Báo cáo khoa học: Factors influencing RNA degradation by Thermus thermophilus polynucleotide phosphorylase ppt

Chetverin, Institute of Protein Research of the Russian Academy of Sciences, Pushchino, Moscow Region, 142290 Russia Fax: +7 495 632 7871 Tel: +7 496 773 2524 E-mail: alexch@vega.protres

Thermus thermophilus polynucleotide phosphorylase

Marina V Falaleeva, Helena V Chetverina, Victor I Ugarov, Elena A Uzlova and

Alexander B Chetverin

Institute of Protein Research of the Russian Academy of Sciences, Pushchino, Moscow Region, Russia

Polynucleotide phosphorylase (PNPase,

polyribonucleo-tide:orthophosphate nucleotidyltransferase, EC 2.7.7.8)

is a 3¢ fi 5¢ exoribonuclease that catalyses the

phospho-rolysis of oligo- and polyribonucleotides producing

ribonucleoside 5¢-diphosphates (ppN), and this can be

reversed by decreasing the inorganic orthophosphate

(Pi) concentration and increasing the ppN

concen-tration [1]:

ðpNÞnþPi$ ðpNÞn1þppN

PNPase has been found in all types of bacteria [2]

except mycoplasma [3] Although PNPase-encoding

sequences are absent from all archaea and yeast

genomes examined to date (which do however contain

homologues of the related RNase PH gene), they are

present in higher eukaryotes [4–6] Plant and animal

PNPases are encoded by nuclear genes, but are mainly localized in organelles [7] In plants, there are two PNPase species, one of which is targeted to chlorop-lasts and the other to mitochondria [4,8]

The PNPase gene (pnp) encodes a 80–100 kDa poly-peptide composed of five evolutionary conserved domains: two N-terminal core domains homologous

to Escherichia coli RNase PH interspaced with an a-helical domain, and two C-terminal RNA-binding domains, KH and S1 [3–6,9,10] RNase PH is a 3¢ fi 5¢ phosphorolytic exonuclease that is responsible for processing of the 3¢ ends of precursor tRNA molecules [11]: the KH (‘K homology’) domain was originally identified in the human RNA-binding

K protein [12], and the S1 domain is named after the RNA-binding protein S1 of the E coli ribosome [13]

Keywords

3¢ fi 5¢ exoribonuclease; oligonucleotide

protection; RNA phosphorolysis; 3¢-terminal

modifications; Thermus thermophilus

Correspondence

A B Chetverin, Institute of Protein

Research of the Russian Academy of

Sciences, Pushchino, Moscow Region,

142290 Russia

Fax: +7 495 632 7871

Tel: +7 496 773 2524

E-mail: alexch@vega.protres.ru

(Received 30 November 2007, revised 29

February 2008, accepted 4 March 2008)

doi:10.1111/j.1742-4658.2008.06374.x

At the optimal temperature (65C), Thermus thermophilus polynucleotide phosphorylase (Tth PNPase), produced in Escherichia coli cells and isolated

to functional homogeneity, completely destroys RNAs that possess even a very stable intramolecular secondary structure, but leaves intact RNAs whose 3¢ end is protected by chemical modification or by hybridization with a complementary oligonucleotide This allows individual RNAs to be isolated from heterogeneous populations by degrading unprotected species

If oligonucleotide is hybridized to an internal RNA segment, the Tth PNPase stalls eight nucleotides downstream of that segment This allows any arbitrary 5¢-terminal fragment of RNA to be prepared with a precision similar to that of run-off transcription, but without the need for a res-triction site In contrast to the high Mg2+ requirements of mesophilic PNPases, Tth PNPase retains significant activity when the free Mg2+ con-centration is in the micromolar range This allows minimization of the

Mg2+-catalysed nonenzymatic hydrolysis of RNA when phosphorolysis is performed at a high temperature This capability of Tth PNPase for fully controlled RNA phosphorolysis could be utilized in a variety of research and practical applications

Abbreviations

Pi, inorganic orthophosphate; PNPase, polynucleotide phosphorylase; Tth PNPase, PNPase of Thermus thermophilus.

The PNPase molecules of E coli and Streptomyces

an-tibioticusare composed of three identical subunits that

form a doughnut-shaped structure with a central hole

capable of accommodating an RNA strand [10,14],

while the spinach chloroplast PNPase seems to be

composed of two such ring [4,15] Interestingly, the

three-dimensional structures of the bacterial PNPase

[10] and of the archaeal exosome core of RNase PH

[16] are highly homologous, suggesting that the

molec-ular mechanisms of action of PNPase and RNase PH

may be very similar [16,17]

As the intracellular Pi concentration is usually high,

the primary physiological role of PNPase is thought to

be the phosphorolytic degradation of RNA [2,18]

Important manifestations of this function include the

control of RNA quality by eliminating defective

mole-cules [19–21], regulation of the virulence and

persis-tence of pathogenic bacteria [22], anticancer activity

due to degradation of tumour-associated mRNAs

[23,24], and RNA processing [25] However, PNPase

sometimes employs its polymerizing ability in vivo to

extend the 3¢ ends of RNAs instead of using a poly(A)

polymerase [26,27]

