Burkholderia cenocepacia Phage BcepMu and a Family of Mu-like Phages Encoding Potential Pathogenesis Factors

doi:10.1016/j.jmb.2004.04.053

J. Mol. Biol. (2004) 340, 49–65

Burkholderia cenocepacia Phage BcepMu and a Family of Mu-like Phages Encoding Potential Pathogenesis Factors Elizabeth J. Summer1, Carlos F. Gonzalez2, Thomas Carlisle3 Leslie M. Mebane3, Andrea M. Cass1, Christos G. Savva4 John J. LiPuma5 and Ry Young1* 1 Department of Biochemistry and Biophysics, Texas A&M University, College Station, TX 77843-2128, USA 2

Department of Plant Pathology and Microbiology Texas A&M University College Station, TX 77843-2132, USA 3

Department of Biology, Texas A&M University, College Station, TX 77843-3258, USA 4

Microscopy and Imaging Center, Texas A&M University College Station, TX 77843-2257, USA 5

Department of Pediatrics and Communicable Diseases University of Michigan Medical School, Ann Arbor, MI 48109, USA

We have isolated BcepMu, a Mu-like bacteriophage whose host range includes human pathogenic Burkholderia cenocepacia (formally B. cepacia genomovar III) isolates, and determined its complete 36,748 bp genomic sequence. Like enteric bacteriophage Mu, the BcepMu genomic DNA is flanked by variable host sequences, a result of transposon-mediated replication. The BcepMu genome encodes 53 proteins, including capsid assembly components related to those of Mu, and tail sheath and tube proteins related to those of bacteriophage P2. Seventeen of the BcepMu genes were demonstrated to encode homotypic interacting domains by using a cI fusion system. Most BcepMu genes have close homologs to prophage elements present in the two published Salmonella typhi genomes, and in the database sequences of Photorhabdus luminescens, and Chromobacterium violaceum. These prophage elements, designated SalMu, PhotoMu and ChromoMu, respectively, are collinear with BcepMu through nearly their entire lengths and show only limited mosaicism, despite the divergent characters of their hosts. The BcepMu family of Mu-like phages has a number of notable differences from Mu. Most significantly, the critical left end region of BcepMu is inverted with respect to Mu, and the BcepMu family of transposases is clearly of a distinct lineage with different molecular requirements at the transposon ends. Interestingly, a survey of 33 B. cepacia complex strains indicated that the BcepMu prophage is widespread in human pathogenic B. cenocepacia ET12 lineage isolates, but not in isolates from the PHDC or Midwest lineages. Identified members of the BcepMu family all contain a gene possibly involved in bacterial pathogenicity, a homolog of the type-twosecretion component exeA, but only BcepMu also carries a lipopolysaccharide modification acyltransferase which may also contribute a pathogenicity factor. q 2004 Elsevier Ltd. All rights reserved.

*Corresponding author

Keywords: Burkholderia; phage; BcepMu; mu; genomics

Present addresses: T. Carlisle & L. M. Mebane, Department of Biology, Building 68-135, Massachusetts Institute of Technology, 31 Ames St., Cambridge, MA 02139, USA; A. M. Cass, Department of Immunology, UT Houston MD Anderson Cancer Center, 7455, Fannin Box 901, Houston, TX 77030, USA. Abbreviations used: EM, electron micrograph(s); BCC, Burkholderia cepacia complex; CF, cystic fibrosis; PFU, plaque-forming unit(s); IST, interactive sequence tag; TMD, N-terminal transmembrane domain; SAR, signal arrest and release; CFU, colony forming unit(s); MOI, multiplicity of infection. E-mail address of the corresponding author: [email protected]

Introduction The Burkholderia cepacia complex (BCC) consists of nine genomovars recently given species status.1,2 BCC members include plant and animal pathogens as well as catabolically active soil saprophytes. Although not typically pathogenic for humans, some BCC members can cause life-threatening respiratory infections, particularly in persons with cystic fibrosis (CF). Recent studies indicate that 85% to 90% of strains isolated from infected CF patients are B. cenocepacia or B. multivorans, but

0022-2836/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.

50 other BCC species are also found infrequently.3,4 Several epidemic clonal lineages of B. cenocepacia have been identified4,5 including ET12, responsible for infecting many CF patients in Canada and the UK,6,7 PHDC,8 responsible for nearly all B. cepacia complex infections in the mid-Atlantic region of the USA, and the Midwest clone, responsible for infecting numerous patients in CF centers in the mid-western region of the USA.9 The factors that account for the apparent enhanced capacity of epidemic clones for human infection are unknown. The present study reports on the isolation, physical characterization and genomic organization of a temperate bacteriophage recovered from a culture of the B. cenocepacia strain J2315 of ET12 lineage. The distribution of the prophage within the BCC, the structure of closely related prophages in other widely divergent bacteria, and potential pathogenicity determinants carried by the phage are described.

Results and Discussion Production of phage particles by lysogenic Burkholderia cenocepacia J2315 The subject phage, originally designated 56, was isolated from an uninduced culture of B. cenocepacia J2315 by plating on another strain also of ET-12 lineage, strain K56-2. Phage purified from turbid plaques in the K56-2 lawn did not plaque on J2315. Electron micrographs (EM) of negatively stained virions revealed an icosahedral head and contractile tail, the features of a typical myophage (Figure 1A). The particles were found to be very unstable in lysates, with a titer dropping rapidly after isolation, and EM images of these

Figure 1. Morphology of intact (A) and inactivated (B) particles of phage 56 (BcepMu). The bar represents 100 nm. Phage lysates were prepared and either imaged immediately (A) or stored for one week at 4 8C prior to imaging (B). Image B shows a typical disintegrating particle with a broken head and partially exposed tail core.

Burkholderia cenocepacia Phage BcepMu

samples showed many empty capsids with contracted tails (Figure 1B). The average particle had a head size of 80 nm and a tail length of 220 nm. A one-step growth curve revealed an average burst of 150 PFU, an eclipse period of approximately 100 minutes, and a latent period of 135 minutes at 28 8C (Figure 2). Sequence analysis of the BcepMu genome The genome size of phage 56 was estimated to be approximately 37 kb by pulsed-field gel electrophoresis, and no genomic ladder, indicating endannealed multimers, was observed, suggesting that the phage uses pac-type DNA packaging, rather than cos (results not shown). A random library of the phage DNA was sequenced until approximately an eightfold coverage was obtained. The assembled reads from 673 reactions resulted in the production of a single contig of approximately 37 kb. The average amount of sequence obtained from each reaction was 473 bp, resulting in greater than eightfold genome coverage. Analysis of the sequence alignments revealed a striking feature of the phage genomic sequence. At the left end side of the consensus sequence, there were 15 clones with chimeric inserts, consisting of up to 180 bp of heterogeneous sequence that ended abruptly with the start of the phage genomic consensus sequence (Figure 3A). All

Figure 2. BcepMu latent period and burst size. B. cenocepacia cells growing at 28 8C were infected at low MOI and virion production was measured in samples taken at various times after infection, as described in Materials and Methods. The squares represent PFU in CHCl3-treated samples (intracellular and free virions), whereas the circles represent infective centers. The average burst size was estimated from the ratio of the plateaus in the infective centers curve, the latent period from the time where the increase in free virions was approximately 50% completed, and the eclipse period from the approximate time when the first virion (per infected cell) is detectable in the CHCl3-treated samples.

51

Burkholderia cenocepacia Phage BcepMu

(Figure 3C). In contrast, in 15 of the 23 clones, the heterogeneous sequences were placed to widely dispersed locations in the three chromosomes of B. cenocepacia J2315. The remaining heterogeneous sequences had no significant homologs in the NCBI databases and were less than 80% identical to J2315 sequences. Presumably these represent sequences specific to the host, strain K56-2, from which the phage stock was prepared. The presence of the variable host sequences at either end of the genomic DNA is a unique feature of phage related to the transposable coliphage Mu and represents packaging from the random integration sites used during transpositional replication. Because of this relationship to Mu, the bacteriophage 56 derived from the lysogenic strain B. cenocepacia J2315 was designated BcepMu. Removal of the variable host DNA sequences from the BcepMu ends resulted in a unique coding region length of 36,748 bp. BcepMu prophages in the B. cepacia complex species

Figure 3. The BcepMu genomic DNA termini are flanked by random host-derived sequences. A, Uppercase letters: alignment of five independently isolated B. cenocepacia K56-2 host/ left end phage DNA junction clones. Lowercase letters: B. cenocepacia J2315 host/ BcepMu prophage left end junction sequence. BcepMu sequences are underlined. B, Uppercase letters: alignment of three independently isolated right end phage/ B. cenocepacia K56-2 host DNA junction clones. Lowercase letters: BcepMu prophage right end/B. cenocepacia J2315 host junction sequence. BcepMu sequences are underlined. C, Position of the BcepMu prophage in B. cenocepacia J2315 chromosome 3. The diagram is not drawn to scale.

