Lon Protease- and Temperature-Dependent Activity of a Lysis Cassette Located in the Insecticidal Island of Yersinia enterocolitica

The investigation of the mechanisms that help pathogens survive in the environment is a prerequisite to understanding their evolution and their virulence capacities. In members of the genus Yersinia, many factors involved in virulence, metabolism, motility, or biofilm formation follow a strict temperature-dependent regulation.

ABSTRACT

The Yersinia genus comprises pathogens that can adapt to an environmental life cycle stage as well as to mammals. Yersinia enterocolitica strain W22703 exhibits both insecticidal and nematocidal activity conferred by the tripartite toxin complex (Tc) that is encoded on the 19-kb pathogenicity island Tc-PAI_Ye. All tc genes follow a strict temperature regulation in that they are silenced at 37°C but activated at lower temperatures. Four highly conserved phage-related genes, located within the Tc-PAI_Ye, were recently demonstrated to encode a biologically functional holin-endolysin gene cassette that lyses its own host W22703 at 37°C. Conditions transcriptionally activating the cassette are not yet known. In contrast to Escherichia coli, the overproduction of holin and endolysin did not result in cell lysis of strain W22703 at 15°C. When the holin-endolysin genes were overexpressed at 15°C in four Y. enterocolitica biovars and in four other Yersinia spp., a heterogenous pattern of phenotypes was observed, ranging from lysis resistance of a biovar 1A strain to the complete growth arrest of a Y. kristensenii strain. To decipher the molecular mechanism underlying this temperature-dependent lysis, we constructed a Lon protease-negative mutant of W22703 in which the overexpression of the lysis cassette leads to cell death at 15°C. Overexpressed endolysin exhibited a high proteolytic susceptibility in strain W22703 but remained stable in the W22703 Δlon strain or in Y. pseudotuberculosis. Although artificial overexpression was applied here, the data indicate that Lon protease plays a role in the control of the temperature-dependent lysis in Y. enterocolitica W22703.

IMPORTANCE The investigation of the mechanisms that help pathogens survive in the environment is a prerequisite to understanding their evolution and their virulence capacities. In members of the genus Yersinia, many factors involved in virulence, metabolism, motility, or biofilm formation follow a strict temperature-dependent regulation. While the molecular mechanisms underlying the activation of determinants at body temperature have been analyzed in detail, the molecular basis of low-temperature-dependent phenotypes is largely unknown. Here, we demonstrate that a novel phage-related lysis cassette, which is part of the insecticidal and nematocidal pathogenicity island of Y. enterocolitica, does not lyse its own host following overexpression at 15°C and that the Lon protease is involved in this phenotype.

INTRODUCTION

Yersinia enterocolitica is a foodborne, psychrotolerant pathogen that causes a range of gastrointestinal diseases (1). Since pigs are the major animal source of yersiniae, yersiniosis is assumed to be transmitted from contaminated pork, especially minced meat (2). Y. enterocolitica is widely spread in nature and can survive for extended periods in aquatic and terrestrial reservoirs at ambient temperature. The species Y. enterocolitica comprises six biovars (3), of which biovar 1A strains lacking Yersinia virulence plasmid pYV (4) are avirulent, whereas biovar 1B strains are highly virulent in humans and mice due to a pathogenicity island responsible for iron acquisition (5). Biovar 2 to 5 strains, which are predominately isolated in Europe and Japan, are weakly pathogenic and unable to kill mice (6). Biovar 2 strain W22703 has been experimentally demonstrated to be toxic against insect larvae and the nematode Caenorhabditis elegans (7–9), illustrating a multiphasic life cycle of this pathogen that allows its oscillation between invertebrates and vertebrates (10–12).

