Identification, sequencing and molecular analysis of Chp4, a novel chlamydiaphage of Chlamydophila abortus belonging to the family Microviridae

The Chlamydiaceae are Gram-negative obligate intracellular bacteria that cause a variety of diseases in a wide range of host species (Longbottom & Coulter, 2003). The family is currently subdivided into two distinct genera, Chlamydia and Chlamydophila. The obligate intracellular lifestyle of these pathogens as well as the fact that they exist in two distinct developmental forms, the infectious extracellular elementary body and the non-infectious intracellular metabolically active reticulate body, has hindered the development of a host-free culture system or a stable gene-transfer system for studying detailed mechanisms of pathogenesis and gene regulation. Over the years, research groups have attempted the direct transformation of chlamydiae using plasmid-based shuttle vectors in combination with chemical treatment or electroporation with no success. Thus, the identification of a bacteriophage that was found to infect chlamydial organisms held promise as a natural transformation system. This bacteriophage, named Chlamydiaphage 1 or Chp1, was first revealed by thin-section transmission electron microscopy of Chlamydophila psittaci (Richmond et al., 1982). Since that initial finding, a further four chlamydiaphages have been identified: Chp2 from Chlamydophila abortus (Liu et al., 2000; Everson et al., 2002), Chp3 from Chlamydophila pecorum (Garner et al., 2004), φCPG1 from Chlamydophila caviae (Hsia et al., 2000a, b) and CPAR39 from Chlamydophila pneumoniae (Read et al., 2000a). In this report, we describe the identification of a new and novel chlamydiaphage infecting C. abortus.

Chlamydiaphage are icosahedral ssDNA phage of the family Microviridae. The family Microviridae consists of two subfamilies, the first containing the well-known and extensively characterized φX174 coliphage, as well as several other phage infecting Enterobacteriaceae species. The chlamydiaphage, owing to the absence of genes encoding major spike and external scaffold proteins, belong to the second subfamily, the Gokushovirinae, other members of which include phage φMH2K, isolated from the intracellular bacteria Bdellovibrio bacteriovorus, (Brentlinger et al., 2002) and SpV4, isolated from Spiroplasma melliferum (Renaudin et al., 1984). As with other members of the Microviridae, chlamydiaphage are lytic to their hosts. It has been demonstrated that the presence of chlamydiaphage φCPG1 may be important in influencing levels of infectivity and pathology of C. caviae (Rank et al., 2009), and that infection with Chp2 can slow cell division of reticulate bodies (RBs) and their subsequent development into elementary bodies (EBs) (Salim et al., 2008).

To identify the presence of chlamydiaphage Chp4, C. abortus (strain S26/3), originating from stocks prepared in 1988 (yolk sac passage 3) and 1991 (yolk sac passage 5), was propagated in McCoy cells (Graham et al., 1995) and sarkosyl/DTT extracted outer membrane (OM) fractions were produced by a method described by Tan et al. (1990). Approximately 5 µg of the OM fractions were solubilized in SDS sample buffer and protein components separated by SDS-PAGE on a 10 % Tris/glycine gel. SDS-PAGE analysis of the OM fractions originating from the low passage stocks revealed the presence of an additional protein band of approximately 60 kDa in size compared with OM fractions from other stocks of C. abortus (Fig. 1a). The 60 kDa protein band was excised from the gel and the proteins destained and reductively alkylated using DTT and iodoacetamide. Gel pieces were digested overnight with trypsin (Promega) at 37 °C and analysed on a Bruker Ultraflex II MALDI–ToF–ToF mass spectrometer (Bruker Daltonics). Proteins were identified by comparing peptide mass fingerprints against a mascot database () of theoretical peptides derived from known protein sequences in the NCBI databases. MALDI–ToF analysis of the excised band revealed the identity of the protein to be a VP1 chlamydiaphage structural protein. Transmission electron microscopy of negatively stained purified C. abortus EBs (Longbottom et al., 1998) infected with semi-purified Chp4 (Everson et al., 2002) confirmed the presence of bacteriophage particles with an icosahedral capsid adhered to the surface of the EB (Fig. 1b).

Molecular characterization of the chlamydiaphage, named Chp4, was performed following the isolation and purification of single-stranded circular genomic DNA from harvested tissue-culture material 72 h post-infection using a QIAprep spin miniprep kit (Qiagen). Phage DNA was amplified by rolling circle amplification using a TempliPhi 100 amplification kit (GE healthcare) and directly sequenced using primers based on the Chp2 chlamydiaphage sequence (GenBank accession no. NP_054647; Supplementary Table S1, available in JGV Online). Following assembly and annotation, the genome sequence of Chp4 was deposited in GenBank (accession no. AY769964).

The Chp4 genome consists of a circular, ssDNA molecule of 4.53 kbp with a nucleotide composition of 28.7 % adenine (A), 22.5 % guanine (G), 30.1 % thymine (T) and 18.7 % cytosine (C). Eight ORFs, each oriented in the same direction and initiated with an ATG start codon, were identified using Genemark Violin (Mills et al., 2003) (Fig. 2). Based on sequence homology with other characterized bacteriophages, ORFs 1–3 were predicted to encode the putative major coat protein VP1, the putative minor spike protein VP2 and the putative scaffolding protein VP3. ORFs 4 and 5 were predicted to encode non-structural proteins homologous to the φX174 A and C proteins involved in the synthesis of viral DNA and packaging of DNA into the viral capsid (Aoyama et al., 1983; Bernal et al., 2004; Eisenberg & Kornberg, 1979), while ORF8 appears to be homologous to the φX174 J protein that is involved in DNA packaging and, interestingly, might also be involved in viral attachment to host cells (Bernal et al., 2004). No homologues have been predicted for ORFs 6 and 7, which overlap ORFs 2 and 1, respectively, other than in the Chlamydiamicroviridae. Putative ribosome-binding sites were identified upstream of the start codon for each of these ORFs.