sequences of independent clones corresponding to the left end of the phage genome were oriented towards the middle of the phage genome, implying that the amount of chimeric sequences at the left end of the phage genome was not significantly more than 180 bp. In contrast, at the right end of the phage genome, there were 16 independent clones reading out of the phage sequences, implying that there must be more than 1000 bp of host DNA in the clones, based on the average insert size (Figure 3B). The lysogenic phage 56 host, B. cenecepacia J2315, is currently being sequenced at the Sanger Institute. Blast analysis with 23 chimeric sequences from the ends of the contigs against the sequenced bacterial genome† revealed that in each case the phage sequences were identical to a region of the B. cenocepacia J2315 sequence, and thus describe the ends of a prophage † http://www.sanger.ac.uk/Projects/B_cepacia/ blast_server.shtml

The BcepMu prophage was found to be located in chromosome 3 of B. cenocepacia J2315 between positions 609,954 and 573,207‡ (Figure 3C). The prophage sequence was inserted after codon 712 of a putative transcriptional regulator with an AAA signature domain. The prophage sequence lacked the characteristic 5 bp direct repeat that flanks Mu prophage sites, suggesting that there had been an additional deletion or rearrangement after the prophage insertion (Figure 3). The distribution of BcepMu among B. cepacia complex species was determined by PCR with BcepMu-specific primers and assaying overnight cultures for PFU on lawns of K56-2. Species (strains) tested were: B. cepacia (ATCC25416), B. multivorans (ATCC17616), B. cenocepacia (J2315), B. stabilis (LMG14294), B. vietnamiensis (PC259), B. dolosa (AU0645), B. ambifaria (AMMD), B. anthina (AU1293) and B. pyrrrocinia (BCO11). Among these nine strains representing the nine species of the BCC, only B. cenocepacia strain J2315 (a representative of the ET12 clonal lineage) produced a positive result with PCR and produced phage particles capable of forming plaques on K56-2. Further testing of 14 B. cenocepacia isolates of ET12 lineage, including BC7 and C542410 and ten ET12 isolates recovered from CF sputum cultures,11 revealed that all except K56-2 were PCR-positive for BcepMu, spontaneously produced an average of 106 PFU/ml in uninduced overnight broth cultures and did not support plaque formation with BcepMu. Fifteen other B. cenocepacia strains, including five isolates each of the PHDC and Midwest clonal lineages and five with unique genotypes12 were found to be PCR-negative for the prophage, did not produce the phage and were not sensitive ‡ http://www.sanger.ac.uk/Projects/B_cenocepacia/

Figure 4 (legend opposite)

53

Burkholderia cenocepacia Phage BcepMu

to BcepMu, suggesting that they are neither lysogenic, nor hosts for BcepMu. These data indicate that BcepMu is restricted to the ET12 clonal lineage of B. cenocepacia. BcepMu-like prophages in Salmonella typhi, Photorhabdus luminescens, and Chromobacterium violaceum From a combination of sequence analysis and experimental approaches, BcepMu was found to encode 53 proteins (Figure 4 and Table 1). Most of the proteins were related to sequences already present in the public databases. Thirty-two of the protein sequences have homologs in a prophage element identified by Deng et al.13 and Parkhill et al.14 in the genomes of two Salmonella typhi strains, Ty2 (NC_004631) and CT18 (NC_003198; Figure 4). In a recent analysis of prophage elements in bacterial genomes, the S. typhi CT18 prophage was given the designation Sti3 and classified as a Mu-like phage with an inverted early region.15 Based on the extensive homology with BcepMu, we designated the Sti3 prophage type as SalMu, and the individual prophages SalMu-Ty2 and SalMu-CT18. Similar contiguous homologies were found in the genomes of P. luminescens (NC_005126, inclusive of entries gi:36786769 and gi:36786730)16 and C. violaceum (NC_005085, inclusive of entries gi:34497620 and gi:34497570)17 and accordingly these elements were designated as the prophages PhotoMu and ChromoMu, respectively (Figure 4). Dot-matrix analysis reveals that, except for a 3.3 kb region near the left end, and the extreme 3.5 kb at the right end, BcepMu shows across its entire length significant DNA sequence similarity with the SalMu, PhotoMu and ChromoMu prophage elements (Figure 4). This is remarkable, considering the differences in GC content of the four hosts (BcepMu 63%, Burkholderia 63%; SalMu 55%, Salmonella 52%; ChromoMu 62%, C. violaceum 65%; PhotoMu 43%, Photorhabdus 43%). Thus, within this group of prophages, there is much less of the mosaicism noted for the lambdoid phages.18 However, there is evidence of gross changes at the level of the functional gene

clusters and also for the insertion of morons, as defined by Hendrix et al.19 (see below). The striking homology of this group of phages from diverse hosts at both the DNA and protein sequence levels and significant dissimilarities with Mu20 (see below) has led us to designate BcepMu, SalMu, ChromoMu, and PhotoMu as the BcepMu family of phages. By comparison with the S. typhimurium LT2 genomic sequence (NC_003197)21 which lacks the SalMu prophage, it was possible to map the insertion site and prophage boundaries of SalMu in the S. typhi TY2 and CT18 genomes (Figure 5). The SalMu prophage insertion site is located within a chromosomal region that is inverted between Ty2 and CT18 but is otherwise identical in the two strains. The Salmonella prophage sequences are almost identical, with two notable exceptions in the endolysin coding region and the invertible tail fiber region13 (and see below) (Figure 5). The SalMu genome ends have similar features to those of BcepMu (Figure 6), including a large imperfect inverted terminal repeat with a perfect inverted repeat of 4 bp at the extreme termini. Both SalMu prophages have modifications at the insertion site: a deletion of 82 nucleotides relative to S. typhimurium LT2 and the creation of flanking 6 nt direct repeats (CCTGAT) (Figure 5). This strongly suggests that these prophages originated from a common insertion event rather than resulting from independent lysogenization events. Interactive sequence tag (IST) analysis of BcepMu A major difficulty in annotating sequenced genomes is providing experimental evidence for the validity of gene prediction. We took advantage of the lambda cI repressor fusion system to screen the BcepMu genome for homotypic interacting domains by generation of an IST library.22 The presence of phage sequences in an IST library is a strong indicator that the sequences correspond to real gene products, as long as certain problematic types of sequences coding for cysteine stretches and other unusual motifs are avoided.23 Sequence

Figure 4. Maps of the BcepMu and related prophage genomes. Genomic maps of the BcepMu phage genome and comparison with the genomes of the related prophages SalMu-Ty2, ChromoMu and PhotoMu (derived from NC_004631, NC_005085_and NC_005126, respectively) are shown. Boxes are predicted genes drawn to scale and placed above or below the genome scale bar, based on transcription orientation (above, left to right). Gene numbers are listed in the boxes and homolog or functional assignments, where possible, are indicated (Table 1). White boxes indicate genes with BcepMu homologues in at least one of the three prophage genomes; the corresponding genes in the prophage sequences are annotated with the cognate BcepMu gene. Gray boxes in BcepMu indicate genes with no homologs within this family of phage genomes. Black boxes in the related prophages are genes, which do not have homologs in BcepMu. To aid in navigation, a few database cross-reference numbers are given for the SalMu, ChromoMu and PhotoMu genes. Below, Dotmatcher plots comparing BcepMu to each of the prophage sequences, with the window size and the threshold both set to 100. The deletion of the lysogenic control region predicted from the Dotmatcher comparison of the SalMu and BcepMu genomes is shown below the BcepMu map as a black bar, with the deletion site in SalMu indicated by an arrow. The SalMu map was generated by DNA Master from a re-annotation of the SalMu coding region based on the analysis of the BcepMu sequence, whereas PhotoMu and ChromoMu maps are based on the original sequences extracted from the GenBank entries.

54

Burkholderia cenocepacia Phage BcepMu

Table 1. BcepMu gene assignments Gene

Gene name

Code.strd.

Left

Right

AA

pI/Mw (PredHel/SP)

Representative homologues (Query genome, accession gene, expect) (C.v. gil34497620, 2e-38) (P.l. gil36786769 PLU3461, 2e-27) (S.e. CT18 gil16760386, STY1591, 2e-32; t1397) (Mu gil9633507 Middle operon reg, Mor, gp17, 9e-08) (Mu gil9633511 (prt. C), 9e-06) C.v. gil34497619, 2e-36) (P.l. gil36786768 PLU3460, 1e-20) (S.e. CT18 gil16760387, STY1592, 1e-19, t1396) (Mu gil9633506 prt. gp16, 1e-09) (Phage lambda gil9626285 ea10, 3e-12) (B. stear.gil227370 DNA binding prt. HU, 6e-24) (C.v. gil34497616,2e-52) (Vibriophage VpV262 gil21234384,2e-04) No hits (C.v. gil34497612,e-151) (P.l. gil36786766, e-111) (S.e. CT18 AF091717 ehe STY1603, e-101, t1385) (Al. hyd. gil1170043 exeA, 6e-09) (P.l. gil36786765, e-170) (C.v. gil34497611, 2e-87) (S.e. CT18 gil16760398,STY1604 0, t1384) (M. tub. CDC1551 gil15842348 IS1604, transposase, 8e-06) (Tn552 gil136133 TN552transposase, 0.005) (P.l. gil36786764, 7e-23) (S.e. CT18 gil16760400,STY1606, 5e-13, t1382) No hits (C.v. gil34497609,4e-35) (P.l. gil36786763, 3e-04) (C.v. gil34497608, 4e-59) No hits No hits (C.v. gil34497606,8e-32) (P.l. gil36786762, 4.e-08) (C.v. gil34497605, 2e-18) (P.l. gil36786761, 4e-05) (E. coli gil15802321 CP-933T put. trans. reg., 9e-15) No hits (X. fast. 9a5c gil15839102, 4e-12)

