Different Expression Patterns of Genes from the Exo-Xis Region of Bacteriophage λ and Shiga Toxin-Converting Bacteriophage Ф24B following Infection or Prophage Induction in Escherichia coli

Lambdoid bacteriophages serve as useful models in microbiological and molecular studies on basic biological process. Moreover, this family of viruses plays an important role in pathogenesis of enterohemorrhagic Escherichia coli (EHEC) strains, as they are carriers of genes coding for Shiga toxins. Efficient expression of these genes requires lambdoid prophage induction and multiplication of the phage genome. Therefore, understanding the mechanisms regulating these processes appears essential for both basic knowledge and potential anti-EHEC applications. The exo-xis region, present in genomes of lambdoid bacteriophages, contains highly conserved genes of largely unknown functions. Recent report indicated that the Ea8.5 protein, encoded in this region, contains a newly discovered fused homeodomain/zinc-finger fold, suggesting its plausible regulatory role. Moreover, subsequent studies demonstrated that overexpression of the exo-xis region from a multicopy plasmid resulted in impaired lysogenization of E. coli and more effective induction of λ and Ф24B prophages. In this report, we demonstrate that after prophage induction, the increase in phage DNA content in the host cells is more efficient in E. coli bearing additional copies of the exo-xis region, while survival rate of such bacteria is lower, which corroborated previous observations. Importantly, by using quantitative real-time reverse transcription PCR, we have determined patterns of expressions of particular genes from this region. Unexpectedly, in both phages λ and Ф24B, these patterns were significantly different not only between conditions of the host cells infection by bacteriophages and prophage induction, but also between induction of prophages with various agents (mitomycin C and hydrogen peroxide). This may shed a new light on our understanding of regulation of lambdoid phage development, depending on the mode of lytic cycle initiation.


Introduction
Bacteriophage l, which is a virus infecting Escherichia coli cells, has been used as a model in studies in the fields of microbiology and molecular biology for over 6 decades (for a review, see [1]). Shortly after its discovery, it appeared that there are other bacteriophages, which genomes' organization and developmental pathways are similar to those of l, therefore, the family of lambdoid bacteriophages has been formally established. One of characteristic features of these viruses is their ability to follow two alternative developmental pathways. The lytic pathway includes phage genome replication and synthesis of phage-encoded regulatory and structural proteins, leading to production and liberation of progeny virions. The lysogenic pathway consists of integration of the phage genome into host chromosome, forming a prophage, and a passive replication of this form of the viral genome together with bacterial DNA (host cells bearing integrated phage genomes are called lysogens). However, the lysogenic stage is not permanent. A developmental switch, which consists of prophage induction, excision of phage DNA from the host chromosome, and initiation of the lytic mode of phage development, can occur under certain conditions causing a DNA damage in the host cell (for reviews, see [2,3]).
Genomes of some lambdoid bacteriophages, apart from genes characteristic for the whole family, contain also genes (abbreviated stx) encoding Shiga toxins. If E. coli is lysogenic with such a phage, it may be highly pathogenic to humans. Bacterial strains bearing such prophages are called Shiga toxin-producing E. coli (STEC), and this group includes enterohemorrhagic E. coli (EHEC) strains that are particularly dangerous pathogens [4][5][6]. Bacteriophages bearing stx genes in their genomes are known as Shiga toxinconverting bacteriophages or Stx phages [7][8][9]. The recent outbreak that occurred in Germany in 2011 resulted in over 4,000 symptomatic infections, including over 50 fatal cases. This can be an indication of severity of STEC-mediated infections and significance of medical problems caused by bacteria lysogenic with these phages [10][11][12][13][14].
In this light, detailed understanding the mechanisms of regulation of lambdoid phages' development appears crucial, particularly because Stx prophage induction and effective replication of the phage genome are indispensible for efficient production of Shiga toxins [15][16][17][18]. This stems from the fact that expression of majority of genes of lambdoid prophages, including stx genes in Stx prophages, is strongly inhibited by the phage-encoded cI repressor (despite the fact that this repression may be weaker in at least some Stx phages than in l) [2,3,19]. The prophage induction may be either spontaneous (but it occurs with a low frequency) or caused by factors and agents provoking DNA lesions, thus provoking the bacterial S.O.S. response which indirectly results in cleavage of the cI protein and de-repression of most of phage genes (for reviews, see [2,3]). In the case of Shiga toxin-converting prophages, among such factor and agents there are UV irradiation, antibiotics interfering with DNA metabolism (like mitomycin C), and hydrogen peroxide (which appears to be the most plausible compound causing induction of Stx prophages) [20,21]. However, despite determination of molecular principles of cI-mediated regulation of gene expression, we are still far from complete understanding of mechanisms influencing efficiency of lambdoid prophage induction and its further lytic development.
