Distribution and function of prophage phiRv1 and phiRv2 among Mycobacterium tuberculosis complex

Mycobacterium tuberculosis complex (MTBC) is notorious for causing diseases, such as tuberculosis. Tuberculosis caused by M. tuberculosis remains a global public health concern. Two prophages, phiRv1 and phiRv2, can be found among most MTBC genomes. However, no precise functions have been assigned for the two prophages. In this paper, to find out the function of these two prophages, the distribution and function of phiRv1 and phiRv2 in MTBC genomes were analyzed from multiple omics data. We found that complex insertion, deletion, and reorganization appeared on the locus of two prophages in MTBC genomes; some genes of the two prophages can be translated and are functional from proteomic data; the expression of other prophage genes, such as Rv1577c, Rv2650c, Rv2652c, Rv2659c, and Rv2658c, can vary with environmental stresses and might enhance the fitness of MTBC. These data will facilitate our in-depth understanding of their function.


Introduction
Prophages are temperate phages or their components, which are integrated into the host bacterial genome during evolution. The full length of prophages or prophage-like elements can be found in the bacterial genomes (Fan, Xie, Li, & Xie, 2014). Full-length prophages have complete excision and integration cassette, lysis cassette, and bacteriophage structure proteins. They can detach from the host genome, replicate, and assemble into phage particles. In contrast, prophage-like elements cannot shed from the host genome due to various deficiencies. The functions of prophages are very intriguing. Their impact on the host genome diversity is widely noted (Lang, Zhaxybayeva, & Beatty, 2012). Some prophages can encode bacterial virulence factors, thereby involving in bacterial pathogenesis (Brussow, Canchaya, & Hardt, 2004;Rabinovich, Sigal, Borovok, Nir-Paz, & Herskovits, 2012;Varani, Monteiro-Vitorello, Nakaya, & Van Sluys, 2013).
Mycobacterium tuberculosis complex (MTBC) refers to a genetically related group of Mycobacterium species that can render tuberculosis in humans or other organisms. Tuberculosis, though largely treatable, claims more people globally, and particularly, in China, than any other single infectious disease. The knowledge gap about its pathogenesis hinders better control (Fan, Tang, Yan, & Xie, 2014). Two prophage-like elements, phiRv1 and phiRv2, are found within the M. tuberculosis H37Rv genome (Bibb & Hatfull, 2002;Cole et al., 1998;Hendrix, Smith, Burns, Ford, & Hatfull, 1999), with few functionality data (Hatfull, 2010;Pedulla et al., 2003). The role of prophages in the pathogenesis of MTBC remains largely untapped. In this study, genome, proteome, and transcriptomic data were mined to survey the distribution of phiRv1 and phiRv2 among MTBC genomes to explore their function in the context of MTBC pathogenicity.

Data collection
All genome sequences (Table S1) of bacteria for analysis were downloaded from Genbank (http://www.ncbi.nlm. nih.gov/genbank). All transcriptomic and proteomic data used here were freely downloaded from the internet (Betts, Lukey, Robb, McAdam, & Duncan, 2002;

Evaluation methods
Comparative genomic analyses of prophages were performed at the NCBI (http://blast.ncbi.nlm.nih.gov/ Blast.cgi). As we know, twofold difference is a widely accepted cut-off for biological differences. For up-and down-regulated prophage genes, a regulation of at least twofold was used as a criterion. And probabilistic corrections (p-value or false discovery rate) of all transcriptomic data used here have been considered in the primary paper.