BcepMu1

Mu protein C/Mor gp17

–

365

736

124

6.12/13,997

BcepMu2

Mu gp16 gemA

–

733

1173

146

9.83/16,725

BcepMu3

ssdbp

–

1157

1561

135

4.34/15,093

BcepMu4

HU dbp

–

1592

1864

91

9.92/9332

BcepMu5 BcepMu6

con. protein, IST con. protein, IST

– –

1928 2668

2545 3105

206 141

5.76/23,298 5.84/16,183

BcepMu7 BcepMu8

hyp. protein exe/ eha subunit, IST

– –

3126 3457

3455 4671

110 405

10.02/11889 6.25/44,578

BcepMu9

Tn552/IS1604 rve transposase

–

4668

6467

600

8.61/65,975

BcepMu10

con. protein, IST

–

6485

7438

318

5.04/34,213

BcepMu11 BcepMu12

hyp. protein con. protein, IST

– –

7449 7672

7661 7974

71 101

4.72/7705 10.13/10878

BcepMu13 BcepMu14 BcepMu15 BcepMu16

con. protein Novel protein, IST Novel protein, IST con. hyp. protein

– þ – –

7971 8571 8638 8856

8450 8807 8970 9098

159 79 111 81

6.38/18,056 6.56/8713 12.00/12082 10.23/8576

BcepMu17

cro/cI repressor tns. reg., IST

þ

9226

9642

139

9.22/14,969

BcepMu18 BcepMu19

hyp. protein con. hyp. protein

þ þ

9639 10,663

10,421 11,199

261 179

BcepMu20 BcepMu21

hyp. protein holin

þ þ

11,264 11,555

11,488 11,902

75 116

BcepMu22

SLT

þ

11905

12516

204

BcepMu23

Rz

þ

12,513

13,112

200

BcepMu24

Rz1, out of frame, prolinerich protein Putative inner membrane prt.

þ

12,820

13,074

84

þ

13,109

13,447

113

7.72/28,155 8.89/20,340 (2/SP) 4.67/8233 9.45/12,658 (3/SP) 10.17/22,285 (SP) 6.09/20,749 (1/SP) 11.76/8870 (SP) 6.59/13,140 (1/SP)

BcepMu26

con. hyp. protein

þ

13,444

13,776

111

5.84/12,854

BcepMu27

TerS, IST

þ

13,778

14,323

182

5.98/19,629

BcepMu28

Mu TerL

þ

14,320

15,822

501

6.02/55,735

BcepMu25

No hits (P.l. gil36786756, 8e-16) (S.e. CT18 gil16760402,STY1608, 3e-17, t1380) (P.l. gil36786755, 2e-50) (S.e. CT18 gil16760403, STY1609, 8e-18, t1379) No hits No hits (P.l. gil36786753,7e-21) (S.e. CT18 gil16760406 prob. membrane prt. STY1612, 2e-20, t1376) (P.l. gil36786752,3e-29) (S.e. CT18 gil16760407,STY1613, 8e-29,t1375) (C.v. gil34497596,CHP, e-59) (P.l. gil36786751,8e-56) (S.e. CT18 gil16760408,STY1614, 2e-60,t1374) (H. inf.gil16273401 FluMu prt. gp27, 2e-09) (C.v. gil34497595 put. portal prt. 0) (P.l. gil36786750, e-175) (S.e. CT18 gil16760409, STY1615, e-176,t1373) (phage Mu gil9633519 put. portal prt. gp28, e-108) (continued)

55

Burkholderia cenocepacia Phage BcepMu

Table 1 Continued Gene

Gene name

Code.strd.

Left

Right

AA

pI/Mw (PredHel/SP)

Representative homologues (Query genome, accession gene, expect)

BcepMu29

Mu gp29 homologue

þ

15,819

17,294

492

4.93/54,396

BcepMu30

Mu F virion morphogenesis protein

þ

17,287

18,123

279

6.85/31,894

BcepMu31

Mu G

þ

18,120

18,647

176

9.57/19,011

BcepMu32

con. protein, IST Mu I equivalent

þ

18,850

19,980

377

5.38/41,115

BcepMu33

con. protein, Mu Z equivalent

þ

19,510

19,980

156

5.77/17,254

BcepMu34

con. hyp. protein Mu T equivalent

þ

20,026

20,949

308

5.94/34,129

BcepMu35

con. hyp. protein

þ

21,024

21,356

111

4.67/10,873

BcepMu36

Mu gp36 homologue, IST

þ

21,358

21,810

151

6.30/16,828

BcepMu37

con. protein, IST

þ

21,810

22,274

155

4.80/16,949

BcepMu38

con. hyp. protein

þ

22,271

22,516

82

4.71/8541

BcepMu39

P2 major tail sheath prt, F1

þ

22520

23953

478

5.51/50926

BcepMu40

P2 major tail tube protein FII

þ

23956

24480

175

5.45/19046

BcepMu41

con. protein IST

þ

24602

24931

110

5.16/11557

BcepMu42

þ

24858

25061

68

4.64/7716

þ

24977

25363

81

6.28/13904

no hits

BcepMu44

Lambda pre-tape measure frameshift prt. alt. gp41 C- termninus, Lambda T eqv. P2 gpT tail protein, IST

C.v. gil34497594 cons. hyp. Prt. 0) (P.l. gil36786749,e-170) (phage Mu gil9633520 gp29, 2e-30) (S.e. CT18 gil16760410,STY1616, e-166,t1372) (C.v. gil34497593, e-126) (P.l. gil36786748, e-97) (S.e. CT18 gil16760411,STY1617, e-97, t1371) (Mu gil9633521 virion morphogenesis late F orf gp30, 5e-69) (C.v. gil34497592,l 4e-27) (P.l. gil36786747, 5e-24) (S.e. CT18 gil16760412 phage prt. STY1618, 6e-32,t1370) (Mu gil9633522 virion morphogenesis late G orf, 3e-06) (C.v. gil34497590, 2e-33) (P.l. gil36786746, 2e-20) (S.e. CT18 gil16760413,STY1619, e-19,t1369) (C.v. gil34497590, e-116) (P.l. gil36786746, 7e-94) (S.e. CT18 gil16760413,STY1619, 2e-92,t1369) (C.v. gil34497589, 5e-79) (P.l. gil36786745, 2e-67) (S.e. CT18 gil16760414,STY1620, 5e-67,t1368) (C.v. gil34497587l, 5.e-25) (P.l. gil36786744, 2e-19) (S.e. CT18 gil16760415, STY1621 3e-21,t1368) (C.v. gil34497586, e-48) (P.l. gil36786743,2e-28) (S.e. CT18 gil16760416 STY1622, 3e-35,t1366) (Mu gil9633527 gp36, 3e-05) (C.v. gil34497585,9e-44) (P.l. gil36786742,1e-37) (S.e. CT18 gil16760417,STY1623, 5e-39,t1365) (C.v. gil34497584, 2e-12) (P.l. gil36786741,2e-08) (S.e. CT18 gil16760418,STY1624, 4e-08,t1364) (C.v. gil34497583, 0) (P.l. gil36786740,0) (S.e. CT18 gil16760419, t1363) (phage P2 gil9630349 gpFI, 2e-34) (C.v. gil34497582, 2e-75) (P.l. gil36786739,2e-69) (S.e. CT18 gil16760420, 3e-67, t1362) (P2 gil9630350 gpFII, 2e7) (C.v. gil34497578,4e-14) (P.l. gil36786738,2e-06) (S.e. CT18 gil16760422,STY1627, 5e-04, t1361) no hits

þ

25336

27876

847

10.26/89218 (6)

BcepMu45

P2 gpU protein

þ

27878

28768

297

8.89/31095

BcepMu46

putative tail fibre protein

þ

28768

28977

70

4.28/7695

BcepMu47

P2 gpD reg.y protein, IST

þ

28965

30170

402

9.64/43073

BcepMu48

P2 gpV base plate, IST

þ

30167

30769

201

5.02/21365

(P.l. gil36786737,e-152) (C.v. gil34497576, e-132) (S.e. CT18 gil16760423 STY1629, e-150), t1359) (phi CTX gil17313245 orf25 similar to T gene of P2, 2e-17) (C.v. gil34497575,9e-66) (P.l. gil36786736,8e-57) (S.e. CT18 gil16760424,STY1630, 2e-56, t1358) (C.v. gil34497574 tail fibre prt, 4e-19) (S.e. CT18 gil16760425 tail fibre prt. STY1631, e-18, t1357) (C.v. gil34497573, e-128) (P.l. gil36786735,8e-91) (S.e. CT18 gil16760426. STY1632, 4e-87, t1356) (phage P2 gil9630355 gpD, 9e-20) (C.v. gil34497572, 9e-75) (P.l. gil36786734,e-36) (S.e. CT18 gil16760427 STY1633, 2e-29, t1355) (phage P2 gil9630342 gpV, 2e-10)

BcepMu43

(continued)

56

Burkholderia cenocepacia Phage BcepMu

Table 1 Continued Gene

Gene name

Code.strd.