One mystery in lambdoid phage biology is the b region in the viral genome. It is dispensable for phage development under standard laboratory conditions (a phenomenon which is unusual among viruses), but contains an evolutionarily conserved fragment, located between exo and xis genes and transcribed from the p L promoter. This fragment is called the exo-xis region, and consists of several open reading frames which functions in phage development are largely unknown. Previous studies demonstrated that overexpression of genes from the exo-xis region caused impairment lysogenization of E. coli by bacteriophage l [22]. Subsequent report indicated that the presence of multiple copies of these genes on plasmids positively influenced efficiencies of induction of prophages l and w24 B , one of Shiga toxin-converting phages [23]. Interestingly, it was also found that two orthologs of the l Ea8.5 protein, encoded by a gene located between exo and xis, contain a fused homeodomain/zinc-finger fold [24]. This strongly suggest a regulatory role of this protein.
In the light of the above facts, we aimed to investigate the exoxis region in more details. Genetic maps of exo-xis regions from genomes of bacteriophages l and F24 B are shown in Fig. 1. We asked what are patterns of expression of genes from this region in E. coli cells either infected with bacteriophage l or F24 B , or lysogenic with these phages (after prophage induction). Unexpectedly, we found that in both phages l and F24 B , these patterns were significantly different not only between conditions of infection of the host cells by bacteriophages and prophage induction, but also between induction of prophages with various agents (mitomycin C and hydrogen peroxide). This may shed a new light on our understanding of regulation of lambdoid phage development, depending on the mode of lytic cycle initiation.

Prophage induction experiments
Bacteria lysogenic with tested phages were cultured in Luria-Bertani (LB) medium at 30uC to A 600 of 0.1. Three induction conditions were tested: 0.2 mg/ml mitomycin C, 1 mM H 2 O 2 , and UV irradiation (50 J/m 2 ). At indicated times after induction, samples of bacterial cultures were harvested, and 30 ml of chloroform were added to 0.5 ml of each sample. The mixture was vortexed and centrifuged for 5 min in a microcentrifuge. Then, serial dilutions were prepared in TM buffer (10 mM Tris-HCl, 10 mM MgSO 4 ; pH 7.2). Phage titer (number of phages per ml) was determined by spotting 2.5 ml of each dilution of the phage lysate on a freshly prepared LB agar (1.5%) or LB agar (1.5%) with 2.5 mg/ml chloramphenicol (according to a procedure described by [27]), with a poured mixture of 1-ml indicator E. coli MG1655 strain culture and 2 ml of 0.7% nutrient agar (prewarmed to 45uC), supplemented with MgSO 4 and CaCl 2 (to a final concentration of 10 mM each). Plates were incubated at 37uC overnight. Analogous experiments but without induction agents were performed (control experiments) with each lysogenic strain. Presented values show phage titer (PFU/ml) normalized to results of control experiments (representing ratios of phage titers in induced and non-induced cultures). Each experiment was repeated three times.