A key determinant that enables Y. enterocolitica strain W22703 to kill invertebrates is the pathogenicity island Tc-PAI_Ye. It encodes the tripartite, high-molecular-weight toxin complex (Tc) that was first characterized in the entomopathogenic bacterium Photorhabdus luminescens and acts on actin and Rho-GTPases of host cells through ADP-ribosylation (13). Tc-PAI_Ye was found in Y. pestis, Y. pseudotuberculosis, and Y. enterocolitica, excluding Y. enterocolitica biovars 1A and 1B, and in Y. mollaretii and Y. similis, whereas it is absent from the genomes of all other Yersinia species (14). The expression of the tc genes tcaA, tcaB, tcaC, and tccC in Y. enterocolitica strain W22703 occurs in a temperature-dependent manner. These genes are completely silenced at 37°C and activated at lower temperatures, with a maximal activation at 10°C (8). The molecular mechanism underlying the regulation of tc gene expression is mainly based on an antagonism between the regulator TcaR2 on the one hand and a complex of the DNA-binding protein H-NS, which is involved in gene silencing, and the Yersinia modulator of virulence, YmoA, on the other hand. At low temperatures, TcaR2 outcompetes H-NS/YmoA and activates transcription of the tc genes. At 37°C, the copy numbers of autoregulated TcaR2 are reduced, whereas the amount of functional YmoA increases, allowing the H-NS/YmoA complex to interact with the promoters of the tc genes and, thus, to inhibit their transcription (15, 16).

In addition to the tc genes, Tc-PAI_Ye carries two small phage-related genes coding for the putative holin HolY and the endolysin ElyY and two open reading frames (ORFs) encoding a putative i-spanin and o-spanin, respectively, which we termed yersinial Rz and Rz1, in line with the phage lambda nomenclature (17, 18). It is assumed that the spanin complex contributes to the disruption of the outer membrane (19). The lysis cassette of strain W22703, with a total length of approximately 1.3 kb, is located between tcaC and tccC and is highly conserved in all Tc-PAI_Ye islands identified so far (14). No other phage-related genes are found in the neighborhood of Tc-PAI_Ye, suggesting the acquisition of Tc-PAI_Ye by a common ancestor of the pathogenic Yersinia strains via phage transduction or the stable propagation of a respective prophage relic due to an unknown selection pressure. Due to the lack of native conditions under which the holin and endolysin (HE) genes are induced, we established the overexpression of the lysis cassette via the pBAD system and demonstrated that the phage-related genes indeed code for a holin and an endolysin. By cell staining, we showed that growth stagnation at 37°C is caused by cell death upon lysis (17).

The presence of the putative i- and o-spanins indicated the attenuating effect of HE on yersinial growth. When heterologous hosts were used, Y. intermedia, Y. aldovae, Y. frederiksenii, P. luminescens, and Salmonella enterica serovar Typhimurium were revealed to be resistant to the overexpressed cassette (17). Although these data demonstrated the functionality of the lysis cassette from Tc-PAI_Ye in yersiniae, its biological role remains unclear.

In this study, we investigated the effects of the holin/endolysin cassette at environmental temperature, e.g., 15°C, following overexpression in Y. enterocolitica W22703 and related biovars and species. We demonstrated that the biological activity of the lysis cassette is abolished at low temperature in strain W22703, whereas other Yersinia species remained susceptible to lysis. Evidence is provided that the Lon protease blocks temperature-dependent lysis, probably by degrading the endolysin ElyY.

RESULTS

The activity of the lysis cassette is high temperature dependent in Y. enterocolitica strain W22703.

To test the function of the lysis determinant, the HE genes alone or together with yRz and yRz1 (HE⁺), respectively, were overexpressed using the plasmids pBAD33-HE and pBAD33-HE⁺ in Y. enterocolitica W22703. At 37°C, the optical density at 600 nm (OD₆₀₀) arrested and decreased within 2 h after the addition of 0.2% arabinose (Fig. 1A). At 15°C, however, the overexpression of the lysis genes had no significant effect on the growth phenotype of strain W22703, indicating that temperature is a major clue of the autolysis (Fig. 1B). Control cultures lacking arabinose and the proof of holin gene induction by arabinose at 15°C are shown in File S1A and B in the supplemental material.

FIG 1.

