The Lytic Enzyme of Bacteriophage PRD1 Is Associated with the Viral Membrane

Bacteriophage PRD1 encodes two proteins (P7 and P15) that areassociated with a muralytic activity. Protein P15 is a solubleß-1,4-N-acetylmuramidase that causes phage-inducedhost cell lysis. We demonstrate here that P15 is also a structuralcomponent of the PRD1 virion and that it is connected to thephage membrane. Small viral membrane proteins P20 and P22 modulateincorporation of P15 into the virion and may connect it to thephage membrane. The principal muralytic protein involved inPRD1 DNA entry seems to be the putative lytic transglycosylaseprotein P7, as the absence of protein P15 did not delay initiationof phage DNA replication in the virus-host system used. Theincorporation of two different lytic enzymes into virions mayreflect the broad host range of bacteriophage PRD1.

INTRODUCTION

Top

Introduction

For successful initiation of infection, an incoming bacteriophagegenome has to cross the barrier formed by the host cell wall.In some cases phage particles are known to contain lytic enzymesthat are capable of producing local openings into the mureinsacculus. In the case of enveloped double-stranded RNA (dsRNA)phage ${phi}$ 6, the lytic endopeptidase, protein P5, is associatedwith the nucleocapsid surface (3, 11, 28, 38). In bacteriophageT4, the baseplate structure of the tail contains protein gp5,which has N-acetylmuramidase activity (29, 42). Amino acid sequenceanalyses have identified putative lytic transglycosylases amongthe structural proteins of bacteriophages T7, PRD1, and P1 (30,33). The putative transglycosylase of phage T7, protein gp16,has recently been shown to increase the efficiency of phagegenome translocation into the host cell (39).

Bacteriophage PRD1 is a lipid-containing dsDNA virus that infectsgram-negative cells carrying a conjugative P, N, or W incompatibilitygroup plasmid (44). The virion consists of an icosahedral proteincapsid surrounding a membrane vesicle that encloses the lineardsDNA genome, which is 14,925 bp long (1, 7, 16). In contrastto other dsDNA phages that have a proteinaceous tail associatedwith the portal vertex, PRD1 particles have spike-penton structuresconsisting of the receptor-binding protein P2, the trimericspike shaft protein P5, and the penton protein P31 at everyvertex (13, 21, 48). The internal phage membrane consists ofapproximately equal amounts of phospholipids and phage-encodedproteins and is capable of transforming from a spherical vesicleinto a tubular structure; it has been proposed that this tubularstructure delivers the phage genome into the host cell cytoplasm(1, 34).

The PRD1 genome encodes two proteins with muramidase activity,a 1,4-ß-N-acetylmuramidase (protein P15), which causeshost cell lysis and liberation of progeny phages (12, 35), andprotein P7, which carries a conserved transglycosylase motifin its amino-terminal half (7, 30). Protein P7 has been shownto be active against peptidoglycan and has an accessory rolein phage DNA entry (46). In this study we found that proteinP15 is a structural component of the virion and that it is associatedwith the phage membrane, most probably via protein-protein interactions.

MATERIALS AND METHODS

Top

Materials and Methods

Bacteria and phages. The bacterial strains, phages, and plasmids used in this studyare listed in Tables 1 and 2 and 3, respectively. Cells weregrown in Luria-Bertani medium (50), and when appropriate, chloramphenicol(25 µg/ml) and/or kanamycin (25 µg/ml) was added.For production of wild-type and mutant phage particles, Salmonellaenterica DS88 cells were infected at a multiplicity of infection(MOI) of 6. When Escherichia coli HMS174(pJB21)(pLM2) cellswere used as hosts, a MOI of 50 was employed. After lysis, viralparticles were concentrated and purified in 5 to 20% sucroserate zonal gradients, as previously described (5). DS88 andHMS174(pJB21)(pLM2) cells infected with phages carrying an ambermutation in the lytic enzyme (protein P15)-encoding gene XV(sus174, sus232, or sus240) were collected by centrifugation(Sorvall GS3 rotor, 3,600 rpm, 10 min, 4°C) 2 h postinfection,resuspended in 20 mM potassium phosphate buffer (pH 7.4) (using1/80th of the original volume), and disrupted by passage througha 0.375-in. French press (300 lb/in²). The original volume wasrestored with buffer, and the viruses were concentrated andpurified in rate zonal gradients as the wild-type virus was.Light-scattering zones formed by PRD1 particles were collectedand either harvested by differential centrifugation, which resultedin 1x-purified virus preparations, or purified further in 20to 70% sucrose equilibrium gradients and collected by centrifugation,which resulted in 2x-purified virus preparations (5, 6). The1x-purified virus preparations or (in the case of mutant particlesdevoid of protein P15) 2x-purified virus preparations were purifiedfurther by ion-exchange chromatography with Sartorius D100 filtercartridges, as described by Walin et al. (52).