One-step-growth experiment
Lytic development of lambdoid phages was studied in one-stepgrowth experiments. Bacteria were grown in LB medium supplemented with MgSO 4 and CaCl 2 (to a final concentration of 10 mM each) at 30uC to A 600 = 0.2. Samples of 10 ml were withdrawn and centrifuged (3,0006g, 10 min). Each pellet was suspended in 1 ml (1/10 of initial volume) of 3 mM NaN 3 in LB. Following 5-min incubation at 30uC, the phage was added to multiplicity of infection (m.o.i.) of 0.05. Phage adsorption was carried out at 30uC for 10 min. The mixture was diluted ten-fold in warm (30uC) 3 mM NaN 3 in LB and centrifuged (3,0006g, 10 min). Bacterial pellet was suspended in 1 ml of LB with 3 mM NaN 3 and centrifuged again (3,0006g, 10 min). This procedure was repeated three times. The suspension was then diluted 1,000fold with LB, prewarmed to 30uC (time 0), and aerated in a water bath shaker at this temperature. The number of infective centers was estimated from nine samples taken in the interval of 0-15 min after the dilution by plating under permissive conditions. The number of intracellular progeny phages (samples previously shaken vigorously for 1 min with equal volume of chloroform and cleared by centrifugation) was estimated by plating on appropriate indicator bacteria. Plates were incubated at 37uC overnight. Each experiment was repeated three times. Dark arrows with continuous outside lines represent highly conserved genes and open reading frames (over 70% nucleotide as well as amino acid sequence identity). Dark arrows with punctuated outside lines represent highly conserved (above 70% sequence identity) open reading frames present in genomes of l and 933W phages, available in the NCBI database but uncharacterized in annotations. The presence of orf73 in the l exo-xis region was indicated by [48]. Light arrows represent genes and open reading frames with low level of sequence identity (,38%). Note the high homology between l and F24 B exo-orf73 regions and low level of identity of other analyzed genes. Arrows indicate positions of promoters predicted with BPROM program. The localizations and 210 and 235 sequences of predicted promoters p 1_l and p 1_W24B are exactly the same (see Table 3). Schematic steam-loop structures [ N ] indicate localizations of predicted transcription terminators, found on the basis of nucleotide seuquence analysis with ARNold software. The localizations and sequences of predicted terminators t 1 and t 2 are exactly the same in case of both phages l and F24 B (see Table 4). Note that in the case of F24 B phage, some ORFs from the exo-xis region: vb_24B_9c, vb_24B_8c, vb_24B_7c, putative C4 zinc finger protein and vb_24B_6c are homologues of l orf60a, orf63, orf61, orf73 and gene ea22 respectively. For clarity of this work, only the names of l ORFs were used. doi:10.1371/journal.pone.0108233.g001

Measurement of bacterial viability during prophage induction experiments
Bacteria lysogenic for tested phages were cultured in LB medium at 306C to A 600 = 0.1 with induction agent, either 0.2 mg/ml mitomycin C or 1mM H 2 O 2 , added at time zero. At indicated times after induction, samples equal to 2610 8 cells/ml were withdrawn and centrifuged at 10,0006g for 10 min. The supernatants were removed and pellets washed in 0.85% NaCl. Bacterial suspensions were stained with LIVE/DEAD BacLight Bactrial Viability Kit (Molecular Probes), according to the manufacturer's protocol. The mentioned kit utilizes mixtures of the green-fluorescent nucleic acid stain SYTO 9 and the redfluorescent nucleic acid dye, propidium iodide. When used alone, the SYTO 9 dye stains bacteria with intact as well as damaged membranes (e.g. arising as a result caused by phage host-cell lysis). In contrast, propidium iodide penetrates only into bacteria with damaged membranes, causing a reduction in the green dye fluorescence when both stains are present. Measurements of fluorescence were performed in microplate reader using excitation wavelength = 485 nm and emission wavelengths = 530 nm for SYTO 9 dye, and 630 nm for propidium iodide dye [28]. Data were analyzed by dividing the fluorescence intensity of stained bacterial suspensions at emission = 530 nm by the fluorescence intensity at emission = 630 nm. Presented values show percent of live bacteria normalized to results of control experiment, noninduced cultures which at each time were assumed as 100% of live bacteria. Each experiment was repeated three times.

Estimation of relative phage DNA amounts
Bacteria lysogenic for tested phages were cultured in LB medium at 30uC to A 600 of 0.1. Induction of prophages was provoked in lysogenic bacteria by addition of mitomycin C to a final concentration of 0.2 mg/ml or H 2 O 2 to a final concentration 1 mM. At indicated times after induction, 2-ml samples with 120 ml of chloroform were vortexed for 10 s and centrifuged in a microcentrifuge for 5 min. The supernatants were collected and filtered through a membrane filter with pore size of 0.22 mm (Sigma-Aldrich) to remove bacterial cells. Filtered samples were first treated with DNase I (20 mg/ml; Sigma-Aldrich) for 30 min at 37uC to remove any free bacterial DNA. The viral DNA was then liberated from virions using the method described by [29], with a minor modification. All DNase-treated samples were heated at 37uC for 60 min in the presence of 20 mM EDTA (Sigma-Aldrich) and 50 mg/ml proteinase K (Sigma-Aldrich). DNA was quantified by staining with Qubit dsDNA BR Assay Kit (Invitrogen), according to the manufacturer's instructions. Concentration of phage DNA (in mg/ml) were calculated relative to analogous experiments but without induction agents (control experiments) with each lysogenic strain. Presented values show phage DNA concentration (mg/ml) normalized to results of control experiments. Each experiment was repeated three times.