Left

Right

AA

pI/Mw (PredHel/SP)

Representative homologues (Query genome, accession gene, expect) (P.l. gil36786733l,7e-32) (S.e. CT18 gil16760428 put. phage baseP.l.ate prt. STY1634, 5e-33, t1354) (T4 gil9632602 gp25 baseplate wedge subunit, 0.64) (C.v. gil34497570, e-116) (P.l. gil36786732,2e-95) (S.e. CT18 gil16760429 STY1635, e-104, t1397) (phage P2 gil9630344 baseplate assembly prt. J, 4e-13) (P.l. gil36786731,7e-47) (S.e. CT18 gil16760430 STY1636, 3e-50, t1352) (P2 gil9630345 gpI, 0.002) (P.l. gil36786730, e-20) (S.e. CT18 gil16760431 STY1637, 8e-33, t1351) (P2 gil9630346 tail fiber prt. gpH, e-06) (lambda gil9911089 tail fiber, 2e-06) (Str. myca. gil285173 3-O-acyltransferase MdmB 9e-12) (Sal. typh. LT2 gil16765560 O-antigen five: acetylation of the O-antigen(LPS), 3e-11)

BcepMu49

P2 gpW, T4 gp25 base plate wedge

þ

30823

31176

118

5.34/13408

BcepMu50

P2 gpJ baseplate assembly protein

þ

31173

32324

384

4.64/41020

BcepMu51

P2 gpI tail fiber protein

þ

32317

32898

194

5.81/21248

BcepMu52

P2 gpH tail fiber protein, IST

þ

32898

35255

786

5.15/80077

BcepMu53

acyltransferase

–

35456

36589

378

9.60/41669 (9/SP)

Abbreviations: Mw, molecular mass; SP, signal peptide; TM, number transmembrane domains; IST, interactive sequence tag domain; put., putative; prt., protein; con., conserved; hyp., hypothetical; eqv., equivalent; S.e., Salmonella enterica; C.v., Chromobacterium violaceum ATCC 12472; P.l., Photorhabdus luminescens subsp. laumondii TTO;, S. typh., Salmonella typhimurum; Str. myca., Streptomyces mycarofaciens; H. inf., Haemophilus influenzae; Ser. marc., Serratia marcescens; N. men., Neisseria meningitidis MC58; X. fast., Xylella fastidiosa; A. hyd., Aeromonas hydrophil; M. tub., Mycobacterium tuberculosis.

was obtained from a total of 68 independently isolated surviving clones. Of these, 20 proved to be derived from host sequences, probably from the variable host ends (results not shown). The inserts of the 48 BcepMu phage-specific IST domain clones were found to be in the reading frames of 16 of the genes initially predicted by Genemark (Table 2). The remaining IST was used to define gene 15 (Tables 1 and 2). Overall organization of the BcepMu genome The BcepMu genome can be divided into functional clusters in the following order from the left end: replication/regulation/pathogenesis, host lysis, capsid formation and tail/tail fiber formation (Table 1; Figures 4 and 7). With the exception noted below, the order of these functional clusters and the arrangement of genes within the clusters is the same as in phage Mu, although the genes encoding tail and base-plate components are much more closely related to the analogous genes in P2 (Figure 7 and Table 1). Inversion of the replication/regulation/ pathogenesis module Compared to Mu, the region of BcepMu spanning genes 1 through 20 is the most diverged module and contains the largest percentage of genes with no detectable homolog in the database (Figure 4 and Table 1). The genes that were found to have homologs in the database are involved in regulation and transposition, as is found in Mu, and also potential modifiers of host pathogenicity.

The entire left end region of BcepMu family phages is inverted with respect to Mu (Figures 4 and 7). In contrast to Mu, where the lysogenic control region and transposase genes are at the extreme left end and the transcriptional regulator genes are on the interior of this segment, the inversion in BcepMu results in homologs of the Mu transcriptional regulator genes at the left end of BcepMu, and the transposase gene in the interior. BcepMu gene 1 encodes a transcriptional regulator with homology to the Mu gp17 middle operon regulator Mor and to the Mu gp21 late operon regulator C. Like these proteins, BcepMu gp1 contains a helix-turn-helix domain typical of this family of DNA binding proteins (results not shown). BcepMu gene 2 is a homolog of the Mu gemA (gene 16), an effector of host gene expression. Moreover, the only good candidate for the lysogenic control region is in the interior of the BcepMu sequence, adjacent to gene 17, which encodes a homolog of a lambdoid repressor and is transcribed from a divergent promoter (Figure 4). Notably, the carboxyl terminal region of BcepMu gp17 was one of the 17 BcepMu genes identified in the IST screen, indicating that its product undergoes the homotypic oligomerization expected for a lysogenic repressor (Table 2). Alignment of BcepMu and SalMu reveals that, beginning at about 7.0 kb from the left end, approximately 4.5 kb of DNA in BcepMu is opposite to an unrelated , 0.5 kb sequence in SalMu (Figures 4 and 7). Analysis of the homologies near the borders of the non-homologous regions indicates there was a , 4 kb deletion of SalMu that would have included genes corresponding to BcepMu genes 11 through 21. This deletion

57

Burkholderia cenocepacia Phage BcepMu

Table 2. Summary of IST analysis Gene BcepMu5 BcepMu6 BcepMu8 BcepMu10 BcepMu12 BcepMu14 BcepMu15 BcepMu17 BcepMu27 BcepMu32 BcepMu36 BcepMu37 BcepMu41 BcepMu44 BcepMu47 BcepMu48 BcepMu52 BcepMu52 a

IST hits

Begina

Enda

3 1 1 3 2 2 1 2 6 3 2 3 3 2 1 3 8 2

1 12 61 161 1 6 1 66 81 130 76 1 1 781 11 1 137 520

142 140 143 297 100 78 110 138 138 186 149 61 65 846 195 105 178 555

Begin and End refer to region of protein sequence cloned as

IST.

Figure 5. The SalMu prophage insertion site and invertible tail fiber genes. A, The position of prophages SalMu-CT18 and SalMu-Ty2 was determined by comparison of sequences from S. typhi strains CT18 (NC_003198) and Ty2 (NC_004631) with that of S. typhimurium LT2 (NC_003197), a Salmonella species that lacks the prophage. Sequences, 82 bp total, of S. typhimurium are deleted in S. typhi Ty2 and CT18 (indicated by italics). Capitalized sequences and arrows indicate the direct duplication. SalMu prophage sequences are shown in bold. SalMu-Ty2 spans nucleotides 1,408,140 to 1,441,928 of entry AE014613 and SalMu-CT18 spans nucleotides 1,538,737 to 1,572,997 of entry AL513382. B, Comparison of the BcepMu and SalMu tail fiber genes by Dotmatcher reveals that the BcepMu 52 has multiple internal duplications of the 50 region of SalMu 48. C, The SalMu-Ty2 and SalMu-CT18 genome maps are identical except for the orientation and subunit composition of an ensemble of paralogous tail fiber C-terminal gene cassettes, separated by sitespecific recombination sites (black stars). Orthologs have the same color in the two maps. The bars terminated by ochre circles indicate one possible set of two independent site-specific recombination events, which would convert SalMu-CT18 to SalMu-Ty2, and the bars terminated by black circles indicate the region which deleted from SalMu-CT18 would convert it to SalMu-Ty2.

removed the lysogenic control region, suggesting that SalMu is an uninducible cryptic prophage (Figures 4 and 7). As the deletion is present in both TY2 and CT18, it most likely occurred before the inversion events in the tail fiber region (see below). Potential pathogenesis determinants BcepMu genes 8 and 53 encode possible

virulence factors. Gene 8 encodes a homolog of ExeA from Aeromonas hydrophila, which participates in the secretion of a subset of toxins, including ExeD, by a type II secretion pathway system.24 The predicted BcepMu gp53 is a membrane protein with homology to the membrane-embedded 3-O-acyltransferase MdmB of Streptomyces mycarofaciens, responsible for resistance against macrolide antibiotics, and to Salmonella typhimurium OafA, which is responsible for acetylation of the lipopolysaccharide O-antigen.25,26 All four BcepMu family members carry the exeA gene, whereas BcepMu alone carries the mdmB homolog (Figure 4 and Table 1). This determinant is likely to be a moron,19 since its GC content (55%) is significantly less than the overall GC of the phage and its host. This is the first time that potential virulence determinants have been associated with a functional transposable phage, a finding that significantly increases the positional diversity available to phage-borne determinants, which have heretofore been restricted to phages that integrate into specific sites in the chromosome. A transposase with Tn552-like binding sites Similarly to the transposases (TnpA proteins) of Mu and the Mu-like prophage elements, BcepMu gp9 and its counterparts in the SalMu, ChromoMu and PhotoMu genomes carry an Rve integrase catalytic core domain (pfam00665)27 and are related to the S. aureus transposon, Tn552. Tn552 shares many features with the Mu and the BcepMu transposases (Figure 6), including the presence of 50 TG…CA 30 dinucleotides at the transposon termini and the ability to transpose into nearly random sites28 – 30 (Figures 3, 5 and 6). The overall identities between these phage TnpA proteins and the Tn552 TnpA are the same (18%; Table 1), but alignment within the catalytic core region reveals clusters of residues conserved between the

58

Burkholderia cenocepacia Phage BcepMu

Figure 6. Transposable ends of BcepMu and SalMu. The left and right transposable ends of BcepMu and SalMu are shown aligned with each other and with the ends of Tn552. Boxes enclose the 23 nt imperfect direct repeats shown to be TnpA binding sites in Tn552.30 The asterisks indicate the nucleotides conserved in the end sequences of the two transposable phages and the transposon, and the carats indicate bases conserved between the two phages. All sequences are written 50 to 30 .