View this table:
[in this window]
[in a new window]
TABLE 1. Bacterial strains used in this study

View this table:
[in this window]
[in a new window]
TABLE 2. Phages used in this study

View this table:
[in this window]
[in a new window]
TABLE 3. Plasmids used in this study

Isolation of the PRD1 membrane fraction. For isolation of PRD1-specific membrane vesicles devoid of theprotein capsid, DS88 cells were infected with sus636 mutantphage at an MOI of 10 (47). Vesicles were precipitated fromthe lysate with polyethylene glycol 6000 by using the standardphage purification method and were separated by isopycnic densityflotation centrifugation in sucrose gradients as previouslydescribed (4, 22, 47). The light-scattering PRD1-specific membranezone at a density of $~$ 1.15 g/cm³ was collected and used in enzymaticassays as described below.

Membranes were obtained from 1x-purified wild-type PRD1 particlesby treating the particles with 2 M guanidine hydrochloride (GuHCl)as described previously (1). Disrupted virus preparation (2mg of protein) was layered on top of a linear 10 to 40% sucrosegradient (in 20 mM potassium phosphate buffer [pH 7.4] supplementedwith 2 M GuHCl), and empty and DNA-filled membrane vesicleswere obtained by centrifugation (Sorvall TH641 rotor, 30,000rpm, 5 h, 20°C). Light-scattering bands at densities of1.14 and 1.17 g/cm³ corresponding to empty and DNA-filled membranes,respectively, were collected. In order to remove GuHCl fromthe samples, the samples were dialyzed against 2 liters of 20mM potassium phosphate buffer (pH 7.4) for 15 h at 4°C beforeenzymatic assays were performed.

Protein contents in mutant infections and phage particles. Production of receptor-binding protein P2 in P15^- mutant (sus174,sus232, or sus240)-infected DS88 cells was analyzed as follows.One-milliliter samples were taken 30, 50, and 120 min postinfection,and cells were harvested and analyzed by sodium dodecyl sulfate(SDS)-polyacrylamide gel electrophoresis (PAGE) and Westernblotting with polyclonal P2 antiserum (21). Similarly, the proteincompositions of purified P15^- (sus174, sus232, or sus240) mutantparticles were analyzed by Western blotting with polyclonalP2 antiserum (21).

Enzymatic assays. Peptidoglycan sacculi of E. coli DH5 ${alpha}$ and Micrococcus lysodeikticuswere prepared as described by Hoyle and Beveridge (27) by boilingcells derived from 500-ml stationary-phase cultures in 4% SDS.The sacculi were washed with 2 M NaCl as described by Bernadskyet al. (9) to remove peptidoglycan-associated proteins, resuspendedin 5 ml of Milli-Q water, and stored at -20°C until theywere used. Zymogram analysis was used for detecting proteinswith muralytic activity as previously described (9). SDS-polyacrylamide(14% polyacrylamide) gels (101 by 82 by 0.5 mm) were cast asdescribed by Olkkonen and Bamford (43), except that an E. colior M. lysodeikticus peptidoglycan preparation homogenized bysonication was added to a final concentration of 15% (vol/vol).Following electrophoresis, the gels were first rinsed and thensoaked for 30 min in Milli-Q water and for 30 min in renaturationbuffer (25 mM potassium phosphate buffer [pH 7.4], 0.1% TritonX-100) with gentle agitation at room temperature. The gels weretransferred to fresh renaturation buffer and incubated for 48to 72 h at 4°C. After incubation, zymograms were stainedwith 0.1% methylene blue in 0.01% KOH at 37°C for 1 h anddestained with Milli-Q water.

Chloroform treatment of E. coli DH5 ${alpha}$ cells and the turbidimetricassay were performed as described by Caldentey and Bamford (11).Chloroform-treated cells were resuspended in 50 mM Tris-HCl(pH 7.2)-0.1% Triton X-100 to an absorbance at 450 nm (A₄₅₀)of about 0.450. Samples containing $~$ 60 µg of purified wild-typeor mutant virus particles or the corresponding amount (majormembrane-associated protein P11 of PRD1 was used for normalization)of virus-specific membrane vesicles were added to the cell suspension(800 µl). Changes in turbidity (A₄₅₀) were monitored witha double-beam Cary 219 spectrophotometer by using cuvettes withan optical path length of 1 cm and a spectral bandwidth of 1nm. Hen egg white lysozyme, used as a control in enzymatic assays,was purchased from Roche.

