The study reported here describes the genomic and proteomic analysis of a bacteriophage isolated from a virulent field strain of H. parasuis. The genome size of the H. parasuis bacteriophage (SuMu) is 37,151 bp and is comparable to the genome sizes of enterobacteriophage Mu of E. coli K-12 (36,717 bp), H. influenzae bacterio-phage FluMu (34,676 bp), and the Neisseria meningitidis Mu-like bacteriophage Pmn1 (39,236 bp) [10, 17] .
Bacteriophage SuMu’s morphology is similar to amber mutants of enterobacteriophage Mu described by Grundy and Howe . An icosahedral head and tail sheath were seen in bacteriophage SuMu by electron microscopy. Our lysates were stored at 4°C for over a week before electron microscopy. Few intact virions for this bacteriophage were found probably due to their disintegration upon storage. These findings are in agreement with those of Summer et al. , who found that phage 56 (BcepMu) particles were unstable in lysates, had decreasing titers with storage, and showed disintegrating particles with broken heads, and partially exposed tails in electron micrographs.
Although bands were observed after cesium chloride centrifugation at the refractive indices (η) reported by Grundy and Howe , no bacteriophages were detected by electron microscopy. Our SM broth had 50% less concentrations of Tris–HCl and magnesium than the final buffer used by Grundy and Howe . We also removed the cesium chloride by ultrafiltration rather than by dialysis. These methods may have also affected the stability of the bacteriophages. However, our electron micrographs clearly showed bacteriophage particles attached to the bacteria in Figure 1 and Figure 1B.
The G + C content of the H. parasuis bacteriophage SuMu at 41.87% is similar to the G + C content of the virulent H. parasuis (39.93%). This close match is also seen in the G + C contents of the Mu-like bacteriophage Pmn1 (53.1%) and its host genome N. meningitidis type A strain Z2491 (51.8%), and enterobacteriophage Mu (52.05%) and the E. coli K-12 genome (50.8%) [10, 20]. In contrast, the G + C content of H. influenzae Mu-like bacteriophage FluMu is approximately 50%, compared to 38% G + C of the H. influenzae Rd chromosome .
A comparison of the bacteriophage SuMu translated nucleotide sequence obtained from its cloned DNA to the enterobacteriophage Mu translated nucleotide sequence using the Artemis Comparison Tool (ACT) revealed that the conserved genes generally were related to lysogeny and lysis, integration and transposition, morphogenesis, and tail proteins. Interestingly, similar bacteriophage sequences were also present in two published H. parasuis genomes. Thirty nine of the 55 bacteriophage SuMu genes were homologous to DNA sequences from H. parasuis strain 29755 (GenBank NZ_ABKM00000000, ctgs_1000001-1000246) or H. parasuis SH0165 . Bacteriophage genes were also reported as upregulated by Melnikow et al.  using a microarray analysis of the H. parasuis genome after growing H. parasuis strain 29755 under iron-limiting, oxygen-limiting, heat, and acidic conditions. The highest gene homologs found were bacteriophage transposase (DQ127936; 99% identity over 1,043 bp; E-value = 0.0), bacteriophage Mu T protein homolog (DQ127950; 96% identity over 924 bp; E-value = 0.0), and bacteriophage Mu I protein GP32 (DQ127939; 97% identity over 1,077 bp; E-value = 0.0).
The partial hemolysin gene found upstream of the SuMu putative bacteriophage repressor gene (CDS 353..1072 c) was 99.3% similar (over 302 bp) to hhdB, which encodes a hemolysin activation/secretion protein (NC_118252, 1417483..1419203) of H. parasuis SH0165, serovar 5. Others have reported a putative hemolysin gene operon, hhdBA, in pigs infected with H. parasuis[23, 24]. The bacteriophage SuMu partial hemolysin gene was probably a remnant of the H. parasuis 34086b host chromosome, which could have been acquired by an illegitimate recombination event.
However, the partial hemolysin gene homolog was not found in the H. parasuis draft 29755, serovar 5 genome (Genbank NZ_ABKM00000000, ctgs 1000001–10000246). Given the draft nature of the H. parasuis 29755 genome, it is possible that assembly breaks, which are not uncommon at bacteriophage and repetitive sequences, may have prevented the annotation of a hemolysin gene if one is present.
The bacteriophage SuMu protease contained a significant deletion (254 amino acids) compared to the Mu I bacteriophage protease. The other two initial inserts of cloned DNA had deletions of 127 amino acids when they were compared to the Mu gp29 sequence.