BcepMu-like TnpAs and the Tn552-IS1604 sequences but not the Mu-like TnpAs (not shown). The organization of the BcepMu termini is similar to the transposable ends of Tn552; both contain short (, 23 nt) imperfect direct repeats, which in Tn552 are tandem transposase binding sites required for optimal transposition,28 as part of terminal inverted repeats spanning 48 nt (Figure 6). Remarkably, the central region of each 23 nt repeat is conserved in the left and right ends of the two transposable phages and the transposon. In contrast, the transposable ends of phage Mu contain three TnpA binding sites spanning regions of over 100 nt (not shown). These comparisons suggest that important features of the Tn552 transposition mechanism have been retained in members of the BcepMu family of transpositional replicases but not in Mu.

Lysis gene cassette In phages of Gram-negative bacteria, lysis gene cassettes are often found consisting of genes encoding a holin, an antiholin, an endolysin and Rz/Rz1 equivalents.31 Of these five proteins, only the endolysin is usually identifiable by homology searches; the other protein functional classes have shown a very wide range of diversity. The rationale for the assignment of lysis functions in BcepMu and its family members is thus somewhat indirect. First, none of the BcepMu genes is significantly related to a known phage lysis protein, but gene 22 encodes a 204 residue protein with a soluble lytic transglycosylase (Slt) domain (residues 32 to 157). The Escherichia coli Slt is thought to participate in murein synthesis and remodeling.32,33 Moreover, gp22 has an N-terminal transmembrane domain

Figure 7. BcepMu shows modular homology with phages Mu and P2. GenomePixelizer (Allometra) representations of the Mu, BcepMu, SalMu and P2 coding regions are shown, with lines connecting homologs. The Mu map was derived from NC_000929 and the P2 map was derived from entry NC_001895. Color coding: yellow, genes conserved between Mu, BcepMu and SalMu; blue, genes conserved between BcepMu and SalMu only; green, genes conserved between P2, BcepMu and SalMu; red, genes homologous in all four phages; gray, genes not shared between any of these phages. Functional gene modules are labeled for Mu and P2. Relevant genes are annotated from the Mu and P2 maps.

Burkholderia cenocepacia Phage BcepMu

(TMD) that resembles the signal arrest and release (SAR) sequence of the phage P1 Lyz endolysin, which has canonical glycosidase or lysozyme muralytic activity (Figure 8). The phage P1 SAR sequence is an unusual TMD that initially directs integration into the inner membrane but is ultimately extracted from the membrane, presumably as a result of holin function, allowing activation of the endolysin and its access to the peptidoglycan.34 SAR sequences have been identified in many phage glycosidase endolysins, indicating that this is an unexpectedly common mode for holinmediated control of endolysin activity, but this is the first report of a transglycosylase with a SAR sequence. This is also the first report of a phage endolysin closely related to a bacterial Slt. Taken

59

together, these perspectives suggest the attractive notion that during evolution, BcepMu acquired its muralytic lysis protein from its host and then replaced its signal sequence with a SAR sequence to subjugate it to holin control. It should be noted that the N termini of the predicted SalMu and PhotoMu endolysins are intriguingly different from the SAR motif; the SalMu protein would have two TMDs, whereas the PhotoMu TMD is sufficiently hydrophobic to make SAR function unlikely, indicating that there must be other modes by which membrane-bound endolysins can be controlled by holin function. Identification of holins is always problematical, given that more than 60 unrelated families of holins are known. However, considering its

Figure 8. Lysis genes. A, The sequences of the putative holins of the BcepMu phage family are shown, with the orthologous BcepMu, SalMu, and PhotoMu sequences aligned. The predicted TMDs of the three orthologous holins are indicated by underline, and the charged residues of the BcepMu holin are shown. B, Alignments of the N-proximal segments of the predicted endolysins of BcepMu, SalMu, and PhotoMu with the Salmonella Slt autolysin (AAL23041) are shown, with the predicted TMDs in the endolysins and the Slt signal sequence underlined. Charged residues in the BcepMu gp22 are shown above the sequence alignment, and residues completely conserved in all four sequences are indicated by asterisks. The arrow marks the signal sequence cleavage site in Slt. C, The predicted mature lipoprotein sequences for the Rz1 analogs from BcepMu, SalMu and PhotoMu are shown, with the Pro residues in bold. The fraction of Pro residues is: BcepMu 15/65; SalMu 11/71; PhotoMu 9/67. The PhotoMu Rz1 sequence is not annotated in the P. luminescens GenBank entry.

60

location adjacent to the putative endolysin gene (Figures 4 and 7), gp21 is the best candidate because its 116 residue sequence has features of a class II holin, including two TMDs, although it has an unusual hydrophobic domain at its extreme C terminus nearly long enough to span the membrane a third time (Figure 8A). The last two genes of the lysis cassette have no sequence homologs in the database but can be securely identified on the basis of unique structural features. Gene 23 encodes a polypeptide predicted to have a cleavable signal sequence and thus to be an exported periplasmic protein. Gene 24 is entirely embedded out of frame within gene 23 and encodes a polypeptide of 85 residues with a signal-peptidase II cleavage site. Thus, processing of gp24 would generate a 65 residue proline-rich lipoprotein anchored to the inner leaflet of the outer membrane by cysteine-linked lipids (Figure 8C). These structural features, combined with the adjacency to the holin-endolysin lysis genes, allow the unambiguous conclusion that the nested genes 23 and 24 are members of the very diverse class of Rz-Rz1 embedded gene pairs, thought to encode proteins involved in attacking the outer membrane or links between the outer membrane and the cell wall during lysis.35 – 37 The Rz/Rz1 gene pair is the only example in biology of embedded genes required for the same biological function.37 Interestingly, analogous embedded gene pairs can be seen in the SalMu and PhotoMu lysis cassettes, encoding polypeptides with the same characteristic features but lacking sequence similarity with each other or with the BcepMu Rz-Rz1 equivalents. Thus, the BcepMu family further extends the diversity of the enigmatic and unusual Rz-Rz1 functional gene class. The BcepMu lysis cassette, although shown by dot matrix analysis to be significantly related to that of the SalMu and PhotoMu prophage elements, is completely unrelated to that of Mu (Table 1). In ChromoMu, the region which by analogy should contain the lysis genes has no sequence related to known proteins, despite the fact that gene 20 encodes a canonical type II holin protein (Figure 8). This suggests that ChromoMu also encodes an endolysin that does not belong to one of the four known classes of muralytic enzymes. Moreover, neither a holin nor an Rz-Rz1 gene pair can be discerned. However, speculation about unique lysis genes in ChromoMu is unwarranted unless the prophage can be shown to give rise to functional lytic virions. Capsid, tail and tail-fiber genes BcepMu genes 27 through 52 are mostly concerned with virion morphogenesis. These genes showed modular homology with Mu and P2, with capsid components being Mu-like and the tail components being P2-like (Figure 7). BcepMu genes 27 through 31 and gene 36 share homology and order with Mu genes 27 through 31 and 36, respectively,

Burkholderia cenocepacia Phage BcepMu

which are genes involved in DNA packaging and capsid formation.20 Functions could be assigned to these BcepMu genes based on the Mu homologies, although with high E-values, conserved gene order, and on the similarity in predicted protein sizes. Moreover, codon 222 of gene 32 is a GTG potential start codon served by a consensus Shine – Dalgarno sequence with perfect spacing (not shown). The positioning and relative sizes of gene 32 and this internal reading frame, which has been designated gene 33, are similar to the organization of Mu genes I and Z, encoding the capsid protease and scaffold, respectively.20 BcepMu genes 39, 40, and 44 through 52 show homology to genes involved in formation of the contractile tail, including the tail tube, tail sheath, base plate and tail fibers. Instead of sharing homology with Mu, these are related primarily to five other myophages regarded as being members of the P2 family: the coliphages P2 and 186, Pseudomonas aeruginosa phage phiCTX, Vibrio harveyi phage VHML,38 – 42 and Yersinia pestis phage L-413C (NC_004745). As has been customary, the BcepMu genes are named after the P2 homolog even when the most significant relationship is with other family members. This is the case with BcepMu gene 44, significantly related at the protein sequence level only to the phiCTX homolog, which is in turn related to the P2 tape-measure protein, T. The gene preceding the tape-measure gene in many phages, including l (genes G and T), Mu (genes 41 and 41.5) and P2 (genes E and E0 ), has an alternate reading frame for the C-terminal coding sequences, resulting in the production of two proteins, both essential for tail assembly, with common N termini and different C termini.20,43,44 Although there is no sequence similarity to these genes, genes 42 and 43 may play this same role in BcepMu, based on the presence of sequence features which might induce frame-shifting (not shown) and the overlap of the 43 reading frame with the C-terminal region of gene 42. Invertible tail fiber region BcepMu gene 52 encodes a 786 residue protein with limited homologies to the P1 S, P2 H and l Stf tail fiber proteins. Most of the limited homology is defined by four 18 amino acid residue repeats spread throughout the N-terminal half of the gene (Figure 5B). The repeats are separated by multiple Pro residues and at least two Gly-Xaa-Xaa-GlyXaa-Xaa motifs, noted previously as ‘collagen-like’ features of other tail fiber proteins, which are thought to trimerize in extended confirmation to constitute each fiber.45 This suggests that the N-terminal domain of the BcepMu tail fiber protein evolved by internal gene duplication (Figure 5B). Tandem duplications were also found in the N-terminal half of the tail fiber gene of the defective P1-like prophage p15B.46 A well-documented feature of Mu is the invertible G region, which, in its two orientations,