Phage adsorption and DNA replication assays. A phage adsorption test was performed by incubating 300 to 600phage particles with 2 x 10⁸ Salmonella cells grown to the optimaladsorption phase (2 x 10⁹ CFU/ml) (31). The infection mixtureswere incubated for 1, 4, and 8 min at 20°C, and the cellswere collected by centrifugation (Heraeus Biofuge Pico centrifuge,5,000 rpm, 3 min, 20°C). The numbers of nonadsorbed phageparticles in the supernatant fractions were determined by platingon lawns of DB7156(pLM2) cells.

Replicating PRD1 DNA was labeled with [³H]thymidine by the methodof Mosig and Colowick (40), except that nalidixic acid (40 µg/ml)was added to inhibit host DNA replication (15). In the caseof sus232 mutant particles an MOI of 30 was employed, whilefor wild-type and sus471 mutant infections an MOI of 15 wasused. The amount of incorporated ³H was determined by scintillationcounting.

Analytical methods. SDS-PAGE was performed as described by Olkkonen and Bamford(43). Western blotting was performed by transferring the proteinsfrom the gels onto polyvinylidene difluoride membranes (Millipore),and the proteins were visualized with the ECL detection system(Amersham) by using horseradish peroxidase-conjugated swineanti-rabbit immunoglobulins (Dako) as secondary antibodies.Protein concentrations were determined with Coomassie brilliantblue by using bovine serum albumin as a standard (10).

RESULTS

Top

Results

Small integral membrane proteins P20 and P22 modulate protein P15 incorporation into the virion. The particle-associated lytic activities of bacteriophage PRD1were investigated by analyzing purified viruses in zymograms.Phage proteins separated by gel electrophoresis were renaturedby incubating the gels in renaturation buffer for 2 to 3 days.Renatured muralytic proteins degraded the peptidoglycan substrateadded to the gels and were detected as clear zones. Two suchzones appeared when wild-type PRD1 particles were analyzed (Fig.1a). The zone that moved more slowly was identified previouslyas protein P7, which has a conserved lytic transglycosylasedomain in its N-terminal half (30, 46). The faster-moving zone,which had a molecular mass of $~$ 15 kDa, was present in all ofthe other particles studied except those devoid of proteinsP15 (sus174, sus232, and sus240 mutants) and P20 (sus30 mutant)(Fig. 1a and b). In addition, the signal caused by the faster-movinglytic activity was clearly decreased in P22-deficient particles(sus42 mutant) when it was compared to the signal in wild-typevirions and normalized to the signal caused by protein P7. Themolecular mass of the faster-moving clear zone was found tocorrespond to that of protein P15 (17.3 kDa) (45), the solublelytic enzyme encoded by PRD1 (12, 35). Proteins P20 and P22are integral membrane proteins that have molecular masses of4.7 and 5.5 kDa, respectively, and it has been proposed thatthese proteins span the membrane once and have no recognizablelytic domains in their amino acid sequences (7, 36; J. K. H.Bamford, J. M. Grimes, D. I. Stuart, and D. H. Bamford, submittedfor publication). Thus, we concluded that the appearance ofthe faster-moving clear zone in zymograms was due to the enzymaticactivity of protein P15. The activity of PRD1 muramidases againstgram-positive peptidoglycan isolated from M. lysodeikticus wasalso analyzed (Fig. 1c). Protein P7 was found to be inactive,while protein P15 degraded peptidoglycan obtained from M. lysodeikticusand caused the appearance of a clear zone in zymograms.

View larger version (53K):
[in this window]
[in a new window]
FIG. 1. Zymogram analysis of lytic activities associated with wild-type and mutant PRD1 particles with different protein compositions. (a and b) Zymograms supplemented with gram-negative E. coli peptidoglycan. (c) Zymogram supplemented with gram-positive M. lysodeikticus peptidoglycan. The positions of PRD1 proteins exhibiting muralytic activities (P7 and P15) are indicated by arrows. wt, wild type.