Distribution of prophages in MTBC genomes
All retrievable MTBC genomes sequenced before 18 September 2014 were included in this study. Based on BLAST-N, we obtained the prophage information in different MTBC strains (Table S1). Data showed that all strains of M. tuberculosis have at least one prophage (either phiRv1 or phiRv2); M. bovis AF2122/97 contains just one prophage, phiRv1; all strains of M. canettii, M. africanum, and M. bovis BCG have neither phiRv1 nor phiRv2, but some strains of M. canettii harbor other prophages .
We investigated the distribution of phiRv1 and phiRv2 in different strains of M. tuberculosis. There are many substrains in M. tuberculosis, such as EAI (East-African-Indian), CAS (Central-Asian), Beijing, 'X', Haarlem, and LAM (Latin-American-Mediterranean) (Comas, Homolka, Niemann, & Gagneux, 2009). Data showed that all strains of EAI, CAS, and Haarlem strain families and a minority of 'Beijing' strains had both phiRv1 and phiRv2; all strains of 'X' and LAM and a majority of 'Beijing' strains carry just one prophage, phiRv2. Two strains isolated from India, M. tuberculosis RGTB423 and M. tuberculosis RGTB327, harbor phiRv1 and phiRv2, respectively. And RGTB423 is the only M. tuberculosis strain carrying just phiRv1. In fact, prophage deletion had been previously reported in the MTBC genome. Two prophages phiRv1 and phiRv2 of M. tuberculosis H37Rv corresponding to regions RD3 and RD11 were absent in BCG (Brosch et al., 2002). Our data showed that phiRv1 or phiRv2 deletion exists in different M. tuberculosis strains.
In short, the deletion or lost of prophages phiRv1 or phiRv2 during evolution can occur in different species of MTBC, and even the diverse strains of the same species; the insertion sites of phiRv1 vary with different M. tuberculosis strains, and even those strains belonged to the same strain grouping (such as Haarlem grouping); homologous recombination or gene deletion phenomenon can be found in the genome of prophages phiRv1 or phiRv2, and the similarities among these prophages seemed irrelevant to the kinship of their hosts. Prophages usually can enhance the virulence of pathogens (Tinsley, Bille, & Nassif, 2006;Wagner & Waldor, 2002). The comparative genomic analysis applied in this study corroborates that of previous studies (Hatfull, 2010;Pedulla et al., 2003), suggesting that there are no significant associations between MTBC and prophages.
3.2. Proteomic data demonstrated that prophages phiRv1 and phiRv2 are functional units In general, not all prophage ORFs can be translated in the dormant stage (lysogenic). Some prophages exist as dormant DNA within the host genome. Repressor-encoding ORFs (Hendrix, Lawrence, Hatfull, & Casjens, 2000), and proteins endowed bacteria adaption (Wang et al., 2010), such as pathogenic factors, stress resistance proteins, can be actively expressed.
Phage capsid protein (Rv1576c and Rv2650c), phage terminase (Rv1578c), and phage integrase (Rv1586c and Rv2659c) are present in both proteome data. These proteins are not conserved in most sequenced mycobacteriophages. They may help M. tuberculosis H37Rv to adapt to environment and be associated with the pathogenesis of M. tuberculosis H37Rv. Since, prophages phiRv1 and phiRv2 are temperate phages, these proteins found in proteomic data may indicate production of phage particles. Furthermore, phage particles may help M. tuberculosis to adapt to environmental fluctuations and be involved in virulence.

Transcriptomic data 3.3.1. Expression of prophage genes can respond to environmental stresses
As intracellular pathogen, it is essential for MTBC member to tackle different stresses such as, acidic environment, low oxygen, and nutritional deficiency to survive and thrive (Fontan et al., 2008). Prophages can improve the capability of pathogens to overcome such stresses (Wang et al., 2010). Mining the variety of Transcriptomic profilings of M. tuberculosis responses to stresses (Betts et al., 2002;Boshoff et al., 2004;Butcher, 2004;Fontan et al., 2008;Keren et al., 2011;Muttucumaru et al., 2004;Schnappinger et al., 2003) can provide information about prophages.
The overview of prophage gene expression under a variety of biological conditions is presented in Figure 3.  Table S3) (Boshoff et al., 2004;Liang et al., 2011Liang et al., , 2012. Oxidative phosphorylation inhibitors, such as valinomyc, KCN, N,N′-dicyclohexylcarbodiimide (DCCD), and 2,4-dinitrophenol (DNP) promoted the expression of most prophage genes under which the electron transfer chain of M. tuberculosis was blocked and the bacteria are supposed to be under hypoxia. The up-regulated prophage genes might implicate in host hypoxia adaptation or set the stage for the prophage to excise from the host chromosome. Since these prophages are functional units, the most parsimonious explanation is the excision of the prophage rather than a putative role in host hypoxia.

Conclusions
In summary, to shed more light on the poorly defined relation between phiRv1 or phiRv2 and MTBC pathogenesis (Hatfull, 2010;Pedulla et al., 2003), genomic, proteomic, and transcriptomic data of MTBC were mined to find the distribution and function of phiRv1 and phiRv2. At the genomic level, complex insertion, deletion, and reorganization appeared on the locus of two prophages. The proteome data showed that some phiRv1 and phiRv2 genes were translated and had function. The transcriptomic data demonstrated that some prophage genes were induced during many stresses. Moreover, oxidative phosphorylation inhibitors can induce expression of most prophage genes, implicating that phiRv1 and phiRv2 elements can sense oxygen status within the host and respond to the variation accordingly. The diverse phenotypes of phiRv1 and phiRv2 suggest important roles. Deletion experiments are needed to reveal phiRv1 and phiRv2 function.

Supplementary material
The supplementary material for this paper is available online at http://dx.doi.10.1080/07391102.2015.1022602.

Disclosure statement
No potential conflict of interest was reported by the authors.