Although their function is currently unknown, homologues for each of the Chp4 ORFs have been predicted in other members of the Chlamydiamicroviridae. Indeed the genomes exhibit very similar organization and structure in terms of both gene order and location, particularly between Chp4, Chp3 and Chp2 (Fig. 2). Comparison of the nucleotide sequences of each predicted ORF in Chp4 with those of the other chlamydiaphages using megalign (lasergene; dnastar) sequence alignment software revealed an overall DNA sequence similarity (excluding the VP1 sequence) of 94–95 %. However, when comparing the VP1 sequences alone it was found that the DNA sequence similarity was greater with CPAR39 and φCPG1 (95 %) than with Chp2 or Chp3 (90–91 %). This was also reflected at the amino acid level with the predicted VP1 protein sequence of Chp4 being 95 % similar to CPAR39 and φCPG1, while the protein sequences of Chp2 and Chp3 were 85 and 83 % similar, respectively.

Phylogenetic analysis of the VP1 protein sequences was performed to determine the degree of relatedness between Chp4 and other members of the families Chlamydiamicroviridae and Microviridae. The translated sequences of the major viral coat protein VP1 were aligned using muscle (Supplementary Fig. S1, available in JGV Online) (Edgar, 2004). MrBayes software (Ronquist & Huelsenbeck, 2003), launched from within the TOPALi version 2 package (Milne et al., 2009), was used to generate a Bayesian phylogenetic tree by using Markov-chain Monte Carlo settings of: two runs of 625 000 generations, burn-in 125 000 generations, trees sampled every 100 generations using the Whelan and Goldman (WAG) substitution model (Ronquist & Huelsenbeck, 2003; Milne et al., 2009). Chp4 was found to be more closely related to CPAR39 and φCPG1 than to either Chp2 or Chp3 (Fig. 3a), forming a monophyletic group supported by a posterior clade probability of 1.0. Differences between the chlamydiaphage VP1 sequences were found to occur mainly in the IN5 and INS regions of the VP1 outer coat protein (aa 217–292 and 446–459 of Chp4, respectively; Fig. 3b, c). Computer analysis predicted the INS region to be located in close contact with IN5 on the capsid protein (Read et al., 2000b), which is thought to play a role in determining the host tropism of the different chlamydiaphage (Everson et al., 2003). The IN5 region is also present in the coat proteins of the related microviruses φMH2K and SpV4, forming an insertion loop that produces globular surface protrusions at each icosahedral threefold axis of symmetry (Chipman et al., 1998), which is believed to be a relic of the external scaffolding protein or major spike protein (Brentlinger et al., 2002). It is thought that in chlamydiaphage these insertion loops are responsible for receptor recognition (Read et al., 2000b). This hypothesis fits with the observations that both Chp2 and Chp3, in which the IN5 region is highly conserved, have an identical host range (i.e. they infect C. abortus, C. pecorum, C. caviae and Chlamydophila felis but not C. pneumoniae or C. psittaci), whereas the chlamydiaphage CPAR39, which has a highly divergent IN5 region, exhibits a different host range (i.e. it infects C. pneumoniae, C. abortus, C. pecorum and C. caviae but not C. felis or C. psittaci) (Everson et al., 2003). In addition, the observations that binding of either Chp2 or Chp3 does not prevent binding of CPAR39 to the susceptible host C. abortus suggests that the microviruses may bind different receptors on the chlamydial outer membrane (Everson et al., 2002, 2003). Given that multiple species of chlamydiaphage may bind, and thus infect, a single chlamydial organism raises the possibility of horizontal gene transfer between chlamydiaphage. The opportunity for horizontal gene transfer could be further increased by the extension of the developmental cycle resulting from chlamydiaphage infection inhibiting cell division (Salim et al., 2008). Indeed, horizontal gene-transfer events have been observed to occur in other members of the Microviridae and so this possibility cannot be ruled out (Rokyta et al., 2006). Thus, the question must also arise as to the origin of Chp4, or indeed any of these chlamydiaphage, as to whether the species they were isolated from was their original host, as well as why, to date, they have been observed to be restricted to Chlamydophila species with no detection of infection in the Chlamydia species Chlamydia trachomatis, Chlamydia muridarum and Chlamydia suis (Everson et al., 2003). The identification of the chlamydial viral receptor protein(s) and further understanding of the differences between the closely related Chlamydia and Chlamydophila species will help to answer this question.

In conclusion, this study identifies Chp4, a unique chlamydiaphage in C. abortus and describes the genome sequence (GenBank accession no. AY769964) and organization of Chp4 in comparison to other members of the Chlamydiamicrovirus. Further studies are required to investigate the prevalence of chlamydiaphage in natural infections, their influence on disease outcomes and whether chlamydiaphage can be utilized as tools for the genetic manipulation of Chlamydophila spp.