Muralytic activity of protein P15 is localized in the phage membrane. In addition to the zymogram analysis, the particle-associatedlytic activities were probed by measuring the changes in A₄₅₀when purified phages were added to chloroform-treated E. colicells. Figure 2a shows the decreases in absorbance when differentamounts of the control hen egg white lysozyme were added tothe substrate cells. Adding freshly purified wild-type particlesdid not result in a sharp decrease in the A₄₅₀. However, theactivity of these particles could be significantly increasedby partially disrupting the protein capsid by three cycles offreezing and thawing (Fig. 2a). Centrifugation and SDS-PAGEanalysis demonstrated that this treatment led to dissociationof $~$ 10% of the major capsid protein P3 and protein P6 from thevirions, while all of the other proteins were still associatedwith the particles (data not shown). Turbidity measurementswere performed with activated wild-type and mutant phage particlesin order to analyze the muralytic activities of PRD1 particleswith different protein compositions (Fig. 2b). The absence ofprotein P7 did not have a detectable effect on the muramidaseactivity. No decrease in the A₄₅₀ was detected when P15^- (sus174,sus232, or sus240) or P20^- (sus30) mutant particles were addedto the assay mixture. Decreases in lytic activity were associatedwith mutants in which a membrane protein (P14 or P22) or a majormembrane-associated protein (P11) was missing. The DNA contentsof viral particles did not affect their muralytic propertiesas the empty P9^- (sus1) mutant particles were found to be asactive as wild-type virions.

View larger version (25K):
[in this window]
[in a new window]
FIG. 2. Lytic activities as determined by turbidity measurements. (a) Control assay in which different amounts of hen egg white lysozyme were added to a suspension of chloroform-treated substrate cells. For comparison, assay mixtures containing freshly purified wild-type particles (PRD1, 60 µg of protein) and particles activated by freezing and thawing three times (PRD1*, 60 µg of protein) were included. (b) Corresponding assay performed with wild-type and mutant PRD1 particles activated by freezing and thawing three times. In this assay 1x- or 2x-purified virus particles (60 µg of protein in approximately 15 µl) were added to substrate cells (in an 800-µl mixture). As a control, 15 µl of 20 mM potassium phosphate buffer (pH 7.4) was added to the assay mixture. wt, wild type.

In order to determine the location of protein P15 in the virionmore precisely, phage membranes were isolated. PRD1-specificmembranes devoid of DNA were obtained from sus636 mutant-infectedpreparations in which protein capsids were not assembled. Disruptionof wild-type viruses with GuHCl yielded both empty and DNA-filledmembrane vesicles. Zymogram analysis revealed the presence ofproteins P7 and P15 in both DNA-filled and empty phage membranes(Fig. 3a). When turbidity was measured, only the DNA-filledmembranes caused a decrease in A₄₅₀ when they were added tochloroform-treated cells (Fig. 3b). In this case the activitydetected was $~$ 50% of the activity associated with similar amountsof wild-type virions, most probably due to the GuHCl treatment.

View larger version (55K):
[in this window]
[in a new window]
FIG. 3. Zymogram analysis (a) and turbidity measurement (b) of lytic activities associated with isolated PRD1 membrane fractions. In both cases the major membrane-associated protein P11 was used to normalize the amounts of virus particles versus membranes used in the assays. P15, partially purified protein P15 (prepared as described by Caldentey et al. [12]); wt, wild-type PRD1; G, wild-type virions disrupted with 2 M GuHCl; V, empty phage-specific membranes isolated from sus636 mutant infections; E and D, empty and DNA-filled membranes obtained from wild-type particles disrupted with GuHCl, respectively.

Decreased amounts of receptor-binding protein P2 in P15^- mutant particles can be restored to wild-type levels by complementation in trans.. SDS-PAGE and Western blot analysis showed that in P15^- mutantparticles the amount of receptor-binding protein P2 was approximately10% of the amount found in wild-type virions (P2^R P15^-) (Fig.4). Also, the specific infectivities obtained for P2^R P15^- mutantparticles (6.5 x 10⁹, 1.1 x 10⁹, and 1.3 x 10⁹ PFU/µgof protein in 2x-purified virus preparations for sus174, sus232,and sus240 mutants, respectively) were approximately 10% ofthe specific infectivity of wild-type particles ( $~$ 1.0 x 10¹⁰PFU/µg of protein in a 2x-purified virus preparation)(6).

View larger version (28K):
[in this window]
[in a new window]
FIG. 4. SDS-PAGE analysis (a) and immunological detection (b) of protein P2 from ion-exchange chromatography-purified wild-type and P15^- mutant PRD1 particles. The positions of some phage proteins are indicated by arrows.