61

Burkholderia cenocepacia Phage BcepMu

allows for the expression of two types of tail fibers.47 While BcepMu lacks an invertible region, SalMu CT18 appears to encode, in addition to a complete tail fiber gene, four alternative tail fiber termini and a homolog of the Gin invertase (Figure 5C). SalMu gene 48 encodes a 342 residue product, the N-terminal portion of which contains the domain repeated multiple times in the much longer BcepMu tail fiber gp52 (Figure 5B). The distal half of gene 48 has four paralogs within SalMu CT18, designated genes 49, 51, 52 and 53 (and three within SalMu Ty2: genes 49, 51 and 52); these paralogs are nearly identical except for an internal hyper-variable region (not show). An analogous multiplex inversion system was found in the defective prophage p15B of E. coli.46 SalMu has a homolog of the Mu Gin and P1 Pin sitespecific recombinases (gene 53 in SalMu-Ty2 and gene 54 in SalMu-CT18), and nucleotide sequences similar to the Gin and Pin recombination sites can be found within the complete tail fiber gene and at the 50 end of each of the paralog sequences (Figure 5C). At least two site-specific recombination events, presumably mediated by the invertase homolog, have occurred subsequent to formation of the SalMu prophage, one between inverted sites and one between tandem sites in SalMu-CT18, leading to the inversion of genes 49-51 and the deletion of gene 52 in SalMu Ty2 (Figure 5C). Thus, as in the case of Mu, the site-specific recombination event can occur in the prophage.47 A puzzling feature of the inversion system in SalMu is the presence of gene 50, encoding a homolog of the l Tfa protein, thought to constitute part of the distal portion of the tail fiber.48 – 50 Unless there is a cryptic promoter associated with gene 50, it would be expressed only in the orientation found in SalMuCT18, in contrast to the invertible regions of Mu and P1, where there are corresponding U and U0 genes cognate for each orientation of the major tail fiber genes (S and S-S0 ). Even if there is a cryptic promoter ensuring expression of gene 50 in both orientations, the Tfa homolog would have to serve all five possible versions of the major tail fiber protein. Another possibility is that like in T4, rather than being part of the mature fiber, the SalMu Tfa homologue functions as a chaparone for tail fiber assembly.48 If this were true then cellular chaparones might replace the Tfa activity for some of the tail fibers.

resulting in Mu virion particles containing 200 and 2000 bp of variable host DNA at the left and right end of the 36,717 bp genome, respectively. Bacteriophages that use transposition as a replication mechanism have been described only infrequently. For E. coli, only two functional Mulike bacteriophages have been isolated, Mu and closely related D108.52 The striking exception is Pseudomonas, for which over 60 transposable bacteriophages have been reported.53 Heteroduplex mapping and recombination frequencies indicate that these belong predominantly to two subgroups with type members D3112 and B3. The genomic sequence of D3112 was recently reported and is particularly remarkable for having replicative genes related to those of Mu and structural genes related to those of lambdoid phages, consistent with its flexible tailed morphology.54 Although Mu-like prophage elements have been identified in the genomes of several Gram-negative bacterial genomes, notably Haemophilus influenzae (FluMu) Neisseria meningitidis (Pnm1) and Deinococcus radiodurans R1 (RadMu)20 and Shewanella oneidensis (MuSo1 and MuSo2),55 none of these have been shown to function as viable phages. In a survey of prophages in sequenced bacterial genomes, Casjens identified several more additional, mostly highly deleted, Mu-like elements.15 These include SalMu CT18 ( ¼ Sti3) described here, and several vestigial elements in the Vibrio cholerae N16961 and Yersinia pestis genomes. While l-like and P2-like prophage elements are widespread, Murelated prophage elements are rare. This indicates that either transposable phages are not widespread, or that such bacteriophages are sufficiently divergent to be unrecognizable by sequence comparison. The recognition that BcepMu is a transposable bacteriophage was initially based on structural features of the genomic sequence assembly, specifically the presence of variable host DNA on either end of the bacteriophage genome, rather than sequence homologies. This suggests that even more classes of transposable phages are present but unrecognized in the sequenced bacterial genomes.

Implications for the distribution of transposable bacteriophages

The complex medium TN broth (TNB)56 was used for liquid cultures of Burkholderia strains. Solid medium was identical except it lacked KNO3 and was supplemented with Difco agar (20 g/l for plates, 7 g/l for top agar). E. coli strains were grown on LB medium, supplemented with ampicillin (Amp) at 0.1 mg/ml, where indicated. All bacterial growth was at 37 8C unless otherwise noted.

Mu and related phages are unique in using transposition as a replication mechanism. Genome amplification during a lytic infection proceeds by replicative transposition, resulting in 50 to 200 copies of Mu integrated nearly randomly throughout the E. coli genome.51 Mu genome packaging starts at a pac site at the left end of the bacteriophage genome and cleavage occurs 200 bp upstream in the host DNA. Packaging terminates about 2000 bp into host DNA from the right end

Materials and Methods Bacterial growth conditions and media

Bacteriophage titration, production and host range testing To determine the plaque-forming unit (PFU) titer, 0.1 ml of a BcepMu lysate sterilized by filtration or CHCl3 and serially diluted in TNB was mixed with

62 0.1 ml of a logarithmic culture (5 £ 108 colony forming units (CFU)/ ml) of B. cenocepacia strain K56-2 grown in TNB at 28 8C, mixed into 5 ml of molten TN top agar containing 1 mM MgSO4, poured on a TN agar plate and incubated overnight at 28 8C. High titer lysates (,1010/ml) were prepared by harvesting TN top agar of plates exhibiting confluent lysis in 5 ml of TN broth, macerating the agar, clearing the lysate by centrifugation (10,000g) and sterilizing through a 0.22 mm filter (Acrodisk). The host range was assessed by spotting 10 ml of serial dilutions of a 1010 PFU/ml stock on lawns of various hosts and incubating overnight. Bacteriophage production was determined in the same manner for selected B. cenocepacia strains of the ET-12 lineage. The eclipse and latent periods, and the burst size for bacteriophage BcepMu were determined by conducting a one-step growth experiment, using K56-2 cells growing exponentially (A600 , 0.3) in TNB at 28 8C for infections and as the plating indicator strain. To initiate the one-step growth experiment, BcepMu was added to K56-2 cells at an input multiplicity of infection (MOI) of 1023. After 15 minutes at 28 8C, the infected cells were diluted 1 : 1000. One ml samples were taken every 15 – 30 minutes for 3.5 hours and a 100 ml sample plated directly in top agar with indicator cells. After addition of chloroform (100 ml) to the remaining 900 ml, the sample was vortexed briefly, centrifuged for one minute and a 100 ml volume removed for plating. Plaques were counted after overnight growth. Library preparation and shotgun sequencing All manipulations were performed according to the instructions of the various manufacturers of the reagents described. Briefly, DNA was isolated from fresh bacteriophage lysate with the Wizard Lambda DNA Isolation Kit (Promega). Bacteriophage DNA was fractionated utilizing the Gene Machines Hydroshear device (Gene Machines) and end repaired using the DNATerminator End Repair Kit (Lucigen). Phage genomic DNA fragments were gel-purified and ligated into the pSmart HC vector (Lucigen). The ligation reaction was transformed by electroporation into “E. cloni” 10G electro-competent cells (Lucigen) and transformants selected on LB agar containing carbenicillin (100 mg/ml). Plasmids were isolated from 384 randomly picked transformants using a Beckman Biomek Robotic Workstation and the Wizard Magnesil Plasmid Purification System (Promega). Each plasmid was sequenced using AmpL1 and AmpR1 primers (Lucigen) and the ABI BigDye Terminator Cycle Sequencing Ready Reaction Kit v2.0 (Applied Biosystems) in 96-well plates in a PE2700 (Applied Biosystems). Unincorporated BigDye was removed by alcohol precipitation; pellets were resuspended in 10 ml of HiDi formamide (Applied Biosystems) and resolved on an ABI3100 capillary sequencer (Laboratory for Plant Genome Technology, Texas A&M University).