Bacteriophage infection
Host bacteria were grown to A 600 of 0.3 at 30uC. Then, 120 ml volume was centrifuged and the pallet was washed with 30 ml of 0.85% NaCl. After centrifugation, the pallet was suspended in 36 ml of LB medium supplemented with MgSO 4 and CaCl 2 (to a final concentration of 10 mM each). The mixture was incubated for 30 min at 30uC and then chilled on ice. Bacteriophage lysate was added to m.o.i. of 5. Following 30 min incubation on ice, at indicated times, 1610 9 bacterial cultures were treated with NaN 3 (Sigma-Aldrich) to a final concentration of 10 mM and harvested. The preparation of RNA and cDNA were performed as described in the subsequent subsection.

Preparation of RNA and cDNA from bacteria
For the preparation of RNA, the induction of temperate bacteriophages from E. coli strain MG1655 was performed with

Real-time PCR Assay
For transcriptional analysis of tested genes by quantitative realtime reverse transcription-PCR (qRT-PCR), the qRT-PCR was performed with the LightCycler 480 Real-Time PCR System (Roche Applied Science), with cDNA samples from lysogenic bacteria. Transcription rates of w24 B and l genes were compared in parallel to those of the icdA (according to a procedure described by [30]) or 16S rRNA housekeeping genes. Primers were developed by Primer3web version 4.0.0 and produced by Sigma-Aldrich or GENOMED. The transcriptional analysis of w24 B and l genes was performed with primers presented in Table 2. Realtime PCR amplifications were performed for 55 cycles in 20-ml reaction volumes by employing LightCycler 480 SYBR Green I Master (Roche Applied Science). Reactions were performed in Roche 96-well plates containing 10 ml 2x SYBR Green I Master Mix, 6.25 ng/ml cDNA and 200 nM of each gene-specific primer ( Table 2). Relative quantification assays were performed with cDNA in an icdA or 16S rRNA and phage genes multiplex assay. All templates were amplified using the following program: 956C for 5 min; 55 cycles of 956C for 10 s; 606C for 15 s and 726C for 15 s. No template control was included with each run. Each  reaction was repeated three times and the specificity of amplified products was examined by melting curve analysis immediately after the final PCR cycle, and confirmed by gel electrophoresis.

Real-time PCR data analysis
To analyze the relative changes in gene expression revealed by quantitative Real-Time PCR experiments, the calibrator normalized relative quantification method with efficiency correction (socalled E-Method) was used. The E-Method provides an efficiency corrected calculation mode by using the determined PCR efficiency of target (E t ) as well as the efficiency of reference (E r ). Relative fold change ratio was calculated by using the following formula, described in the application manual of Roche Light-Cycler Real-Time PCR Systems [31]: Normalized relative ratio = E t CT(t) calibrat u r -CT(t) sample /E r CT(r) calibrat u r -CT(r) sample (where ''t'' is target, and ''r'' is reference).
The sample at the time point ''zero'' was a calibrator. The raw run data for w24 B and l genes were transferred from the LightCycler 480 to the LinRegPCR 12.5 software using the ''LC480 Conversion: conversion of raw LC480 data'' software (available at http://www.hartfaalcentrum.nl/index. php?main=files&sub=0). PCR efficiency was determined for each gene by LinRegPCR program [32,33]. This software was successfully used previously to calculate PCR efficiency [34][35][36][37][38].

Prediction of the presence of transcription promoters and terminators in phage genomes
Prediction of the presence of promoters in the sequences of genomes of l and w24 B phages was performed using BPROMthe bacterial s 70 promoter recognition program available at: http://linux1.softberry.com.