For a long time, due to its synthetic activity in

the absence of template, PNPase was widely used

as a tool for producing a variety of model nucleic

acids to solve important biological problems, such

as establishing the genetic code and studies on

physicochemical properties of polyribonucleotides

[28] These applications have been largely replaced

by more versatile chemical synthesis of

oligori-bonucleotides with defined sequences and in vitro

transcription of DNA templates for the synthesis of

longer RNAs

In contrast to nucleotide polymerization, the

exonu-clease activity of PNPase was rarely exploited This is

mainly due to the fact that, despite the high

processivi-ty of the enzyme, phosphorolysis is not easily

control-lable The rate and extent of degradation depend on

the stability of the RNA structure and vary greatly

among RNA species [2] Increasing temperature results

in destabilization of the RNA structure and more

thor-ough and uniform degradation [29] However, E coli

PNPase is unstable above 55C and is rapidly

inacti-vated at 65C [2] Conditions have not been

estab-lished that would allow the desired RNA or the

desired fragment of RNA to selectively survive the

phosphorolysis reaction The only successful example

of a controlled phosphorolysis by PNPase has been the

faithful removal of poly(A) tails from mRNAs This

was achieved by carrying out the reaction at 0C,

when the rest of mRNA was structured and hence

resistant to PNPase [30]

In this regard, PNPases from thermophilic organ-isms deserve special attention as they are expected to

be active at high temperatures at which the second-ary and tertisecond-ary structures of RNA are melted There have been a few reports on such PNPases In one of them, PNPases from thermophilic bacteria Bacillus stearothermophilus and Thermus aquaticus showed highest polymerizing activities at 69 and

80C, respectively However, their ability to phosp-horolyse structured RNAs was not characterized, the enzyme from T aquaticus was not isolated, and the purified enzyme from B stearothermophilus was reported to be composed of atypically small subunits

of 51 kDa [31] In another paper, PNPase from Thermus thermophilus (Tth PNPase), a close relative

of T aquaticus, was purified to apparent homogene-ity as tested under non-denaturing conditions, but the purified enzyme completely lacked phosphorolytic activity, presumably due to proteolytic degradation that had occurred during the isolation procedure The purified enzyme was reported to be composed

of three unequal polypeptides of 92, 73 and 35 kDa [32]

In 1997, the nucleotide sequence of a putative pnp gene from T thermophilus was submitted to the Gen-Bank database (accession number Z84207) This open reading frame has the capacity to encode a 78.2 kDa polypeptide (a molecular mass that is similar to that of the E coli PNPase, 77.1 kDa [33]) in which all struc-tural domains characteristic of typical PNPases can be identified (NCBI Protein Database accession number CAB06341) However, details of the expression, isola-tion and characterizaisola-tion of the enzyme, although noted in the NCBI Protein Database submission, have not been published

In this paper, we report cloning of the pnp gene of

T thermophilus, its expression in E coli cells, and characterization of biochemical properties of the iso-lated enzyme We show that the isoiso-lated Tth PNPase

is capable of phosphorolysis, and that even RNAs with very stable secondary structures are readily degraded

to completion We further show that RNA can be pro-tected from phosphorolysis by either modifying its 3¢ end or annealing its 3¢-terminal sequence to a comple-mentary oligonucleotide If a mixture of protected and unprotected RNAs is treated with Tth PNPase, then only unprotected RNA is degraded Finally, we show that hybridization of oligonucleotide to an internal segment of RNA protects that segment, the upstream portion and a 8 nt downstream sequence of the RNA from Tth PNPase These features could make Tth PNPase a useful tool for controlled RNA degradation

in vitro

Isolation of Tth PNPase from E coli cells

The pnp gene was PCR-amplified using the T

thermo-philuschromosomal DNA as a template, cloned within

a plasmid downstream of the T7 promoter, and

expressed in T7 RNA polymerase-producing E coli

cells The expression product has a higher molecular

mass (78 203 Da, as calculated from the amino acid

sequence deposited in the NCBI Protein Database

under accession number CAB06341) than its E coli

counterpart (77 122 Da [33]), and this allowed removal

of the E coli enzyme from the Tth PNPase

prepara-tion to be monitored by PAGE analysis (see

supple-mentary Fig S1) The Tth PNPase isolation procedure

included heating of the cell lysate at 70C, resulting in

denaturation and precipitation of most of the host

proteins, including the E coli PNPase; ion-exchange

chromatography on a DEAE-Sepharose column;

incu-bation in the presence of Pi and Mg2+ (‘autolysis’,

during which any endogenous RNA was completely

degraded by PNPase; phosphatase was included to

eliminate any possible 3¢ phosphoryl groups on RNA

that might interfere with the phosphorolysis [34]);

and gel filtration through a highly porous Superose 6

column

The full-sized 78 kDa Tth PNPase polypeptide

con-stituted > 50% of the final preparation, with the rest

of protein comprising multiple shorter bands grouping

at around 60–70 and 30–40 kDa These minor

poly-peptides co-purified with PNPase, and similar products

accumulated in the E coli cells upon the induction of

T7 RNA polymerase (supplementary Fig S1)

There-fore, they are most likely fragments of the 78 kDa

polypeptide, which is highly susceptible to proteolysis

[32] As the enzyme preparation did not contain

con-taminating activities (as shown below), we did not

attempt to further purify it The enzyme yield (based

on the content of the 78 kDa polypeptide) was 3.3 mg

from 9 g of the Tth PNPase-producing E coli cells

The ADP-polymerizing activity of the final preparation

was approximately 200 unitsÆmg)1, which is higher

than reported for the PNPase isolated from T

thermo-philus cells (approximately 70 unitsÆmg)1 [32]), and,

unlike that preparation, our Tth PNPase retained

phosphorolytic activity (see below)