Temperature-dependent activity of the lysis cassette in Y. enterocolitica W22703. W22703 cells were transformed with pBAD33, pBAD33-HE, and pBAD33-HE⁺. An overnight culture was diluted 1:1,000 in LB medium. Following incubation either at 37°C (A) or at 15°C (B) with shaking, arabinose was added at a final concentration of 0.2% when the cultures reached an OD₆₀₀ of 0.25, as indicated by the arrow. Time point zero was set immediately after dilution. Each experiment was biologically independently and performed thrice; error bars show the standard deviations.

To further investigate the functionality of the lysis determinant in response to temperature, the HE genes were overexpressed using the plasmids pBAD33 and pBAD33-HE in the heterologous host E. coli strain LMG194 growing at 37°C and 15°C. At 37°C, growth arrested 2 h after the addition of arabinose and restarted 4 h later (Fig. 2A). Growth resumption can probably be attributed to the all-or-none induction of the arabinose promoter P_BAD (20), resulting in selection for recombinant plasmid loss (17). At 15°C, the OD₆₀₀ stagnated within the first 3 h after addition of 0.2% arabinose and did not resume until at least 27 h later, probably due to the longer generation time at low temperature (Fig. 2B). Taken together, these data demonstrate that the lysis cassette from Tc-PAI_Ye is functional in Y. enterocolitica W22703 at 15°C but not at 37°C. In contrast, the heterologous host E. coli LMG194 is lysed by the yersinial cassette at both temperatures.

FIG 2.

Activity of the holin-endolysin cassette in E. coli LMG cells. E. coli LMG194 cells were transformed with pBAD33 and pBAD33-HE. An overnight culture was diluted 1:1,000 in LB medium. Following incubation at 37°C (A) or 15°C (B) with shaking, arabinose was added to a final concentration of 0.2% at an OD₆₀₀ of 0.2 (arrows). Time point zero was set immediately after dilution. Each experiment was independently performed thrice; standard deviations are indicated by error bars.

Overexpression of the lysis cassette in Yersinia spp. at 15°C.

To investigate the activity of the lysis genes at 15°C in Yersinia spp., pBAD-33 and pBAD33-HE⁺, including the i-spanin yRz and the o-spanin yRz1, were tested in representatives of Y. enterocolitica biovars 1A, 2, 3, and 5, in Y. pseudotuberculosis, and in the apathogenic Yersinia species Y. intermedia, Y. frederiksenii, and Y. kristensenii. At 15°C, the overexpression of the lysis cassette HE⁺ had no significant effect on the growth of the Y. enterocolitica biovar 5 strain compared to the control strain equipped with pBAD33 (P = 0.2131) but reduced the maximal cell density of biovar 2 and biovar 3 strains and attenuated the growth of a biovar 1A strain during the exponential phase (P = 0.0040) (Fig. 3A). These data are in contrast to similar experiments performed at 37°C that showed more drastic effects of HE⁺ overexpression (17).

FIG 3.

Overexpression of the holin-endolysin cassette in Yersinia spp. at 15°C. (A) Plasmids pBAD33 and pBAD33-HE⁺ were transformed into Y. enterocolitica biovar 1A, 2, 3, and 5 strains. Overnight cultures were diluted in LB medium to an OD₆₀₀ of 0.07, and measurements were performed in microtiter plates incubated at 15°C with shaking. Induction with 0.2% (wt/vol) arabinose was performed at an OD₆₀₀ of 0.2. (B) Y. intermedia, Y. frederiksenii, Y. kristensenii, and Y. pseudotuberculosis were tested in a similar manner. Cultivation was performed in the absence or presence of arabinose as indicated. The standard deviations from three independently performed experiments are shown.

When the plasmids were tested in other Yersinia species, a significant growth retardation upon HE⁺ induction was observed in Y. intermedia (P = 0.0399) and Y. frederiksenii (P = 0.0171) (Fig. 3B). In Y. pseudotuberculosis, we monitored a strong decrease of the OD₆₀₀ from 0.82 to 0.49, and growth did not resume during growth monitoring for 48 h. The strongest effect of HE⁺ was observed with Y. kristensenii (Fig. 3B) and S. Typhimurium (File S2), whose growth arrested immediately after the addition of arabinose and resumed 15 h later only in the case of the S. Typhimurium culture. Growth retardation to different extents was also observed with Y. ruckeri (P = 0.0404), Y. aldovae (P = 0.0042), Y. bercovieri (P = 0.0441), and Y. mollaretii (P = 0.0252) (File S2). When the construct lacking yRz and yRz1 (HE) was used, no qualitative deviations from the phenotypes seen with pBAD-HE⁺ were observed (data not shown). The nonrecombinant plasmid pBAD33, used as a control, exhibited no effect on the growth behavior of all strains tested in medium with or without arabinose. The addition of arabinose stimulated the growth of most strains, as observed previously (17).