In order to determine whether the reduction in the amount ofprotein P2 in P15^- mutant particles was due to a decrease inthe amount of protein P2 in mutant-infected cells or to an inabilityto assemble protein P2 into P15^- mutant particles, we analyzedthe production of protein P2 during wild-type and mutant infectionsby SDS-PAGE and Western blotting with P2 antiserum. In wild-typeinfections, P2 was detected in samples taken 30 min postinfectionor later. Similarly, in P15^- mutant infections, protein P2 wasdetected at the same times, but the levels were significantlylower.

PRD1 particles containing wild-type amounts of protein P2 butdevoid of protein P15 (P2⁺ P15^-) could be obtained by providingprotein P2 in trans. sus232 mutant phages were propagated inE. coli HMS174(pJB21)(pLM2) carrying wild-type gene II clonedin the pSU18 vector. Incorporation of protein P2 into mutantparticles was verified by SDS-PAGE and by Western blotting (Fig.4). The specific infectivity obtained for P2⁺ P15^- particleswas 6.5 x 10⁹ PFU/µg of protein, which was $~$ 50% of thespecific infectivity of ion-exchange chromatography-purifiedwild-type viruses (1.8 x 10¹⁰ PFU/µg of protein) (52).

Viral DNA replication is not delayed in the absence of protein P15. The ability of sus232 P2⁺ P15^- mutant particles, as well asP2^R P15^- mutant particles, to bind to a host cell surface wasdetermined by an adsorption assay (Fig. 5a). This analysis showedthat P2^R P15^- mutant particles bind to the host cells approximately50% less efficiently than the wild-type viruses bind. However,the level of adsorption of P2⁺ P15^- mutant particles was thewild-type level. Phage DNA replication was studied by monitoringthe incorporation of [³H]thymidine into infected cells. sus471mutant particles devoid of the putative lytic transglycosylaseprotein P7 were used as a control because in sus471 mutant infectionsthe initiation of DNA replication is known to be delayed andasynchronous (46). In order to compensate for the reductionin adsorption, the MOI used in P2⁺ P15^- and P2^R P15^- mutantinfections was twice the MOI used in P7^- (sus471) mutant andwild-type phage infections. In P2^R P15^- mutant infections theabsence of protein P15 had no effect on the initiation or therate of phage DNA replication (Fig. 5b). In P2⁺ P15^- mutantinfections the level of viral DNA synthesis was substantiallydecreased. Accordingly, host cell lysis was delayed and incomplete,as shown by the one-step growth curves in Fig. 5c.

View larger version (19K):
[in this window]
[in a new window]
FIG. 5. (a) Adsorption of wild-type and mutant particles to DB7156(pLM2) cells. Symbols: ${circ}$ , wild-type PRD1; , P7^- mutant (sus471); ${blacktriangledown}$ , P2⁺ P15^- mutant (sus232, protein P2 provided in trans); ${triangledown}$ , P2^R P15^- mutant (sus232). (b) [³H]thymidine incorporation into replicating DNA in DB7156(pLM2) cells treated with nalidixic acid. Nalidixic acid inhibits host DNA replication by inactivating DNA gyrase. PRD1 DNA replication is not affected. ${blacksquare}$ , DB7156(pLM2) cells not infected with PRD1. Other symbols as described above for panel a. (c) One-step growth curves for wild-type and mutant PRD1 phages in a suppressor host. Symbols as described above for panel a.

DISCUSSION

Top

Discussion

The host cell lysis induced by bacteriophage PRD1 is causedby the muralytic activity of protein P15 (35). Previously, recombinantprotein P15 was purified and characterized, and it was foundto be a soluble 1,4-ß-N-acetylmuramidase with stronglytic activity against gram-negative peptidoglycan (12). Inthis study we found that it is also a structural component ofthe PRD1 virion.