Conditions optimal for exonuclease activity

The isolated Tth PNPase degraded unprotected RNA,

but did not degrade RNA whose 3¢ end was hybridized

to a complementary oligodeoxyribonucleotide (Fig 1)

This observation indicated that the enzyme preparation does indeed possess the 3¢ fi 5¢ exonuclease activity and was free from endoribonucleases Also, RNA remained intact in the absence of Pi, suggesting that all degradation was due to phosphorolysis, rather than hydrolysis of polyribonucleotides

Although RNA was not degraded in the presence of

56 lm EDTA, which would sequester trace amounts of divalent cations, complete degradation of unprotected RNA occurred when the concentration of added

Mg2+was equal to that of EDTA (Fig 1) Given the value of the dissociation constant for the Mg2+:EDTA complex (approximately 10)9m [35]), this indicates that the requirement of Tth PNPase for free Mg2+ is very low (£ 1 lm), which is much lower than is the Mg2+ requirement of a mesophilic (Micrococcus luteus) enzyme, for which the apparent Km value for

Mg2+ is 0.8 mm [36] In subsequent experiments described here, the free Mg2+concentration was main-tained at about 40 lm, which allowed enzymatic phos-phorolysis of RNA to be performed at a high temperature without significant concomitant Mg2+ -catalysed hydrolysis Under such conditions, the opti-mal temperature for phosphorolysis was around 65 C, whereas, in the absence of PNPase, RNA remained substantially intact even at 85C (Fig 2)

Figure 3(A) shows that, at a saturating enzyme con-centration, when all RNA strands are phosphorolysed synchronously, a 109 nt 3¢-terminal fragment of a highly structured RQ135 RNA [37,38] is degraded to near completion within 2 min, suggesting a rate of approximately 50 nt per min at 65C This is much faster than the rate of 3.5 nt per min that is achievable

Fig 1 Mg2+ dependence of RNA phosphorolysis Nondenaturing PAGE patterns for the 3¢ fragment of RQ135 RNA treated with Tth PNPase after it had been annealed in the absence or presence of the oligodeoxyribonucleotide R-38 that is complementary to the RNA 3¢ end Each sample contained 0.056 m M EDTA, which was present in the RNA and enzyme preparations, and the indicated concentration of added MgCl2.

with the E coli PNPase at its optimal temperature

(37C) using a similarly structured tRNA-like

3¢-ter-minal domain of tobacco mosaic virus RNA [2]

How-ever, in contrast to the E coli enzyme, which saturates

RNA at a 1 : 1 m ratio, a larger amount of Tth

PNPase is required for saturation For the

concentra-tion of RNA used in the experiments shown in Fig 3

(50 nm), saturation occurs at a fourfold higher enzyme

concentration; the Tth PNPase concentration was

cal-culated assuming that one enzyme molecule active in

phosphorolysis consists of three intact (78 kDa)

subun-its [34] The same saturating enzyme-to-RNA ratio

was determined when complex formation in the

pres-ence of Mg2+ (but in the absence of Pi to prevent

phosphorolysis) was monitored by a shift of the

32P-labelled RNA band during PAGE (Fig 3B) Based

on the data shown in Fig 3, a Kd value of

approxi-mately 100 nm was estimated for the Tth

PNPa-se:RNA interaction Hence, the affinity of Tth PNPase

for RNA is approximately 10 times lower than that of the E coli enzyme (Kd= 10–20 nm [39]) The lower affinity does not necessarily mean that larger amounts

of Tth PNPase have to be used Instead, the complete degradation of RNA can be achieved by either reduc-ing the reaction volume to concentrate the reactants,

or increasing the incubation time

Selective protection of RNA by terminally hybridized oligonucleotides

As determined in an RT-PCR assay, under the optimal conditions established here (temperature, Mg2+ and EDTA concentrations, enzyme-to-RNA ratio, reaction time), Tth PNPase eliminated more than 99% of a highly structured unprotected RNA without affecting RNA whose 3¢-terminal sequence was protected by annealing with a complementary oligonucleotide (Fig 4A) This indicated that hybridization of an oligonucleotide at the 3¢ end of an RNA could selec-tively protect it from degradation by Tth PNPase Figure 4B demonstrates that this is indeed the case In this experiment, a mixture of two RNA species, 5¢ and 3¢ fragments of RQ135 RNA [37,38], was annealed with an oligodeoxyribonucleotide complementary to the 3¢ end of the 3¢ fragment Subsequent phosphoroly-sis with Tth PNPase resulted in complete degradation

of the 5¢ fragment and yielded a virtually pure 3¢ frag-ment (except for the oligodeoxyribonucleotide itself, which can be removed by electrophoresis or DNase treatment)

Fig 2 Temperature dependence of RNA phosphorolysis

Denatur-ing PAGE patterns of 0.5 pmol of the 3¢ fragment of RQ135 RNA

incubated at the indicated temperature with or without 0.5 pmol of

Tth PNPase.

Fig 3 Interaction of Tth PNPase with RNA (A) Denaturing PAGE patterns of 0.5 pmol of the 3¢ fragment of RQ135 RNA treated with the indicated relative amounts of Tth PNPase for the indicated time periods (B) Gel-shift analysis by nondenaturing PAGE of the complex formed between Tth PNPase and the [a- 32 P]-labelled 3¢ fragment of RQ135 RNA upon incubation for 15 min at 65 C in a phosphate-free phosphorolysis buffer RNA bands were visualized by scanning the gel using the Cyclon TM storage phosphor system (Packard Instruments, Meriden, CT, USA) The concentration of RNA in the incubation mixtures was constant (50 n M ), and the concentration of PNPase varied as indicated.