Promoters were searched within ,400 bp long sequence fragment before the start of the orf73 coding region, encompassing whole sequence of orf61 and the region between orf61 and orf73. BPROM has accuracy of E. coli promoter recognition about 80%, and considers promoters with score above 0.20 [39].
Predicted terminators were found within the exo-xis region of l and w24 B using ARNold, the online analysis tool which predicts the existence and location of rho-independent transcription terminators using RNAmotif and ERPIN complementary programs [40][41][42][43]. The ARNold program is available at: http://rna. igmors.u-psud.fr/toolbox/arnold/.

Effects of the multicopy exo-xis region on l and w24 B development and host survival at 30uC
Previous studies demonstrated that the presence of the exo-xis region on a multicopy plasmid in the host cells caused enhanced efficiency of induction of prophages l and w24 B under standard laboratory conditions (rich medium, 37uC). Since the main aim of this work was to determine detailed patterns of expression of genes from the exo-xis regions of the tested lambdoid phages, after preliminary experiments, we decided to culture host bacteria at 30uC, rather than at 37uC, to slow metabolic processes down which made the analyzes more precise. However, such a change in cultivation conditions relative to previously reported studies (temperature 30uC instead of 37uC) made it necessary to check whether effects of multiple copies of the exo-xis region on phage development are similar in both experimental systems.
We found that lytic developments of both tested phages, l and w24 B , after infection of the E. coli cells growing at 30uC, were not significantly affected by the presence of corresponding exo-xis regions on multicopy plasmids (Fig. 2). In the case of phage l, effects of the presence of particular fragments of this region were also tested, again with no considerable changes detected (Fig. 2).
This corroborates results of analogous experiments performed at 37uC and reported previously [23].
Studies on l and w24 B development after prophage induction were carried out in the host cells treated with either mitomycin C or hydrogen peroxide. Again, results obtained in experiments conducted at 30uC (Fig. 3) were generally similar to those at 37uC, described in the previous article [23]. However, positive effects of multiple copies of the exo-xis region in both l and w24 B were even more pronounced at 30uC (compare Fig. 3 with [23]). These results also confirmed our presumption that performing the planned experiments at 30uC, rather than 37uC, may allow us to detect any putative differences more efficiently.
Accordingly to previous studies [23], we have also tested development of both tested phages in the presence of the plasmidborne exo-xis region after prophage induction with UV irradiation. Similarly to experiments with mitomycin C and hydrogen peroxide (Fig. 3), production of phage progeny of both l and w24 B started earlier and was more efficient when additional copies of the exo-xis region were present in host cells (data not shown). Because of these similarities, in further experiments we have focused on prophage induction conditions which are more likely to occur in the intestine, a natural environment of the host bacterium [4][5][6][7][8][9], i.e. the presence of an antibiotic (mitomycin C) or H 2 O 2 , rather than UV irradiation. Finding experimental conditions (30uC rather than 37uC) causing more pronounced differences in development of bacteriophages between hosts containing and lacking additional copies of the exo-xis region allowed us to address the problem whether expression of genes and/or ORFs from this region is responsible for the observed effects. This was reasonable as one might assume that the tested fragment of phage genome might bind regulatory factor(s), titrating it/them out, and thus causing changes is the control of viral development.
To test such a possibility, we have employed the l model with two experimental systems. First, instead of using the whole exo-xis region, host cells were transformed with plasmids bearing particular genes or ORFs or their combinations. Second, we have constructed a series of derivatives of pGAW3775tet (a plasmid bearing the whole exo-xis region) where each plasmid contains a frameshif mutation in particular gene or ORF. In the first experimental system, the absence of particular fragments of the exo-xis region relative to pGAW3775tet resulted is intermediate effects, i.e. the phage development was more effective than in the host bearing the vector, but less effective than in the pGAW3775tet-bearing host (Fig. 3 A and B). However, in some cases, specific effects were detected, namely, the presence of either pJWea22 (bearing ea22) or pJWorfea22 (bearing orf61, orf60a, orf63 and ea22) did not result in more rapid induction or phage development in the mitomycin C-treated host (Fig. 3A), and the presence pJWea22 (bearing ea22) caused a delay in the phage development in hydrogen peroxide-treated cells (Fig. 3B).