Several lines of evidence suggest that protein P15 is associatedwith the internal phage membrane. Freezing and thawing purifiedphage particles three times led to partial dissociation of theprotein capsid surrounding the membrane. Simultaneously, anapproximately fivefold increase in particle-associated lyticactivity was detected by turbidity measurements. When isolatedPRD1 membranes were studied, the presence of protein P15 inempty and DNA-filled phage membranes could be detected by zymogramanalysis. In turbidity assays only membranes containing thephage genome were associated with muralytic activity. Basedon these results, we propose that the proper (spherical) conformationof the membrane vesicle is critical for positioning proteinP15 correctly in order to bind and cleave the peptidoglycansubstrate. The proper conformation can be maintained by supportingthe membrane vesicle either internally (with the phage genomein DNA-filled membranes) or externally (with the protein capsidin empty sus1 mutant particles). The release of DNA from PRD1particles is associated with formation of a tubular membranestructure that is inactive in the turbidity assay. In zymogramanalysis proteins are separated by electrophoresis before theyare renatured, and the presence of P15 in membrane vesiclesdevoid of DNA is revealed. The missing (P20^- particles) or considerablyreduced (P22^- particles) P15 activities in mutants devoid oftwo small integral membrane proteins indicate that these proteinsare involved in positioning P15 in the virus particle or areresponsible for providing the phage membrane integrity neededfor proper binding or activity of P15. As the absence of proteinsP20 and P22 results in DNA leakage from virions (36), we envisionthat these proteins, as well as protein P15, are located ator near the vertex. This is also consistent with protein P15lytic enzyme activity.

Proteins P2 and P15 are translated from the same mRNA (20),and the reduced amounts of protein P2 in P15^- mutant-infectedcells and the resulting particles are probably due to the polareffects of the amber mutation in gene XV. However, the polarity-causedP2 deficiency could be bypassed by providing P2 in trans. Althoughsuch viruses (P2⁺ P15^-) had normal receptor-binding efficiency,they were compromised in terms of initiating infection. Thisimplies either that in the absence of P15 protein P2 does notproperly function in signaling initiation of the DNA deliveryprocess or that P15 somehow is part of the signal transductionpathway. This also supports the idea that P15 is located atthe vertex.

The lytic transglycosylase activity exhibited by protein P7was previously shown to have an accessory role in PRD1 DNA entry(46). In the absence of this activity phage DNA replicationand host cell lysis are both delayed and asynchronous, althoughthe activity is not absolutely essential for infectivity. Thiscould be due to the presence of another lytic enzyme in thevirion. However, in the absence of protein P15 no delay in phageDNA replication was detected, suggesting that protein P7 isthe main muralytic enzyme involved in bacteriophage PRD1 DNAentry into Salmonella cells. As our attempts to isolate P7^-P15^- double mutants have been unsuccessful so far, we do notknow whether the lytic activities of proteins P7 and P15 canreplace each other or whether they are dispensable for PRD1DNA entry. E. coli cells are known to contain at least six differentlytic transglycosylases located at the outer membrane or inthe periplasmic space (17–19, 25, 26, 32, 51), and ithas been proposed that these enzymes function in enlargementand division of the murein sacculus (23–25). We supposethat it would be advantageous for the infecting bacteriophageto use a lytic transglycosylase activity to locally digest thepeptidoglycan layer as the enzymatic machinery needed to repairthe damage would be readily available in the host cell. As aresult, the host cell integrity needed for production of progenyphages would be maintained.

Bacteriophage PRD1 can infect a great variety of gram-negativehosts, including Escherichia, Pseudomonas, Salmonella, Proteus,Vibrio, and Acinetobacter species (44). It is the type virusof the family Tectiviridae, which also contains phages thatinfect gram-positive Bacillus species (AP50 and ${phi}$ NS11) (2, 41,49). Interestingly, in zymograms protein P15 of PRD1 was ableto degrade peptidoglycan isolated from the gram-positive M.lysodeikticus. We propose that the presence of two lytic enzymesin PRD1 might be a common feature in tectiviruses that infectgram-negative and gram-positive hosts; i.e., the tools for penetratingboth types of cell walls are always available. However, it remainsto be seen how closely related tectiviruses that infect gram-negativeand gram-positive hosts are at the genome sequence level.

ACKNOWLEDGMENTS

We thank Marja-Leena Perälä for her excellent technicalassistance.

Helsinki Graduate School in Biotechnology and Molecular Biologyis acknowledged for a fellowship to P.S.R. This work was supportedby research grants 162993, 164298, and 168694 [Finnish Centreof Excellence Programme (2000-2005)] to D.H.B. from the Academyof Finland.

FOOTNOTES

* Corresponding author. Mailing address: Viikki Biocenter, P.O. Box 56 (Viikinkaari 5), FIN-00014 University of Helsinki, Finland. Phone: 358-9-19159100. Fax: 358-9-19159098. E-mail: dennis.bamford{at}helsinki.fi.

REFERENCES

Top