Burkholderia cenocepacia Phage BcepMu

dicted proteins were then compared to the NCBI protein database with Blastp at the mirror site located at XBlast‡. Structural features of the proteins were determined with proteomic tools at ExPASy§. Genome comparisons were performed using GenomePixelizerk. DNA pair-wise comparisons were performed with the EMBOSS Dotmatcher program{. Phage genome maps were drawn utilizing the program DNA Mastera. Comparisons of BcepMu sequences with the J2315 host genome were performed using the BLAST server at the B. cenocepacia sequencing projectb. The complete DNA sequence of bacteriophage BcepMu can be found in GenBank under accession number AY539836. IST library construction and analysis The IST library of phage BcepMu was produced and analyzed as described for l and yeast IST libraries.22,23 Briefly, BcepMu DNA was fragmented randomly by hydro-shearing, end-repaired, and ligated into SmaIdigested pLM100 (GenBank accession no. AF308740). The pLM100 vector encodes, under a weakly constitutive promoter, the N-terminal DNA binding domain but not the C-terminal dimerization domain of the l cI repressor. The unique SmaI site of this plasmid is near the end of a fragment of the l cI. DNA fragments, which encode a polypeptide sequence capable of supporting homotypic interactions (dimerization or higher oligomerization), confer l-immunity on cells carrying the plasmid. The ligation mixture was transformed into electro-competent JH787 (AG1688 [F128 lacIq lacZ < Tn5/araD139 D(ara-leu)7697 D(lac)X74 galE15 galK16 rpsL hsdR2 mcrA mcrB1 (f80 su-3))58 and selected for immunity to l by plating on LB agar containing ampicillin (40 mg/ml) and seeded with two l phages (KH54 and KH54h80; each at 108 PFU/plate) deleted for the cI gene, as described.22 Plasmid DNA was purified from immune colonies arrayed into 96-well culture plates and sequenced across the entire length of the inserted DNA. Insert sequences in frame with cI were extracted and compared to the predicted BcepMu protein sequencesc. BcepMu-specific PCR assay Preparation of bacterial genomic DNA was performed as described.5 Oligonucleotide primers F8.3 (50 -TGTTCA GAGATGCGTTCGAC-30 ) and R9.2 (50 -ATGGCGCTTGA CAGGTAATC-30 ) were used in a PCR assay to determine the distribution of BcepMu within species of the B. cepacia complex. PCR was performed using the Taq PCR Core Kit (Qiagen) in 50 ml reactions containing 20 ng of template DNA and 25 pmol of each primer using amplification conditions of one cycle at 95 8C for one minute, 25 cycles at 95 8C for 30 seconds, 50 8C for 30 seconds, 72 8C for one minute, and one cycle at 72 8C for five minutes. PCR products were analyzed by agarose gel electrophoresis.

Sequence assembly and analysis The program Sequencher (Gene Codes Corporation) was used for sequence assembly. Areas of low quality sequence were re-sequenced using primer walking. The final consensus sequence was analyzed for the presence of protein coding regions using GeneMark†.57 The pre† http://opal.biology.gatech.edu/GeneMark/

‡ http://xblast.tamu.edu/pise/ § http://us.expasy.org k http://www.atgc.org/GenomePixelizer/ { http://www.sarsresearch.ca/index. php?page ¼ toolssite a http://cobamide2.bio.pitt.edu/computer.htm b http://www.sanger.ac.uk/Projects/B_cenocepacia/ c Protocols can be found at http://tofu.tamu.edu/ doodle/protocols

63

Burkholderia cenocepacia Phage BcepMu

Transmission electron microscopy (TEM) TEM was performed using high titer lysates, with 1010 PFU/ml spotted onto 400 mesh carbon-coated copper grids and negatively stained with 2%(w/v) uranyl acetate. The samples were visualized with a JEOL 1200 EX.

6.

7.

Acknowledgements Support for this work was provided primarily from a grant from the National Science Foundation, MCB-0135653, to establish a research and instruction program in phage genomics for undergraduate students. The authors T.C., L.M. and A.C. were participants in the program and conducted all the genomic sequencing and the primary annotation in partial fulfillment of the requirements of this program. This work was also supported, in part, by grant LIPUMA00A0 from the Cystic Fibrosis Foundation (to J.J.L. and C.F.G.). The assistance of Jim Hu and Leonardo Marin˜o-Ramirez in generating the IST library was essential. We are grateful for sequencing and robotics facilities provided to this program through the cooperation of John Mullet, Director of the Center for Plant Genomics and Biotechnology and Robert Klein, Southern Plains Agricultural Research Center, USDA-ARS. Other support for this work was derived from funding provided by the US Army Medical Research and Material Command Disaster Relief and Emergency Medical Services Program (DREAMS), the Texas Agricultural Experiment Station, and PHS grant 27099 (to R.Y.). The FE-SEM acquisition was supported by the National Science Foundation under grant number DBI-0116835.

8.

9.

10.

11.

12.

13.

14.

15.

References 1. Coenye, T., Vandamme, P., Govan, J. R. & LiPuma, J. J. (2001). Taxonomy and identification of the Burkholderia cepacia complex. J. Clin. Microbiol. 39, 3427–3436. 2. Vandamme, P., Henry, D., Coenye, T., Nzula, S., Vancanneyt, M., LiPuma, J. J. et al. (2002). Burkholderia anthina sp. nov. and Burkholderia pyrrocinia, two additional Burkholderia cepacia complex bacteria, may confound results of new molecular diagnostic tools. FEMS Immunol. Med. Microbiol. 33, 143–149. 3. LiPuma, J. J., Spilker, T., Gill, L. H., Campbell, P. W., III, Liu, L. & Mahenthiralingam, E. (2001). Disproportionate distribution of Burkholderia cepacia complex species and transmissibility markers in cystic fibrosis. Am. J. Respir. Crit. Care Med. 164, 92– 96. 4. Vandamme, P., Holmes, B., Coenye, T., Goris, J., Mahenthiralingam, E., LiPuma, J. J. & Govan, J. R. (2003). Burkholderia cenocepacia sp. nov.—a new twist to an old story. Res. Microbiol. 154, 91 – 96. 5. Coenye, T. & LiPuma, J. J. (2002). Multilocus restric-

16.

17.

18.

19. 20.

tion typing: a novel tool for studying global epidemiology of Burkholderia cepacia complex infection in cystic fibrosis. J. Infect. Dis. 185, 1454– 1462. Johnson, W. M., Tyler, S. D. & Rozee, K. R. (1994). Linkage analysis of geographic and clinical clusters in Pseudomonas cepacia infections by multilocus enzyme electrophoresis and ribotyping. J. Clin. Microbiol. 32, 924–930. Ledson, M. J., Gallagher, M. J., Jackson, M., Hart, C. A. & Walshaw, M. J. (2002). Outcome of Burkholderia cepacia colonisation in an adult cystic fibrosis centre. Thorax, 57, 142– 145. Chen, J. S., Witzmann, K. A., Spilker, T., Fink, R. J. & LiPuma, J. J. (2001). Endemicity and inter-city spread of Burkholderia cepacia genomovar III in cystic fibrosis. J. Pediatr. 139, 643–649. LiPuma, J. J., Mortensen, J. E., Dasen, S. E., Edlind, T. D., Schidlow, D. V., Burns, J. L. & Stull, T. L. (1988). Ribotype analysis of Pseudomonas cepacia from cystic fibrosis treatment centers. J. Pediatr. 113, 859 –862. Mahenthiralingam, E., Coenye, T., Chung, J. W., Speert, D. P., Govan, J. R., Taylor, P. & Vandamme, P. (2000). Diagnostically and experimentally useful panel of strains from the Burkholderia cepacia complex. J. Clin. Microbiol. 38, 910– 913. Coenye, T. & LiPuma, J. J. (2003). Population structure analysis of Burkholderia cepacia genomovar III: varying degrees of genetic recombination characterize major clonal complexes. Microbiology, 149, 77 –88. Coenye, T., Spilker, T., Martin, A. & LiPuma, J. J. (2002). Comparative assessment of genotyping methods for epidemiologic study of Burkholderia cepacia genomovar III. J. Clin. Microbiol. 40, 3300 –3307. Deng, W., Liou, S. R., Plunkett, G., III, Mayhew, G. F., Rose, D. J., Burland, V. et al. (2003). Comparative genomics of Salmonella enterica serovar Typhi strains Ty2 and CT18. J. Bacteriol. 185, 2330– 2337. Parkhill, J., Dougan, G., James, K. D., Thomson, N. R., Pickard, D., Wain, J. et al. (2001). Complete genome sequence of a multiple drug resistant Salmonella enterica serovar Typhi CT18. Nature, 413, 848– 852. Casjens, S. (2003). Prophages and bacterial genomics: what have we learned so far? Mol. Microbiol. 49, 277 –300. Duchaud, E., Rusniok, C., Frangeul, L., Buchrieser, C., Givaudan, A., Taourit, S. et al. (2003). The genome sequence of the entomopathogenic bacterium Photorhabdus luminescens. Nature Biotechnol. 21, 1307 –1313. Brazilian National Genome Project Consortium (2003). The complete genome sequence of Chromobacterium violaceum reveals remarkable and exploitable bacterial adaptability. Proc. Natl Acad. Sci. USA, 100, 11660– 11665. Juhala, R. J., Ford, M. E., Duda, R. L., Youlton, A., Hatfull, G. F. & Hendrix, R. W. (2000). Genomic sequences of bacteriophages HK97 and HK022: pervasive genetic mosaicism in the lambdoid bacteriophages. J. Mol. Biol. 299, 27 – 51. Hendrix, R. W., Lawrence, J. G., Hatfull, G. F. & Casjens, S. (2000). The origins and ongoing evolution of viruses. Trends Microbiol. 8, 504– 508. Morgan, G. J., Hatfull, G. F., Casjens, S. & Hendrix, R. W. (2002). Bacteriophage Mu genome sequence: analysis and comparison with Mu-like prophages in Haemophilus, Neisseria and Deinococcus. J. Mol. Biol. 317, 337– 359.