Because Tc-PAI_Ye is present only in Y. enterocolitica biovars 2 to 5, Y. pseudotuberculosis, and Y. mollaretii but absent from all other strains investigated here, there is no correlation between the presence of Tc-PAI_Ye and the lytic capacities of the HE cassette. Taken together, a temperature-dependent activity of the lysis cassette was observed only in Y. enterocolitica strains, pointing to a species-specific mechanism underlying this lysis phenomenon.

Lon protease contributes to temperature-dependent activity of the lysis cassette.

The proteases Lon and ClpP are known to contribute to temperature-dependent gene regulation, as exemplified by RovA (21, 22). In particular, RovA acts as a proteinaceous thermometer whose conformation at body temperature results in a higher susceptibility to the Lon protease. Therefore, we tested whether or not these two proteases play a role in the lysis phenotypes described above. Nonpolar deletion mutants of their genes were constructed, and the resulting W22703 Δlon and W22703 ΔclpP strains were transformed with pBAD33-HE. Compared with results shown in Fig. 1A, the growth curve of W22703 ΔclpP/pBAD33-HE did not exhibit a strong phenotype in preliminary experiments (File S3).

The growth behavior of strain W22703 Δlon/pBAD33-HE at 37°C was monitored in the absence or presence of arabinose (Fig. 4A). Cell proliferation was stopped upon HE overexpression due to the addition of arabinose. Compared with the graphs shown in Fig. 1A, the effect of HE overexpression was more pronounced, as growth did not restart until at least the end of the experiment, 14 h after inoculation. However, when the same strain was incubated at 15°C, arabinose-induced overexpression of HE resulted in growth arrest and cell lysis until at least 27 h after inoculation, when the noninduced culture reached the stationary phase. As a control, the growth of strain W22703 Δlon/pBAD33-HE was not restricted in the absence of arabinose (Fig. 4B). This activity of the lysis cassette at 15°C in the absence of lon is in clear contrast to the data shown in Fig. 1B. To further validate this finding, strain W22703 Δlon/pBAD33-HE was equipped with plasmid pBR322-lon to complement the lack of the lon gene. We observed that cell lysis of strain W22703 Δlon/pBAD33-HE at low temperature and with arabinose was completely abolished by adding the medium-copy-number plasmid pBR322 carrying lon (Fig. 4B). These data indicate that the Lon protease contributes to the lysis effect at 15°C, possibly by degrading the holin and/or the endolysin. In the following section, we focus on endolysin, because holin is a relatively small protein of approximately 12 kDa that is hardly separated or visible on SDS gels.

FIG 4.

Effect of holin-endolysin overexpression in Y. enterocolitica W22703 cells in the absence of protease Lon. (A) Using the plasmid pBAD33-HE, the overexpression of holin and endolysin in the W22703 Δlon mutant was tested at 37°C in three biologically independent experiments. The resulting strain was transformed with pBR322-lon for complementation of the lon deletion. As controls, both strains were grown in the absence of arabinose. Growth was monitored in 50 ml LB medium under shaking, and 0.2% arabinose was added at an OD₆₀₀ of 0.2. (B) The growth experiment was performed at 15°C. Controls with nonrecombinant plasmid pBR322 are shown in File S6. Three biologically independent cultures were measured at both temperatures.

Longer half-life of ElyY in the absence of Lon protease.