Protection of RNA by 3¢-terminal modifications

Figure 5 shows that 3¢-terminal modifications of RNA

make it more resistant to phosphorolysis by Tth

PNPase, further confirming that RNA degradation is

due to the 3¢ fi 5¢ exonuclease activity Oxidation of

RNA with periodate, which opens the terminal ribose

ring and generates a dialdehyde [40], results in selective

protection of the modified RNA species (Fig 5A,B)

This protection is moderate, but it is increased upon

treatment of the oxidized RNA with aniline (Fig 5C),

which eliminates the broken nucleoside and yields

RNA whose terminal 3¢ hydroxyl is phosphorylated [41] The high resistance of the 3¢-phosphorylated RNA to Tth PNPase is not due to chemical modifica-tion of the RNA body during periodate or aniline treatment, because it is fully released upon removal of the 3¢-terminal phosphate by a phosphatase (Fig 5C)

A similar increase in protection is achieved by treat-ment of the oxidized RNA with biotin hydrazide, which adds the biotin group to the RNA 3¢ end (Fig 5D)

Comparison with PNPase of E coli Figure 6 compares some properties of Tth PNPase with those of the E coli PNPase While both the enzymes readily degraded poly(A), an unstructured polyribonucleotide, they behaved differently with the 3¢-terminal fragment of RQ135 RNA, which possesses

a strong secondary structure [37,38] Unlike the

T thermophilus enzyme, the E coli PNPase only phos-phorolysed a fraction of the RNA at its optimal tem-perature, and the rest of the RNA remained apparently intact This resembles the degradation pat-tern of another structured RNA, tRNA [29], which exists in two conformations, of which only one is sus-ceptible to E coli PNPase attack at 37 C [42] How-ever, in contrast to tRNA [29,42], E coli PNPase was unable to completely degrade the 3¢-terminal fragment

of RQ135 RNA even when temperature was raised to

55C (supplementary Fig S2) Another difference between these two enzymes concerns the effects of the 3¢-terminal phosphoryl group While this group effi-ciently protected RNA from Tth PNPase, such an effect was not observed with the E coli PNPase, prob-ably because it was masked by the inability of the latter enzyme to completely degrade the unprotected RNA (Fig 6)

Protection of RNA by internally hybridized oligonucleotides

Figure 7 shows degradation patterns of three nested RNAs that were differently extended at the 3¢ end All three RNAs were subjected to phosphorolysis by Tth PNPase after annealing in the presence or the absence

of a 38 nt oligodeoxyribonucleotide that hybridized to the 3¢-terminal sequence of the shorter RNA (SmaI) and to internal sequences of the longer RNAs (EcoRI and PvuII)

The oligonucleotide completely protected the shorter RNA, and partially protected the two longer RNAs, whose degradation yielded a resistant product that migrated during PAGE slightly more slowly than did

A

B

Fig 4 Selectivity and completeness of RNA degradation (A)

Ethidium bromide-stained nondenaturing PAGE patterns for the

products of RT-PCR of RNA that survived incubation with Tth

PNPase of the indicated number of molecules of CT1n1 RNA

after it had been annealed in the absence or presence of the

oli-godeoxyribonucleotide R-38 that is complementary to the RNA

3¢ end PNPase was present in all samples except those marked

‘–PNPase’ (B) Nondenaturing PAGE patterns of a mixture of the

3¢ and 5¢ fragments of RQ135 RNA incubated with or without

Tth PNPase after annealing in the presence of the

oligodeoxyri-bonucleotide R-38 that is complementary to the 3¢ end of the 3¢

fragment.

C

D

B

Fig 5 Effects of 3¢-terminal modifications.

(A–C) Denaturing PAGE patterns for a

mix-ture of 0.5 pmol each of the 3¢ fragment

(upper band) and 5¢ fragment (lower band)

of RQ135 RNA, of which either the 5¢

ment (A and C, initial mixture) or the 3¢

frag-ment (B) was oxidized with sodium

periodate (3¢oxi) while the other possessed

the native 3¢ end (3¢OH) The mixture was

either not treated (N ⁄ T) or treated with the

indicated amounts (in pmol) of Tth PNPase,

either directly (A–C, initial mixture) or after

treatment with aniline, resulting in a

conver-sion of the oxidized fragment to

3¢-phos-phorylated (C, after aniline) and then with

shrimp alkaline phosphatase, removing the

3¢-phosphate (C, after phosphatase) (D)

Nondenaturing PAGE patterns for 3¢

frag-ments whose 3¢ end was either modified

with biotin (3¢bio) or not modified (3¢OH),

and which were either treated (+) or not

treated ( )) with Tth PNPase, or mixed with

a fivefold molar excess of streptavidin

before electrophoresis (Str).

Fig 6 Comparison of Thermus thermophilus and Escherichia coli

PNPases Nondenaturing PAGE patterns for 20 ng of poly(A) and

0.5 pmol of the 3¢ fragment of RQ135 RNA whose 3¢ end was

either intact (3¢OH) or esterified with phosphate (3¢P) as a result of

consecutive reactions with periodate and aniline, and which was

either not treated ( )) or treated in the presence of 1 m M Piand 0.1

m M Mg2+with 4 pmol of PNPase from E coli (E) at 37 C or with

1 pmol of Tth PNPase (T) at 65 C.