As has been reported in the literature [10, 25–27], a bacteriophage may contribute to the virulence of its host. Avian pathogenic E. coli causes respiratory infection and septicemia of poultry, involving bacteriophage-related sequences coding for proteins such as DNA stabilization, portal, and integrase proteins . Bacteriophage SuMu gp29 and gp36 are a homologs of enterobacteriophage Mu gp29 and gp36, found in H. influenzae Rd . SuMu gp29 and gp36 are also homologous to two genomic fragments identified by Townsend et al.  which hybridize to virulent hemorrhagic septicemia isolates of P. multocida, but not to other P. multocida isolates. Translated DNA of one fragment (clone 6b) of P. multocida identified by Townsend et al.  had 62% homology  over 87 AA to SuMu_34, similar to gp36 of enterobacteriophage Mu (E-value = 1e-17) by blastp analysis . The other translated DNA fragment (clone A3b)  had 72% homology over 83 AA to SuMu_27, similar to enterobacteriophage Mu gp29 (E-value = 4e-31). SuMu proteins may also be associated with virulence in swine infected with H. parasuis carrying the SuMu bacteriophage .
Canchaya et al.  have reviewed work on lysogenic conversion. Some bacteriophages carry DNA that can alter the phenotype of the bacterial host (lysogenic conversion genes) and this lysogeny is a form of short-term bacterial evolution. When lysogenic Gram-negative bacteria were grown in animals, the bacteriophage-specific genes were upregulated in the bacteria . This study described Mu-like genes that mirrored those found in enterobacteriophage Mu and MuMenB, a bacteriophage of N. meningitidis strain MC58 . Corresponding Mu-like genes found in Mu, MuMenB and SuMu (reported here) included Mu A, Mu B, gp29, gp30, Mu G, Mu I, gp36, gp37, gpL, Mu M, gp42, gp45, gp46, and gp47. Two strains of H. parasuis, strain 29755  and strain SH0165 , whose genomes have been sequenced, also carry Mu-like bacteriophages. The Mu-like bacteriophage GenBank NZ_ABMK00000000, ctg_1000002 sequence of strain 29755 is similar to that of the SuMu DNA sequence. The tail assembly protein of N. meningitidis (Mu G) has been postulated to be membrane-associated. In these studies , the authors hypothesized that bacteriophage-encoded membrane-associated proteins of H. influenzae and N. meningitidis contributed to the variability of the bacterial envelope structure and may therefore influence the virulence and pathogenicity of the organisms.
DNA sequencing revealed a peptidoglycan recognition protein (PGRP) which hydrolyzes peptidoglycans of bacterial cell walls at the analogous position of the lysis protein of enterobacteriophage Mu. Mass spectrometry also suggested a lysozyme/muramidase/endolysin.
The mass spectrometry analysis of the proteins from the 2-D gel suggested that H. parasuis strain 34086b may contain more than one bacteriophage as was shown by others. For example, H. parasuis strain SH0165  also carries Lambda, P2, and CP-933 K bacteriophage genes while strain 29755  has remnants of Lambda, CP4-57-like, P4-like, phi-C3, CP-933 C, and Lj965 bacteriophage genes. However, the SuMu genome sequence was based on cloned DNA and contained only one bacteriophage.
Bacteriophage SuMu was putatively distantly related to the enterobacteriophage Mu by combined DNA and proteomic evidence. Only 17 out of 54 putative proteins were more than 50% homologous between enterobacteriophage Mu and bacteriophage SuMu. The percent homology and E-values associated with them when comparing their translated nucleotides and their proteins was low compared to similar comparisons of bacteriophage SuMu to other Mu-like bacteriophages. These results are in agreement with the evolutionary diversity of bacteriophages [32, 33]. Although the double-stranded DNA tailed-bacteriophages are extremely diverse, they share common modules, such as terminase and portal proteins as well as tail proteins . Either by illegitimate recombination or by integrase-mediated site specific recombination, a structural gene operon can be changed so it is not homologous to other genes in related bacteriophages . The presence of variable terminal sequences in bacteriophage SuMu supports the classification of it as a Mu-like bacteriophage. The rearrangement of SuMu prophage sequences in H. parasuis 34086b is corroborated by a report of CampMu prophages mediating genomic rearrangements in Campylobacter jejuni. Since bacteriophage SuMu’s G + C content is not much different than its host’s chromosome G + C content, it can be concluded that SuMu has been associated with H. parasuis for a relatively long length of time .
Based on the results of others [27, 30, 35], it is suggested that some virulent bacteria harbor bacteriophages while many avirulent organisms do not have genome-associated bacteriophage. Our recent study  showed the distribution of SuMu’s Mu-like portal bacteriophage gene, gp29, among 15 reference strains and 31 field isolates of H. parasuis. The gene was present in most of the virulent isolates and absent in most of the avirulent isolates. This nested PCR assay detected 28 of 31 field isolates designated as virulent and five of six reference strains designated as virulent by utilizing the gp29 gene of bacteriophage SuMu.
Since the bacteriophage SuMu genes and proteins identified are related to bacteriophage Mu, they could potentially confer the necessary factors for this bacteriophage to be able to transduce virulence factors, which could affect the epidemiology of H. parasuis field isolates [9, 10]. A potential diagnostic test such as a bacteriophage-specific PCR test was developed to assay for the presence of bacteriophage genes in bacterial isolates from H. parasuis-infected animals .