To test the hypothesis that the Lon protease targets ElyY, we cultivated the strains W22703/pET28b-elyY/pTARA and W22703Δlon/pET28b-elyY/pTARA in LB medium at 30°C until an OD₆₀₀ of 0.6. The helper plasmid pTARA provides T7 RNA polymerase required for transcription of elyY, which was induced with 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG). The cultures were incubated at 30°C for 4 h and split into four aliquots, two of which were further cultivated at 15°C and 37°C, respectively. To inhibit protein biosynthesis, 50 µg/ml tetracycline was added to one culture at each temperature. Samples of 1 ml each were taken after 15, 30, 60, 90, 120, 180, 240, and 300 min, and the amounts of ElyY-His₆ were determined by Western blotting.

In the absence of tetracycline, a steady amount of endolysin was measured for 5 h at both temperatures. After adding the antibiotic to the culture of strain W22703/pET28b-elyY/pTARA, successive degradation of ElyY was detected to 20.7% at 37°C and to 58.8% at 15°C with respect to the amount measured at time zero (Fig. 5A). In the case of strain W22703Δlon/pET28b-elyY/pTARA lacking the Lon protease, almost no reduction of ElyY amounts was observed at either temperature, indicating that the Lon protease is responsible for ElyY degradation (Fig. 5B). The short half-life of the endolysin ElyY at 37°C could explain why, at this temperature, growth of W22703 restarts following lysis (Fig. 1), in contrast to that of the W22703 Δlon strain at the same temperature.

FIG 5.

Stability of ElyY. Cultures of Y. enterocolitica W22703/pET28b-elyY/pTARA (A), W22703 Δlon/pET28b-elyY/pTARA (B), and Y. pseudotuberculosis (C) were grown in 200 ml LB medium to exponential phase (OD₆₀₀ of 0.6) at 30°C before IPTG was added to a final concentration of 1 mM. After incubation for an additional 3 h, the cultures were divided into four samples, and tetracycline (50 µg/ml) was added to two of them to inhibit protein biosynthesis. The samples were then incubated at 37°C or at 15°C. Aliquots were taken at the indicated time points, and equilibrated amounts of total protein were loaded onto an 18% Tricine-SDS polyacrylamide gel. The amount of ElyY was quantified by Western blotting. Three independent experiments were performed for each temperature.

The stability of Lon protease was investigated in the same manner. At 15°C, nearly identical amounts of the proteolytic enzymes were monitored in the presence or absence of tetracycline. In contrast, a rapid decline of the protease amount was detected at 37°C until complete degradation 120 min after the addition of the antibiotic (File S4). These data support the hypothesis that different stabilities of the Lon protease contribute to the temperature-dependent lysis of Y. enterocolitica cells upon HE overexpression.

Finally, we tested the stability of ElyY overexpressed in Y. pseudotuberculosis, in which lysis occurred at both temperatures upon overexpression of HE (Fig. 3B) (17). We transformed the plasmids pET15b-elyY and pTARA into Y. pseudotuberculosis and performed the same experiment as that described above. Following the addition of tetracycline, no significant decrease of the amount of ElyY was observed at either 15°C (P = 0.2231) or 37°C (P = 0.3054) (Fig. 5C). This result is in contrast to the findings made with Y. enterocolitica strain W22703 in Fig. 5A and B but in line with the lysis of Y. pseudotuberculosis upon HE overexpression at 15°C (Fig. 3B).

DISCUSSION

Members of the genus Yersinia are well known to regulate their virulence toward mammals in a temperature-dependent manner. For example, the Yop effector proteins and the adhesin YadA are produced at body temperature (23, 24). In Y. pestis and in Y. pseudotuberculosis, the transcript of the virulence regulator gene lcrF serves as an RNA thermometer. While the mRNA of lcrF forms a secondary structure that masks the Shine-Dalgarno sequence in a stem-loop at 25°C, it melts at higher temperatures and allows the translation of the lcrF mRNA (22, 25). Another example is the transcription of the invasion gene inv, which is activated by the regulatory protein RovA at 25°C in both Y. enterocolitica and Y. pseudotuberculosis (26–28). Levels of active RovA are low at 37°C and high at 25°C due to the characteristic of RovA as a proteinaceous thermometer that undergoes conformational changes at 37°C, reducing its DNA-binding capacity and rendering it more susceptible to degradation by the Lon protease (21).