In the second experimental system, frameshift mutations in orf61 and ea22 (in the exo-xis fragment present in the plasmid) resulted in a delay in phage development after prophage induction with mitomycin C (Fig. 3C), while frameshit mutations in orf61, orf73 and ea8.5 caused similar effects in hydrogen peroxidetreated bacteria (Fig. 3D). In addition, we have measured the frequency of spontaneous (without addition of any specific agent) prophage induction to find similar values in cells bearing pGAW3775tet (with wild-type exo-xis region) and most of constructs with frameshift mutations in one of genes or ORFs (the frequencies were about 10 25 per cell), which were about 10 times higher than in bacteria bearing a control plasmid pJW0tet (about 10 26 per cell). However, a frameshift mutation in orf60a abolished the effect caused by the presence of a plasmid with the exo-xis region (the frequency was about 10 26 per cell). Although these results did not exclude a possibility for titrating out regulatory factors(s) by the exo-xis region, they suggested specific roles of expression of at least some genes and ORFs, particularly ea22, ea8.5, orf61, orf73, and orf60a.
Additional confirmation of the more efficient development of phages l and w24 B in cells treated at the lysogenic stage with mitomycin C or hydrogen peroxide was provided by measurement of an increase in bacteriophage DNA amount. Again, more efficient increase in the level of DNA of both tested phages was observed after induction of corresponding prophages with both tested inductors in cultures of hosts bearing plasmids with appropriate exo-xis region relative to those containing plasmid vector (Fig. 4). As expected, in the same experimental system, survival rate of bacteria with the vector was always higher than cells bearing a plasmid with the exo-xis region (Fig. 5). Survival of a high percentage of bacterial cells after the induction might seem surprising. However, one should note that contrary to UV irradiation, the efficiency of prophage induction after treatment of the host (lysogenic with l or w24 B ) with mitomycin C or hydrogen peroxide may be moderate. In fact, experimental data indicated that in mitomycin C-or hydrogen peroxide-treated cultures of such bacteria, prophage induction occurred in less than 50% or even in only a few percent of cells, respectively [20,21]. Therefore, relatively large fraction of cells may survive, and due to lysogenic state, they are immune to superinfection by the same phage. This may allow these cells to grow and divide, resulting in values even over 100% in the employed experimental system (Fig. 5).
In summary, these results clearly indicated that the presence of the exo-xis region on a multicopy plasmid stimulates development of bacteriophages l and w24 B after treatment of lysogenic cells with either mitomycin C or hydrogen peroxide at 30uC.
Expression of genes from the exo-xis regions of bacteriophages l and w24 B in infected host cells To assess patterns of expression of genes from the exo-xis region in phage-infected bacteria, E. coli wild-type (MG1655) cells were infected with either l or w24 B , and at certain times after infection total RNA was isolated and levels of tested transcripts were determined by quantitative real-time reverse transcription PCR. The exo-xis region is believed to be transcribed from the leftward p L promoter, one of two major lytic promoters of lambdoid phages [3,8]. Positions and sequences of predicted promoters and terminators located in the exo-xis regions of genomes of both tested phages were predicted, and they are indicated in Fig. 1 and summarized in Tables 3 and 4, respectively.
Lytic development of bacteriophage l is quicker than that of w24 B , as demonstrated in Fig. 3, therefore, it is not a surprise that expression of the major genes of phage lytic development, N and cro, occurred earlier after the infection with the former phage than in the latter one (Fig. 6). The presence of low level signals, followed by a maximal amount of the transcript, and finally by decreased intensities of signals in the reactions of detection of Nand crospecific mRNAs indicates that appropriate time-frames were chosen to assess the expression efficiency. Another proof of the properly chosen times of sample withdrawn are low levels of mRNAs for Q and R genes, coding for proteins synthesized at the late stage of bacteriophage development.
Interestingly, some of the genes from the exo-xis regions were expressed (at the RNA level) as efficiently as, or with a similar efficiency to, the N gene. This was true for orf60a, orf63, and orf61 in phage l, and orf60a, orf63, orf73, and ea22 in phage w24 B (Fig. 6). Moreover, although levels of mRNA for orf73, ea22 and ea8.5, which are downstream of the predicted terminator (located between orf61 and orf73), were significantly lower than genes located upstream of the terminator in l, considerably different expression pattern has been observed in w24 B . In host cells infected by the latter phage, orf61 was poorly expressed, while levels of mRNAs for orf73 and ea22 were as high as those for N, orf61a and orf63 (Fig. 6). Interestingly, upstream of orf73 and ea22, a promoter p 1_W24B was predicted by BPROM program (Fig. 1 and Table 3). The p 1_W24B promoter sequences 210 and 2 35 are identical with analogous sequences of the p 1_l promoter, localized upstream of the l orf73.