Fig 7 Effects of internally hybridized oligonucleotide Nondena-turing PAGE patterns for nested transcripts of the same plasmid digested at consecutive sites with SmaI (transcript length

109 nt), EcoRI (127 nt) or PvuII (303 nt), and treated (+) or not treated ( )) with Tth PNPase after annealing in the absence or presence of oligodeoxyribonucleotide R-38 complementary to positions 72–109 White arrows indicate the resistant phosphorol-ysis products.

the hybrid of the oligonucleotide with the shorter

RNA (left part of Fig 7) Such products were not

observed in the absence of the oligonucleotide (right

part of Fig 7)

The fine structure of the phosphorolysis product was

analysed by high-resolution (sequencing) PAGE under

denaturing conditions (Fig 8) Before phosphorolysis,

each of the three RNAs was dephosphorylated and

then labelled at the 5¢ end using [c-32P]ATP and

poly-nucleotide kinase Figure 8A shows that the resistant

product had rather heterogeneous 3¢ ends and was up

to eight nucleotides longer than the terminally

pro-tected RNA, whose 3¢ ends were also not uniform, but

this is common for run-off transcription [43] The

het-erogeneity of the phosphorolysis product was mainly

due to the heterogeneity of the protecting

oligonucleo-tide, and was substantially reduced when the 5¢ ends of

this oligonucleotide were made uniform by digesting a

longer oligonucleotide with restriction endonuclease

Phosphorolysis of the longer RNAs, hybridized with

this 5¢-terminally polished oligonucleotide, yielded a

product whose 3¢-terminal heterogeneity was similar to

that of a product of the run-off transcription, and which was 8 nt longer than the terminally protected shorter RNA (Fig 8B) This indicates that Tth PNPase stalls on a phosphorolysed RNA eight resi-dues from the 5¢ end of the internally hybridized oligo-nucleotide, which parallels the observation that PNPase of E coli stalls six to nine residues from the base of a stable stem-loop structure in an RNA strand [44] Similarly, the exosome core of a

hyperthermophil-ic archaeon Sulfolobus solfataricus, consisting of RNase PH subunits, also stalled eight or nine nucleo-tides from an upstream stem, which has been attrib-uted to the inability of a double-stranded structure to enter the central hole of the doughnut-shaped enzyme structure [45] The presence of a phosphate group at the 5¢ end of the oligonucleotide did not affect the length of the protected RNA fragment, but slightly reduced its heterogeneity (Fig 8B; compare lanes 3 and 4 with lanes 7 and 8)

Discussion

This paper reports the isolation of a thermophilic PNPase whose subunits have a molecular mass typical

of PNPases and which possesses phosphorolytic activ-ity The isolated enzyme is free from other

ribonucleas-es Biochemical characterization of the isolated enzyme revealed several important features that were not observed with PNPases from other sources

First, Tth PNPase has maximal phosphorolytic activity at a temperature at which even the most stable hairpins are melted Therefore, it degrades to completion those RNAs that mesophilic PNPases fail

to phosphorolyse Second, the requirement of Tth PNPase for free Mg2+ ions is extremely low This feature is beneficial, because it allows directed 3¢ fi 5¢ phosphorolysis to be performed at high tem-peratures without a risk of breakage of RNA strands due to Mg2+-catalysed hydrolysis Moreover, it is this feature that permitted us to discover that RNA can survive the high-temperature phosphorolysis con-ditions if its 3¢ end is protected Protection can be achieved either by hybridizing the terminal sequence with a complementary oligonucleotide (care must be taken to ensure that the hybrid is not melted at the reaction temperature) or by modifying the terminal nucleoside

The results show that 3¢-terminal modifications pro-tect RNA from Tth PNPase Previously, a phosphoryl group in either the 3¢ or 2¢ position of the terminal ribose, as well as the 2¢,3¢-cyclic phosphate, were found

to prevent phosphorolysis of short (up to 6 nt) oligo-nucleotides by PNPases from E coli [46] and M luteus

Fig 8 RNA sequence protected by internally hybridized

oligonu-cleotide High-resolution denaturing PAGE patterns for [5¢- 32

P]-labelled SmaI, EcoRI and PvuII transcripts protected from Tth

PNPase by hybridization with oligodeoxyribonucleotide R-38 (cf.

Fig 7) (A) The oligonucleotide was prepared by chemical

synthe-sis without further treatment Lanes 1 and 2 show untreated

SmaI and EcoRI transcripts respectively; lanes 3 and 4 show

PNPase-treated EcoRI and PvuII transcripts, respectively The

untreated PvuII transcript migrated too slowly to be shown here

(cf Fig 7) (B) Oligonucleotides were prepared from a longer one

by digestion at the SmaI site to generate uniform 5¢ ends (lanes

2–4) and by further removing the 5¢-terminal phosphate (lanes

6–8) Shown are untreated SmaI transcripts (lane 1),

PNPase-treated SmaI (lanes 2 and 6), EcoRI (lanes 3 and 7) and PvuII

(lanes 4 and 8) transcripts, and the EcoRI transcript partially

degraded in mild alkali to produce a 1 nt ladder [51] (lane 5) RNA

bands were visualized by scanning the gel using the Cyclon TM

storage phosphor system.