In contrast, a large part of the Yersinia gene set is transcriptionally activated at temperatures below 37°C, pointing to environmental surroundings that yersiniae encounter during their life cycles (9, 11). For example, genes of yersiniae that are more strongly expressed at temperatures below 25°C than at 37°C are involved in the synthesis of lipopolysaccharides, iron scavenging systems, and biofilms (29–31). Low-temperature-induced determinants of pathogenic Yersinia species are also involved in chemotaxis, central metabolism, anaerobic respiration, hemin storage, amino acid metabolism, and substrate transport (12, 32, 33). The most prominent example is the Tc of Y. enterocolitica that is responsible for the insecticidal and nematocidal activity of strain W22703 (7, 8). The tc genes, which are located on the pathogenicity island Tc-PAI_Ye, are transcriptionally silenced at 37°C but maximally activated at 10°C to 15°C. The underlying molecular mechanism is based on an antagonism between the inducer TcaR2 and the repression complex YmoA/H-NS (15, 16). In contrast to inv regulation by RovA, the conformation of TcaR2 is not affected by temperature changes. The regulatory model proposes a higher oligomerization rate of H-NS at lower temperature, allowing the YmoA/H-NS complex to interact with multiple DNA binding sites of the tc promoters, thereby forming a DNA bridge that blocks insecticidal gene transcription at low temperatures (15).

The four phage-related genes located on the insecticidal island are a biologically functional holin-endolysin system upon overexpression via the arabinose-inducible vector pBAD at 37°C (17) and, therefore, might play a role in the release of toxic factors. Here, we performed growth assays at 15°C and observed lysis in different Yersinia species and in E. coli, but not in Y. enterocolitica strains, following HE overexpression. An explanation is that the murein of Y. enterocolitica, the primary target of ElyY (17), is resistant to HE activity at low temperature. Another possibility tested here is a temperature-dependent proteolytic degradation of components of the lysis cassette. Indeed, a lack of the Lon protease resulted in growth arrest of strain W22703 at 15°C that had not been observed for the parental strain. At 37°C, the lack of lon intensified the phenotype of growth arrest and cell lysis, as monitored in Fig. 1A. Both of these lysis phenotypes were abolished when the lon gene was provided in trans via plasmid pBR322-lon, which has a middle copy number (34). These findings led to the assumption that the Lon protease degrades the holin and/or the endolysin at 15°C but not at 37°C. The data provided in Fig. 5 do not indicate a higher stability of ElyY at 37°C, but the Lon protease seems to be less stable at higher temperature. Although this effect seems to be compensated by sufficient protease expression, as demonstrated in File S4 in the supplemental material, a higher Lon protease stability at 15°C might contribute to a lack of W22703 cell lysis despite HE overexpression.

At both temperatures tested here, the overexpression of ElyY revealed the high stability of this enzyme in Y. pseudotuberculosis, in contrast to its increased proteolytic susceptibility in Y. enterocolitica strain W22703 at both temperatures, an observation that is in line with the lytic phenotypes of Y. pseudotuberculosis at 15°C (Fig. 3B) and 37°C (17). The reason for the longer half-life of ElyY in Y. pseudotuberculosis is unknown. In Y. pseudotuberculosis, the Lon protease is significantly more strongly expressed at 37°C than at 25°C (21). With respect to RovA, it is assumed that thermally induced conformational changes and/or DNA binding blocks degradation of RovA by Lon and Clp proteases at 25°C and that partial defolding/inactivation of the RovA protein further improves cofactor binding and allows better access of the proteases (21). Given the overall identity of the Lon proteins from Y. enterocolitica and Y. pseudotuberculosis, additional factors rather than substrate specificity might be responsible for the different half-lives in the two species. A difference between the two pathogens in protein stability was also observed with YmoA, which is stable in Y. enterocolitica strain W22703 at both 15°C and 37°C (15) but degraded in Y. pseudotuberculosis at body temperature (25).

