FEN1 nuclease
cannot exonucleolytically cleave a triphosphorylated 5?-end,
andRNaseHIcannotremovethejunctionribonucleotide(11).
However, FEN aptly stands for ‘‘flap endonuclease’’ because
its preferred substrate is a primer-template in which the primer
has an unannealed 5?-region (14, 26). The nuclease enters the
5?-end of the unannealed tail, or flap, and slides to the point
of annealing. There it performs a structure-specific endonu-
cleolytic cleavage. The exonucleolytic activity is much less
active and has been proposed to be a manifestation of endo-
nucleolytic cleavage of the transiently unannealed 5?-
nucleotide of a primer (27). Furthermore, yeast FEN1 nucle-
ase functionally interacts with the Dna2 helicase (28, 29). This
supports the interpretation that the nuclease acts in vivo after
formation of a partially unannealed 5?-strand. It has puzzled us
that RNase HI appears to act on fully annealed Okazaki
fragments,whereastheFEN1nucleaseprefersasubstratewith
an unannealed tail.
The demonstration that the calf RNase HI is also a single-
strand specific junction RNase suggests an alternative pathway
for the removal of initiator RNA, as illustrated in Fig. 5. In this
case, strand displacement by DNA synthesis or DNA helicase
activity creates an unannealed tail at the 5?-end of every
Okazaki fragment. This is an efficient substrate for the junc-
tion RNase activity of RNase HI. Next, cleavage by FEN1
removes a segment of unannealed tail containing the 5?
monoribonucleotide. This fully endonucleolytic mechanism
not only takes advantage of preferred substrates for both
enzymes but could explain why pol ? does not display high
fidelity synthesis. Base substitution errors made by pol ? would
disappear from the chromosome if all of the nucleotides added
by that polymerase were removed by the strand displacement
mechanism. pol ? directed synthesis from upstream primers
would replace this region with accurately base-paired nucle-
otides. Unlike an exonucleolytic mechanism, cleavage of the
unannealed tail to remove RNA and a substantial region of
DNA requires that each nuclease make only a single cleavage.
When we created an Okazaki fragment that had a mis-
matched5?-regionincludingtheinitiatorRNAandanadjacent
region of DNA, FEN1 nuclease efficiently removed the entire
unannealed region (19). If so, what is the need of RNase HI?
We propose that RNase HI will cleave all initiator RNAs
regardless of secondary structure. FEN1 is inhibited by a
primer annealed to the tail flap (27, 30). Therefore, if the
initiator RNA forms a stable secondary structure with another
sequence on the displaced region of the Okazaki fragment, the
RNA may not be removable except by RNase HI.
In light of our current work, we can interpret results
obtained by Eder et al.(16) in more general terms. They found
that human RNase HI can cleave just 5? to a single ribonu-
cleotide in double-stranded DNA, even if that ribonucleotide
were mismatched. It is now evident that RNase HI recognizes
the RNA-DNA junction itself and not a unique helical struc-
ture. This allows the single ribonucleotide to be recognized
even though it is likely to have minimal impact on the
surrounding helical configuration. Furthermore, the ability to
FIG. 4.
HI. (A) An aliquot of purified RNase HI was subjected to sedimen-
tation on a 15–35% glycerol gradient. After fractionation, aliquots
were analyzed by electrophoresis on an 8–14% SDS-polyacrylamide
gel and stained with silver. Gel bands corresponding to RNase HI are
indicated (arrows). The peak of RNase HI is fraction 26 that corre-
sponds to a sedimentation value of ?4.3 and a native molecular mass
of ?68 kDa as expected for RNase HI. (B) Glycerol gradient fractions
(diluted 1:400) were tested for activity on [3H]RNA?DNA hybrids and
Okazakisubstratesaspreviouslydescribed.Thepercentofacidsoluble
radioactivity (?) and the percent of double-stranded (F) and single-
stranded (E) Okazaki substrate to converted to product were plotted
vs. fraction number. Above the graph, positions of protein standards
for the glycerol gradient are: (1) RNase (13.7 kDa), (2) BSA (68 kDa),
and (3) catalase (232 kDa).
Junction RNase activity is tightly associated with RNase
2248Biochemistry: Murante et al.Proc. Natl. Acad. Sci. USA 95 (1998)