[47] However, there were conflicting reports on

phos-phorolysis of the longer (up to 20 nt)

3¢-phosphory-lated fragments that constitute most commercial RNA

preparations, of which between 20% [48] and 100%

[49] was found to be degradable by E coli PNPase

This disagreement could be explained by a different

average fragment length in the RNA preparations or

by different content of contaminating enzymes (such

as a phosphatase that would remove the 3¢-terminal

phosphate from the RNA fragments) in the PNPase

preparations used

The results further show that Tth PNPase is

sensi-tive to 3¢-terminal modifications that cannot result in

positioning of a phosphate or a similar bulky

nega-tively charged group in the subsite of the PNPase

active site normally occupied by the terminal

nucleo-side, the mechanism that was hypothesized to be the

basis for the inhibitory action of the 3¢-terminal

phos-phate [50] The fact that periodate oxidation of RNA,

resulting a conversion of the 2¢,3¢-cis-glycol to

dialde-hyde, also has an inhibitory effect indicates that

inter-actions with the free 3¢ and ⁄ or 2¢ hydroxyls of the

terminal ribose may be important for correct binding

of substrate RNA to the active site of Tth PNPase, as

was found for the S solfataricus exosome [45]

Terminal protection allows desirable RNA species

to be isolated from heterogeneous populations by

degrading other (unprotected) strands This could

also be used for reducing artifactual sequence

recom-bination and increasing the specificity of RT-PCR

assays The observation that biotin at the 3¢ end of

an RNA strand efficiently protects it from

phospho-rolysis suggests an efficient way of obtaining a 100%

pure biotinylated RNA preparation Many other Tth

PNPase applications of this sort can undoubtedly be

conceived

Finally, the results of this study show that any

arbi-trary 5¢-terminal fragment of an RNA strand can be

obtained by a controlled phosphorolysis with a

preci-sion similar to that of run-off transcription, but

with-out the need for a restriction site This can be done by

protecting the RNA strand with an oligonucleotide

complementary to an internal RNA segment that lies

8 nt upstream of the desired 3¢ end

In summary, this paper demonstrates that the

prop-erties of Tth PNPase allow fully controlled RNA

deg-radation, which could be used in various research and

practical applications In view of the structural

similar-ities between bacterial PNPase [10] and the S

solfatari-cus exosome [16], one may expect that the archaeal

enzyme, which is also thermophilic, will possess

bio-chemical properties similar to those reported here for

Tth PNPase

Experimental procedures

Cloning and expression of the pnp gene of

T thermophilus The pnp gene was PCR-amplified using the chromosomal DNA of T thermophilus VK1 as a template and primers 5¢-GACGTCGACATATGGAAGGCACACCCAATG-3¢ [matching the start of the PNPase-coding sequence, positions 977–995 of GenBank accession number Z84207 (underlined) and introducing an NdeI site (bold)], and 5¢-TTCGAATTC TTACTTGCGCCGCCTGGG-3¢ [matching the end of the PNPase-coding sequence, positions 3101–3117 of GenBank accession number Z84207 (underlined) and introducing an EcoRI site (bold)] The PCR product was digested at the NdeI and EcoRI sites, and ligated into plasmid vector 7 [52] between these sites The resulting plasmid, pT7-PNP, was cloned [53] and expressed in E coli B834(DE3)⁄ pLysS cells [54] by inducing T7 RNA polymer-ase synthesis using 1 mm isopropyl b-d-thiogalactoside for

4 h at 37C

Isolation of Tth PNPase Nine grams of Tth PNPase-producing E coli cells pel-leted from 6 L of culture were suspended in 40 mL of ice-cold buffer A (20 mm Hepes⁄ NaOH pH 8.2, 100 mm NaCl, 0.1 mm Na-EDTA, 2 mm b-mercaptoethanol, 1 mm phenylmethanesulfonyl fluoride) and disrupted in a Gaulin Micron Lab 40 homogenizer (APV Thermotech GmbH, Artem, Germany) at 1400 bar The lysate was 10-fold diluted with buffer A and incubated for 60 min

at 70C with stirring After pelleting of the cell debris and precipitated proteins for 40 min at 25 900 g in rotor JA-14 of a J2-21 centrifuge (Beckman, Vienna, Austria), the lysate was applied to a 11 mL DEAE-Sepharose FF (Amersham Biosciences, Vienna, Austria) column equili-brated with buffer B (20 mm Tris⁄ HCl pH 8.0, 1 mm Na-EDTA, 0.1 m NaCl, 2 mm b-mercaptoethanol, 5%

w⁄ w glycerol) After washing the column with buffer B, the enzyme was eluted with a linear gradient of NaCl (0.1–0.5 m) in the same buffer Fractions with a high PNPase activity were pooled, supplemented with DNase (1 lgÆmL)1) and calf intestinal alkaline phosphatase (1 unitÆmL)1), and dialysed overnight against buffer B containing 3 mm MgCl2 and 5 mm Na-phosphate After

an additional incubation for 30 min at 37C, followed

by 60 min at 66C and then chilling on ice, the precipi-tated proteins were pelleted as above The supernatant was concentrated to a volume of 0.25 mL using a Centr-icon-30 filter (Amicon, Beverly, MA, USA), cleared by centrifugation in a microcentrifuge at 12 100 g, and sub-jected to gel filtration through a 24 mL Superose 6 HR column (Amersham Biosciences) in buffer B Fractions with the highest PNPase activity were pooled, dialysed

overnight against the storage buffer (50 mm Tris⁄ HCl

pH 8.0, 100 mm NaCl, 10% w⁄ w glycerol, 0.1% Nonidet

P40 (BDH Ltd, Poole, England), 2 mm dithiothreitol),

cleared by centrifugation, and stored at )70 C after the

addition of glycerol to 50% w⁄ w

Enzyme assay

Tth PNPase activity was determined colorimetrically as Pi

[55] released from ADP during the synthesis of poly(A)

at 70C in a mixture containing 50 mm Tris ⁄ HCl

pH 8.2, 20 mm ADP, 0.1 m NaC1, 1 mm MgCl2 and no

primer One unit of activity was defined as the amount

of enzyme that polymerized 1 lmol of ADP per hour

Protein was measured by the method described by Lowry

et al [56] Enzyme purity was analysed by SDS–PAGE

(supplementary Fig S1) The percentage of full-sized

PNPase polypeptide was estimated from Coomassie

G-250-stained gel images using program imagej (http://

rsb.info.nih.gov/ij/)