Native conditions under which the HE cassette of Tc-PAI_Ye is expressed have not yet been identified. This prompted us to establish an artificial system based on the overexpression of the lytic genes that, despite its limitations, allowed several conclusions. (i) The Lon protease degrades the endolysin ElyY. (ii) Adding higher copy numbers of Lon via pBR322 prevents cell lysis completely, suggesting that the protease has similar activity at 15°C and 37°C. (iii) The Lon protease is less stable at 37°C than at 15°C, a finding that points to a conformational change, as observed for RovA. (iv) Y. enterocolitica cells are less susceptible to HE-mediated cell lysis at 15°C than at 37°C. Under native conditions, cell lysis at 37°C is probably prevented by a tight repression of the HE cassette. To conclude, we propose that the HE cassette is induced only during infection under environmental conditions and contributes to the insecticidal activity in a manner that does not require cell lysis for Tc release.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

Bacterial strains and oligonucleotides used in this study are listed in Table 1 and in File S5 in the supplemental material, respectively. Cultures were grown in LB broth (containing 10 g/liter tryptone, 5 g/liter yeast extract, and 5 g/liter NaCl) or on LB agar (LB broth supplemented with 1.5% agar). E. coli was grown at 37°C and Yersinia spp. at 30°C or as indicated. For growth studies, overnight cultures were diluted to an OD of 0.07 in 200-ml flasks with 50 ml medium and incubated under vigorous shaking or in microtiter plates with 200 µl per well. If necessary, kanamycin (50 µg/ml), streptomycin (50 µg/ml), chloramphenicol (20 µg/ml), tetracycline (12 µg/ml), nalidixic acid (20 µg/ml), ampicillin (150 µg/ml), IPTG (1 mM), or 0.2% (wt/vol) arabinose was added to the medium.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Genotype or relevant feature(s)	Reference or source
E. coli
LMG194	F− ΔlacX74 galE thi rpsL ΔphoA (PvuII) Δara714 leu::Tn10	Invitrogen, Karlsruhe, Germany
SM10	lacY tonA recA Muc+ thi thr leu supE RP4-2−Tc::Mu Kmr λpir	35
Y. enterocolitica
W22703	Nalr Res− Mod+ pYV−	38
W22703 Δlon	Nonpolar deletion mutant of lon encoding ATP-depending protease Lon	This study
SZ4331/97	Biovar 1A, O:14, Tc-PAIYe negative	Institut für Hygiene und Umwelt, Hamburg, Germany
H692/94	Biovar 2, O:9, Tc-PAIYe positive	Institut für Hygiene und Umwelt, Hamburg, Germany
H230/89	Biovar 3, O:5, 27, Tc-PAIYe positive	Institut für Hygiene und Umwelt, Hamburg, Germany
519-36/88	Biovar 5, O:2a, 2b, 3	Institut für Hygiene und Umwelt, Hamburg, Germany
Yersinia spp.
Y. pseudotuberculosis	CIP 55.85	Collection Institute Pasteur
Y. intermedia	CIP 80.28	Collection Institute Pasteur
Y. kristensenii	CIP 80.30	Collection Institute Pasteur
Y. frederiksenii	CIP 80.29	Collection Institute Pasteur
Y. ruckeri	CIP 80.20	Collection Institute Pasteur
Y. aldovae	CIP 103162	Collection Institute Pasteur
Y. bercovieri	CIP 103323	Collection Institute Pasteur
Y. mollaretii	CIP 103324	Collection Institute Pasteur
Plasmids
pKNG101	Conditionally replicating vector; R6K origin, mobRK2 transfer origin, sucrose-inducible sacB, Strr	37
pBAD33	Arabinose-inducible promoter, Camr	20
pBAD33-HE	772-bp fragment comprising the holin and endolysin genes cloned into pBAD33 via SacI and XbaI	17
pBAD33-HE+	1,310-bp fragment comprising the HE cassette and yRz and yRz1 cloned into pBAD33 via SacI and XbaI	17
pET28b	Expression vector; IPTG-inducible T7lac promoter, N- and C-terminal polyhistidine tag, Kanr lacI; pBR322 origin	Merck, Darmstadt, Germany
pET28b-elyY	Plasmid for endolysin gene overexpression; elyY cloned into IPTG-inducible pET28b via NcoI and XhoI; C-terminal polyhistidine tag	17
pET15b	Expression vector; T7 promoter, N-terminal polyhistidine tag, Ampr lacI; pBR322 origin	Merck, Darmstadt, Germany
pET15b-elyY	Plasmid for endolysin gene overexpression; elyY cloned into IPTG-inducible pET15b via NdeI and BamHI; N-terminal polyhistidine tag	This study
pTARA	pBAD33-derived plasmid carrying the gene for arabinose-inducible T7 RNA polymerase; Camr	Addgene, Cambridge, MA
pBR322	Ampr Tetr	39
pBR322-lon	Plasmid to complement gene deletion of protease Lon; lon cloned into pBR322 via PstI; Amps Tetr	This study