Expression of genes from the exo-xis regions of bacteriophages l and w24 B after prophage induction with various agents Expression of phage genes after prophage induction was assessed in host cells treated with either mitomycin C or hydrogen peroxide. In both tested phages, l and w24 B , characteristic timecourse of mRNA amounts for N and cro genes, encompassing low level, maximum, and decreased levels, was achieved at significantly later times after induction than after infection ( Fig. 7 and Fig. 6, respectively). In phage l, levels of N and cro transcripts were significantly lower than those of genes from the exo-xis region, which differs from the pattern determined in phageinfected cells. Moreover, N and cro were expressed at similar times to Q expression, which again indicates the difference between two variants of initiation of the phage lytic development (infection vs. induction). Additionally, in the case of mitomycin C-induced l prophage, the level of expression of ea8.5, ea22 and orf73 was significantly decreased in comparison with other ORFs from the exo-xis region (Fig. 7A) what could be explained by the presence of predicted t 2_l transcription terminator, localized between orf61 and orf73 ( Fig. 1 and Table 4).
In phage F24 B , expression of all tested genes was delayed after the prophage induction relative to initiation of the lytic development by infection ( Fig. 7 and 6, respectively). Nevertheless, expression patterns were quite similar between mitomycin Ctreated lysogenic cells and bacteria infected with viruses. Surprisingly, when the prophage induction was caused by treatment of host cells with hydrogen peroxide, levels of mRNAs for orf73 and ea22 were significantly higher than other tested genes. This expression pattern differs considerably from that observed in mitomycin C-induced lysogens and might be explained by the presence of the p 1_W24B promoter upstream of orf73 ( Fig. 1 and Table 3). On the other hand, in both cases, hydrogen peroxide-and mitomycin C-induced F24 B prophages, the change in the expression level was observed between orf61 and orf73 ( Fig. 7C and D). At this point, it should be noted that localizations and sequences of predicted transcription promoters p 1_l and p 1_W24B , as well as terminators t 2_l and t 2_W24B are exactly the same ( Fig. 1 and Table 4).

Discussion
Although bacteriophage l and related phages have been used as models for genetic studies, including regulation of gene expression, reports on global analyses of gene expression (especially time course of the expression) of lambdoid viruses are rare in the literature. Recently, expression pattern of phage l genes after thermal induction of the prophage bearing a temeperaturesensitive mutation in the cI gene was investigated by ribosome profiling [44], and genes' expression of phage w24 B in the lysogenic host was studied using a proteomic approach [19]. In both cases, unexpected results were obtained. In bacteriophage l developing after prophage induction, global gene expression analysis revealed the activity of various previously unappreciated open reading frames [44]. In both l and w24 B , apart from cI, rexA, rexB, lom and bor -genes previously known as those transcribed from a prophage, several other genes were found to be expressed in non-induced lysogens; those included ea8.5 in l, and exo in w24 B , genes studied also in our work. The recently published reports, mentioned above, indicated quite an unexpected complexity in the patterns of expression of genes of lambdoid phages, suggesting the existence of multiple regulatory systems, not yet identified in these viruses.
In this work, we have investigated expression patterns of genes from the exo-xis regions of phages mentioned above (l and w24 B ) after infection of host cells or induction of corresponding prophages. We have chosen this genome region because recent studies indicated that genes included there had significant effects on efficiency of lysogenization and prophage induction in both tested phages [22,23]. Moreover, the Ea8.5 protein, encoded in this region, was found to contain a specific motif, strongly suggesting its regulatory role [24].
Similarly to two other recent analyses [19,44], our studies led to unexpected results. First, time courses of expression of the investigated genes, including those coding for major regulatory proteins, N and cro, were significantly different in phage-infected cells and in induced lysogens. Second, despite homologous regulatory sequences (promoters and terminators), identified and predicted in genomes of l and w24 B , gene expression patterns were significantly different between these two tested phages. Third, even in the same phage, considerably different patterns of gene expression were detected, depending on the nature of agent (mitomycin C or hydrogen peroxide) used to induce the w24 B prophage.