Isolation of E coli PNPase

Unless otherwise stated, all procedures were carried out at

4C

Homogenization

Two hundred grams of pelleted E coli B cells were

sus-pended in 200 mL of buffer C (50 mm Tris⁄ HCl pH 8.0,

1 mm Na-EDTA, 10 mm MgCl2, 0.2 m KCl, 5 mm

b-mer-captoethanol, 5% w⁄ w glycerol, 1 mm

phenylmethane-sulfonyl fluoride) and disrupted in homogenizer 15M-8TA

(APV Gaulin Inc., Everett, MA, USA) at 600 bar

Phase partitioning

After removal of cell debris (for 40 min at 30 100 g in

rotor JA-14 of the Beckman J2-21 centrifuge), the lysate

was subjected to the liquid polymer phase partitioning

procedure previously developed for the isolation of Qb

replicase, using a mixture of polyethylene glycol 6000

(Merck KGaA, Darmstadt, Germany) and Dextran

T500 (Amersham Biosciences), followed by extraction of

RNA-bound proteins from the dextran phase using NaCl

[57]

DEAE cellulose

After dialysis against buffer D (10 mm Tris⁄ HCl pH 7.5,

1 mm Na-EDTA, 5 mm MgCl2, 5 mm b-mercaptoethanol,

5% w⁄ w glycerol), until conductivity of the extract

dropped to that of buffer D100 (buffer D plus 100 mm

NaCl), it was cleared by centrifugation for 40 min at

30 100 g in rotor JA-14 and applied to a 200 mL DE-52 column (Whatman, Florhom Park, NJ, USA) equilibrated

in buffer D100 The column was washed with 2.5 vol-umes of D100, then with 2.5 volvol-umes of D150 (buffer D plus 150 mm NaCl) and eluted with a gradient of NaCl (150–400 mm) in buffer D Fractions with a high PNPase activity were pooled, protein was concentrated by precipi-tation with ammonium sulphate (40 gÆ100 mL)1), and the pellet was solubilized in 20 mL of buffer and dialysed overnight against buffer D100

Autolysis After clearing the dialysed preparation by centrifugation, it was supplemented with 50 lg of DNase, 50 units of calf intestinal alkaline phosphatase, 1 mm ZnCl2 and 1 mm phenylmethanesulfonyl fluoride, and incubated for 60 min

at 37C Then 1 m K-phosphate (pH 8.0) was added to a concentration of 100 mm, and incubation was continued for an additional 30 min

Sepharose CL-6B The autolysed preparation was cleared by centrifugation and subjected to gel filtration through a column of Sepha-rose CL-6D, Amersham Biosciences (2.6· 100 cm) equili-brated in buffer D100

Mono Q Fractions with high activity were pooled and chromato-graphed in several portions on a FPLC system through a Mono Q HR 5⁄ 5 column (Amersham Biosciences) equili-brated in buffer D100 and utilizing a 100–1000 mm NaCl gradient in buffer D

Superose 6 Active Mono Q fractions were pooled, concentrated by passing through a Centricon-30 filter (Amicon), and sub-jected to gel filtration through a 24 mL Superose 6 HR column (Amersham Biosciences) Fractions were analysed

by SDS–PAGE, and those containing the least amount of impurities were pooled, and stored at )70 C after addition

of glycerol to 50% w⁄ w

The final preparation (2.2 mg of protein) was > 90% pure (supplementary Fig S1C, lanes 2 and 3) and was free from endonuclease activity The specific activity of the enzyme was approximately 450 unitsÆmg)1, with 1 unit being defined as the amount of enzyme that released 1 lmol

of Pifrom ADP per hour at 37C in a mixture containing

100 mm Tris⁄ HCl pH 8.0, 10 mm ADP, 0.5 mm Na-EDTA,

10 mm MgCl2and no primer

RNA sequences

CT1n1 RNA is a derivative of RQ135)1 RNA [37]

carry-ing a 43 nt insert (underlined), generated in unpublished

experiments on intermolecular RNA recombination The

RNA preparation was obtained by transcription of a

SmaI-digested plasmid, encoding the recombinant sequence:

GGGGUUCCAACCGGAAGUUGAGGGAUGCCUAGG

CAUCCCCCGUGCGUCCCUUAAAGCUUCAUUCUUC

CUUUCUUUAAAAGAGAGAGAGAGAAAGCGAGGG

AUUUGAGAGAUGCCUAGGCAUCUCCCGCGCGCC

GGUUUCGGACCUCCAGUGCGUGUUACCGCACUG

UUAGCCC

The 5¢ fragment of RQ135 RNA is a 52 nt 5¢ terminal

sequence of RQ135)1 RNA [37] carrying a 23 nt extension

(underlined), obtained by transcription of the 5¢

fragment-encoding plasmid [58] digested at site BamHI: GGGG

UUCCAACCGGAAGUUGAGGGAUGCCUAGGCAUC

CCCCGUGCGUCCCUUCUGCAGCUCGAGUCUAGAG

GAUC

The 3¢ fragment of RQ135 RNA (transcript SmaI) is a

81 nt 3¢ terminal sequence of RQ135)1RNA [37] carrying a

28 nt extension (underlined), obtained by transcription of

the 3¢ fragment-encoding plasmid [58] digested at the SmaI

site: GGCGCUGCAGCUCGAGUCUAGAGGAUCCUA

CGAGGGAUUUGAGAGAUGCCUAGGCAUCUCCCG

CGCGCCGGUUUCGGACCUCCAGUGCGUGUUACC

GCACUGUUAGCCC

Transcript EcoRI (the 3¢ fragment of RQ135 RNA

extended at the 3¢ end by a 18 nt sequence, underlined) was

obtained by transcription of the 3¢ fragment-encoding

plas-mid [58] digested at site EcoRI: GGCGCUGCAGCUCGA

GUCUAGAGGAUCCUACGAGGGAUUUGAGAGAUG

CCUAGGCAUCUCCCGCGCGCCGGUUUCGGACCU

CCAGUGCGUGUUACCGCACUGUUAGCCCGGGUA

CCGAGCUCGAAUU

Transcript PvuII (the 3¢ fragment of RQ135 RNA

extended at the 3¢ end by a 194 nt sequence, underlined)

was obtained by transcription of the 3¢ fragment-encoding

plasmid [58] digested at site PvuII: GGCGCUGCAGCUC

GAGUCUAGAGGAUCCUACGAGGGAUUUGAGAGA

UGCCUAGGCAUCUCCCGCGCGCCGGUUUCGGAC

CUCCAGUGCGUGUUACCGCACUGUUAGCCCGGG

UACCGAGCUCGAAUUCGUAAUCAUGGUCAUAGC

UGUUUCCUGUGUGAAAUUGUUAUCCGCUCACAA

UUCCACACAACAUACGAGCCGGAAGCAUAAAGU

GUAAAGCCUGGGGUGCCUAAUGAGUGAGCUAA

CUCACAUUAAUUGCGUUGCGCUCACUGCCCGCUU

UCCAGUCGGGAAACCUGUCGUGCCAG

RNA preparations

RNA samples used in this work were prepared by run-off

transcription of plasmids digested at appropriate restriction

sites and purified by PAGE as described previously [58]

Where indicated, RNA was oxidized with sodium perio-date, followed by treatment with aniline and shrimp alka-line phosphatase as described previously [58] Biotinylated RNA was prepared by incubation of periodate-oxidized RNA at 4C for 20 h with 5 mm biotin hydrazide (Sigma,

St Louis, MO, USA) in the presence of 0.1% SDS and 0.2 m Na-acetate (pH 4.8) Unincorporated biotin hydra-zide was removed by gel filtration through a Sephadex G-25 (Amersham Biosciences) spin column

RNA:oligonucleotide hybridization Immediately before hybridization, RNA (in 0.1 mm Na-EDTA pH 8.0) and oligodeoxyribonucleotides (in

10 mm Tris⁄ HCl pH 9.0, 0.01 mm Na-EDTA) were sepa-rately melted in a boiling bath for 2 min followed by rapid chilling on ice Hybridization was carried out by incuba-tion, at 65C for 15 min, of a 5 lL mixture containing

80 mm Hepes-KOH pH 8.1, 150 mm KCl, 0.2 mm Na-EDTA, 0.5 pmol RNA and 2.5 pmol of complementary oligodeoxyribonucleotide

RNA degradation and analysis Unless otherwise stated, RNA degradation and analysis were performed using a 10 lL reaction mixture containing

50 mm Hepes-KOH pH 8.1, 75 mm KCl, 1 mm dithiothrei-tol, 1 mm Na-phosphate, 0.06 mm Na-EDTA, 0.1 mm MgCl2, 0.5 pmol RNA, 2 pmol Tth PNPase, and, where indicated, 2.5 pmol of a protective oligodeoxyribonucleo-tide Reactions were carried out at 65C for 15 min and terminated by the addition of 10 lL of 10 mm Na-EDTA Nucleic acids were isolated by phenol extraction [53] and analysed by nondenaturing PAGE (in the presence of TBE buffer: 100 mm Tris⁄ HCl, 100 mm boric acid, 2 mm Na-EDTA) or by denaturing PAGE (in the presence of buffer TBE containing 7 m urea) Gels were stained after electrophoresis using silver [59]

Preparation of oligodeoxyribonucleotides with uniform 5¢ ends

To prepare oligodeoxyribonucleotide R-38 with uniform 5¢ end, two other oligonucleotides were synthesized: a longer oligodeoxyribonucleotide R-50 (3¢-AAAGCCTGGAGGTC ACGCACAATGGCGTGACAATCGGGCCCCCTTAAG GT-5¢) and a shorter complementary oligodeoxyribonucleo-tide F-24 (5¢-CACTGTTAGCCCGGGGGAATTCCA-3¢) Upon annealing, these produced a partial duplex (last 24 nucleotides of R-50 with complete sequence of F-24), which was digested with restriction endonuclease SmaI at the site underlined in R-50 After digestion, the preparation was melted, and the 38 nt long oligodeoxyribonucleotide R-38 was isolated by PAGE