Standard procedures.

DNA manipulation and chromosomal DNA isolation were performed according to standard procedures (34) or to the manufacturer's protocol. PCR was performed using Thermoprime Taq polymerase (ABgene, Hamburg, Germany) via the following protocol: 95°C for 2 min; 30 cycles at 95°C for 10 s, annealing temperature depending on oligonucleotides for 30 s, and 72°C for 20 to 180 s depending on the expected fragment length; and final extension at 72°C for 10 min. Chromosomal DNA (100 ng) was used as the template.

Conjugational transfer was performed using the mobilizing E. coli strain SM10 as the donor for matings (35, 36). The T7 RNA polymerase-expressing helper plasmid pTARA was used to allow pET-plasmid driven gene expression within Yersinia spp.

Protein biosynthesis of bacterial cultures in exponential phase was stopped by adding 50 µg/ml tetracycline. Subsequently, cultures were incubated at either 15°C or 37°C, and samples were taken at the indicated time points. Protein concentration was determined using RotiQuant solution (Carl Roth GmbH, Karlsruhe, Germany) based on Bradford's method (37).

Student's t test was applied for statistical evaluations; P values of ≤0.05 were considered significant.

Recombinant plasmids and strains.

The plasmid pET15b-elyY was generated using NdeI and BamHI. The in-frame deletion of lon was performed using the suicide plasmid pKNG101, into which two fragments flanking lon were cloned. Strain W22703 Δlon then was generated via a previously described procedure (8). The deletion was confirmed by sequencing. All oligonucleotides used for these constructs are listed in File S5.

Immunoblotting.

To prepare Y. enterocolitica protein extracts, culture samples were pelleted at 13,200 rpm for 1 min. The pellet was boiled in 100 μl 1× Laemmli buffer for 5 min at 100°C and stored at −20°C. After centrifugation at 4°C, aliquots of the supernatant were subjected to 18% Tricine-SDS polyacrylamide gel electrophoresis, and separated proteins were transferred to a polyvinylidene difluoride membrane (Merck, Darmstadt, Germany) with a semidry blot apparatus. The prestained protein ladder of Fermentas/ThermoFisher Scientific, Darmstadt, Germany, was used as a marker. If appropriate, the membrane was blocked overnight at 4°C in Tris-buffered saline (TBS) with 0.1% Tween 20 containing 5% milk powder. The antibodies (primary antibodies, monoclonal mouse anti-His₆ tag antibody [Invitrogen], polyclonal rabbit anti-LONP1 [Sino Biological], and monoclonal mouse anti-GroEL [Enzo Life Sciences]; secondary antibodies, horeseradish peroxidase [HRP]-conjugated goat anti-mouse antibody [Dianova, Hamburg, Germany; ThermoFisher] and HRP-conjugated goat anti-rabbit antibody [ThermoFisher Scientific]) were diluted 1:2,000 or 1:10,000 in TBS with Tween 20 (TBST) containing 10% fetal calf serum (FCS). Proteins were detected using the SuperSignal West Pico chemiluminescent substrate (ThermoFisher, Waltham, MA). The signals were monitored with a Lumina in vivo imaging system (IVIS) and quantified with Living Image 3.0 software (Xenogen, CA).