64

21. McClelland, M., Sanderson, K. E., Spieth, J., Clifton, S. W., Latreille, P., Courtney, L. et al. (2001). Complete genome sequence of Salmonella enterica serovar Typhimurium LT2. Nature, 413, 852– 856. 22. Marin˜o-Ramirez, L. (2001). Using lambda repressor fusions to isolate and characterize self-assembling domains. In Protein– Protein Interactions: A Laboratory Manual (Golemis, E., ed.), Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 23. Marin˜o-Ramirez, L. & Hu, J. C. (2002). Isolation and mapping of self-assembling protein domains encoded by the Saccharomyces cerevisiae genome using lambda repressor fusions. Yeast, 19, 641–650. 24. Ast, V. M., Schoenhofen, I. C., Langen, G. R., Stratilo, C. W., Chamberlain, M. D. & Howard, S. P. (2002). Expression of the ExeAB complex of Aeromonas hydrophila is required for the localization and assembly of the ExeD secretion port multimer. Mol. Microbiol. 44, 217– 231. 25. Hara, O. & Hutchinson, C. R. (1992). A macrolide 3-O-acyltransferase gene from the midecamycin-producing species Streptomyces mycarofaciens. J. Bacteriol. 174, 5141– 5144. 26. Slauch, J. M., Lee, A. A., Mahan, M. J. & Mekalanos, J. J. (1996). Molecular characterization of the oafA locus responsible for acetylation of Salmonella typhimurium O-antigen: oafA is a member of a family of integral membrane trans-acylases. J. Bacteriol. 178, 5904– 5909. 27. Marchler-Bauer, A., Anderson, J. B., DeWeese-Scott, C., Fedorova, N. D., Geer, L. Y., He, S. et al. (2003). CDD: a curated Entrez database of conserved domain alignments. Nucl. Acids Res. 31, 383– 387. 28. Leschziner, A. E., Griffin, T. J. & Grindley, N. D. (1998). Tn552 transposase catalyzes concerted strand transfer in vitro. Proc. Natl Acad. Sci. USA, 95, 7345– 7350. 29. Rowland, S. J. & Dyke, K. G. (1989). Characterization of the staphylococcal beta-lactamase transposon Tn552. EMBO J. 8, 2761– 2773. 30. Rowland, S. J., Sherratt, D. J., Stark, W. M. & Boocock, M. R. (1995). Tn552 transposase purification and in vitro activities. EMBO J. 14, 196– 205. 31. Wang, I.-N., Smith, D. L. & Young, R. (2000). Holins: the protein clocks of bacteriophage infections. Annu. Rev. Microbiol. 54, 799– 825. 32. Holtje, J. V. & Heidrich, C. (2001). Enzymology of elongation and constriction of the murein sacculus of Escherichia coli. Biochimie, 83, 103– 108. 33. Holtje, J. V. (1996). Lytic transglycosylases. EXS, 75, 425– 429. 34. Xu, M., Struck, D. K., Deaton, J. F., Wang, I.-N. & Young, R. (2004). A new type of signal, the signal arrest-release (SAR) sequence, mediates export and control of the phage P1 endolysin. Proc. Natl Acad. Sci. USA,. 35. Kedzierska, S., Wawrzynow, A. & Taylor, A. (1996). The Rz1 gene product of bacteriophage lambda is a lipoprotein localized in the outer membrane of Escherichia coli. Gene, 168, 1– 8. 36. Young, R., Way, S., Yin, J. & Syvanen, M. (1979). Transposition mutagenesis of bacteriophage lambda: a new gene affecting cell lysis. J. Mol. Biol. 132, 307– 322. 37. Zhang, N. & Young, R. (1999). Complementation and characterization of the nested Rz and Rz1 reading frames in the genome of bacteriophage lambda. Mol. Gen. Genet. 262, 659– 667. 38. Nakayama, K., Kanaya, S., Ohnishi, M., Terawaki, Y.

Burkholderia cenocepacia Phage BcepMu

39.

40.

41. 42.

43.

44.

45. 46.

47. 48.

49. 50.

51. 52.

53.

54.

55.

& Hayashi, T. (1999). The complete nucleotide sequence of phi CTX, a cytotoxin-converting phage of Pseudomonas aeruginosa: implications for phage evolution and horizontal gene transfer via bacteriophages. Mol. Microbiol. 31, 399– 419. Oakey, H. J., Cullen, B. R. & Owens, L. (2002). The complete nucleotide sequence of the Vibrio harveyi bacteriophage VHML. J. Appl. Microbiol. 93, 1089– 1098. Temple, L. M., Forsburg, S. L., Calendar, R. & Christie, G. E. (1991). Nucleotide sequence of the genes encoding the major tail sheath and tail tube proteins of bacteriophage P2. Virology, 181, 353– 358. Xue, Q. & Egan, J. B. (1995). Tail sheath and tail tube genes of the temperate coliphage 186. Virology, 212, 218– 221. Xue, Q. & Egan, J. B. (1995). DNA sequence of tail fiber genes of coliphage 186 and evidence for a common ancestor shared by dsDNA phage fiber genes. Virology, 212, 128– 133. Christie, G. E., Temple, L. M., Bartlett, B. A. & Goodwin, T. S. (2002). Programmed translational frameshift in the bacteriophage P2 FETUD tail gene operon. J. Bacteriol. 184, 6522– 6531. Levin, M. E., Hendrix, R. W. & Casjens, S. R. (1993). A programmed translational frameshift is required for the synthesis of a bacteriophage lambda tail assembly protein. J. Mol. Biol. 234, 124– 139. Smith, M. C., Burns, N., Sayers, J. R., Sorrell, J. A., Casjens, S. R. & Hendrix, R. W. (1998). Bacteriophage collagen. Science, 279, 1834. Sandmeier, H., Iida, S., Huebner, P., Hiestand-Nauer, R. & Arber, W. (1991). Gene organization in the multiple DNA inversion region Min of plasmid p15B of E. coli 15T(2 ): assemblage of a variable gene. Nucl. Acids Res. 19, 5831–5838. Plasterk, R. H., Ilmer, T. A. & Van de, P. P. (1983). Site-specific recombination by Gin of bacteriophage Mu: inversions and deletions. Virology, 127, 24 – 36. Hashemolhosseini, S., Stierhof, Y. D., Hindennach, I. & Henning, U. (1996). Characterization of the helper proteins for the assembly of tail fibers of coliphages T4 and lambda. J. Bacteriol. 178, 6258–6265. Hendrix, R. W. & Duda, R. L. (1992). Bacteriophage lambda PaPa: not the mother of all lambda phages. Science, 258, 1145– 1148. Montag, D., Schwarz, H. & Henning, U. (1989). A component of the side tail fiber of Escherichia coli bacteriophage lambda can functionally replace the receptor-recognizing part of a long tail fiber protein of the unrelated bacteriophage T4. J. Bacteriol. 171, 4378– 4384. Symonds, N., Toussaint, A., Van De Putte, P. & Howe, M. M. (1987). Phage Mu, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Gill, G. S., Hull, R. C. & Curtiss, R., III (1981). Mutator bacteriophage D108 and its DNA: an electron microscopic characterization. J. Virol. 37, 420– 430. Akhveridian, V. Z., Khrenova, E. A., Bogush, V. G., Gerasimova, T. V., Kirsanov, N. B. & Krylov, V. N. (1984). Wide distribution of transposable phages in natural populations of Pseudomonas aeruginosa. Genetika, 20, 1612 –1619. Wang, P. W., Chu, L. & Guttman, D. S. (2004). Complete sequence and evolutionary genomic analysis of the Pseudomonas aeruginosa transposable bacteriophage D3112. J. Bacteriol. 186, 400– 410. Heidelberg, J. F., Paulsen, I. T., Nelson, K. E., Gaidos,

Burkholderia cenocepacia Phage BcepMu

E. J., Nelson, W. C., Read, T. D. et al. (2002). Genome sequence of the dissimilatory metal ion-reducing bacterium Shewanella oneidensis. Nature Biotechnol. 20, 1118 – 1123. 56. Hansen, J. B. & Olsen, R. H. (1978). Isolation of large bacterial plasmids and characterization of the P2 incompatibility group plasmids pMG1 and pMG5. J. Bacteriol. 135, 227– 238.

65

57. Besemer, J. & Borodovsky, M. (1999). Heuristic approach to deriving models for gene finding. Nucl. Acids Res. 27, 3911 –3920. 58. Hu, J. C., Newell, N. E., Tidor, B. & Sauer, R. T. (1993). Probing the roles of residues at the e and g positions of the GCN4 leucine zipper by combinatorial mutagenesis. Protein Sci. 2, 1072 –1084.

Edited by M. F. Summers (Received 4 February 2004; received in revised form 5 April 2004; accepted 6 April 2004)