At the current stage of our knowledge, it is difficult to predict the mechanisms of the differential expression of phage genes during lytic development initiated by different ways, either infection or prophage induction. Even harder to understand is different expression of genes from the exo-xis region of phage w24 B , when prophage induction is caused by either mitomycin C or hydrogen peroxide. One would expect that both agents should induce the S.O.S. response in the host cells, which should lead to degradation of the cI repressor and subsequent prophage excision, followed by expression of phage genes as in the lytic cycle initiated by infection. Definitely, regulations of these processes are significantly more complicated than assumed.
One example of unexpected specific regulation arises from analysis of the patterns of expression of genes from the exo-xis region of phage w24 B . In silico analysis predicted the existence of both promoter and terminator between orf61 and orf73, homologous to those in l. In bacteriophage l, levels of transcripts for ORFs located upstream of the predicted terminator are lower than those located downstream of this terminator, irrespective of the way of initiation of the lytic development. However, orf73 and ea22 of w24 B are efficiently expressed despite the presence of this terminator, which is especially well pronounced in cells treated with hydrogen peroxide, where orf73 and ea22 are expressed at significantly higher levels than the rest of the exo-xis region. This might suggest the presence of a promoter upstream of orf73, and it was confirmed by the analysis with the use of BPROM program which allowed us to localize a predicted s 70 -dependent promoter upstream of orf73. Despite unknown mechanisms responsible for differential expression of genes from the exo-xis region of bacteriophages l and w24 B , the fact that transcripts of ORFs from this region occur at the significant levels during phages' development suggests that they can play important regulatory roles in development of these viruses, as suggested previously on the basis of biological experiments [22,23]. Such a proposal is corroborated by results presented in this report (Figs. [2][3][4][5], indicating more efficient development of both bacteriophages after prophage induction in cells bearing additional copies of the exo-xis region on plasmids, and by the recent finding that Ea8.5, encoded in this region, contains fused homeodomain/zinc-finger fold [24]. The experiments with plasmids bearing particular ORFs or genes, rather than the whole exo-xis region, as well as with plasmids bearing frameshift mutations in particular ORFs or genes, did not exclude a possibility that an excess of phage-derived sequences could cause effective binding of some regulatory factors and titrating them out. This could be responsible for observed effects on phage development. However, such a scenario seems unlikely as a sole mechanism of this phenomenon, especially in the light of specific effects of particular mutations (Fig. 3). Therefore, we suggest that it is more probable that stimulation of development of l and w24 B in cells bearing additional copies of the exo-xis region arises from effects of expression of certain genes, and possible cooperative actions of at least some of their products. In fact, only two genes from this region, ea8.5 and ea22, were confirmed to date to encode proteins, thus others are named as orfs. However, specific effects of frameshift mutations in orf61, orf73, and orf60a ( Fig. 3C and D) suggest that they may code for biologically active polypeptides.
Recent bioinformatics and microarray analyses have indicated a growing number of genes encoding small proteins in the range of 20-130 amino acids [45][46][47]. Increasing amount of experimental data demonstrates that such small proteins have variety of roles and different mechanisms of action. They can regulate functions of larger proteins, act as signaling factors or structural proteins [47]. Expression patterns of the l genome from ribosome profiling [44] as well as our results from qRT-PCR analyses show increased expression of ORFs of unknown function during l phage lytic development (in between these localized in the exo-xis region as shown in Fig. 1 from work [44] and Fig. 7 in this work). Although expression is observed at different times after induction, it could be explained by application of different inductors, temperatures of cultivation and measurement methods. Additionally, in this work we first present the increased expression of ORFs from exo-xis region during lytic development of phage w24 B . It is important to note that the level of expression of some of the analyzed ORFs from the exo-xis region is comparable to that of known genes or even higher. As suggested previously [44] such observation allow to suppose that these ORFs might be translated into active polypeptide products. Therefore, our further research will focus on determination of biological and biochemical roles of products of genes included in the exo-xis region, as well as on determination of regulatory mechanisms operating in the process of the controlling of expression of these genes.