Development of the recombinase-based in vivo expression technology in Streptococcus thermophilus and validation using the lactose operon promoter


  • M. Junjua,

    1. Unité de Recherche, ‘Animal & Fonctionnalités des Produits Animaux’, Équipe ‘Protéolyse et Biofonctionnalités des Protéines et des Peptides’, UC INRA 340, Université de Lorraine, Vandœuvre-lès-Nancy, France
    Search for more papers by this author
  • W. Galia,

    1. Unité de Recherche, ‘Animal & Fonctionnalités des Produits Animaux’, Équipe ‘Protéolyse et Biofonctionnalités des Protéines et des Peptides’, UC INRA 340, Université de Lorraine, Vandœuvre-lès-Nancy, France
    Search for more papers by this author
  • N. Gaci,

    1. Unité de Recherche, ‘Animal & Fonctionnalités des Produits Animaux’, Équipe ‘Protéolyse et Biofonctionnalités des Protéines et des Peptides’, UC INRA 340, Université de Lorraine, Vandœuvre-lès-Nancy, France
    Search for more papers by this author
  • O. Uriot,

    1. Unité de Recherche, ‘Animal & Fonctionnalités des Produits Animaux’, Équipe ‘Protéolyse et Biofonctionnalités des Protéines et des Peptides’, UC INRA 340, Université de Lorraine, Vandœuvre-lès-Nancy, France
    Search for more papers by this author
  • M. Genay,

    1. Unité de Recherche, ‘Animal & Fonctionnalités des Produits Animaux’, Équipe ‘Protéolyse et Biofonctionnalités des Protéines et des Peptides’, UC INRA 340, Université de Lorraine, Vandœuvre-lès-Nancy, France
    Search for more papers by this author
  • H. Bachmann,

    1. NIZO food research, Health Department, Ede, the Netherlands
    2. VU University Amsterdam, Systems Bioinformatics, Faculty of Earth and Life Sciences, Amsterdam, the Netherlands
    Search for more papers by this author
  • M. Kleerebezem,

    1. NIZO food research, Health Department, Ede, the Netherlands
    2. Wageningen University, Host Microbe Interactomics Group, Wageningen, the Netherlands
    Search for more papers by this author
  • A. Dary,

    1. Unité de Recherche, ‘Animal & Fonctionnalités des Produits Animaux’, Équipe ‘Protéolyse et Biofonctionnalités des Protéines et des Peptides’, UC INRA 340, Université de Lorraine, Vandœuvre-lès-Nancy, France
    Search for more papers by this author
  • Y. Roussel

    Corresponding author
    1. Unité de Recherche, ‘Animal & Fonctionnalités des Produits Animaux’, Équipe ‘Protéolyse et Biofonctionnalités des Protéines et des Peptides’, UC INRA 340, Université de Lorraine, Vandœuvre-lès-Nancy, France
    • Correspondence

      Yvonne Roussel, Faculté des Sciences et

      Technologies, B.P. 70239 54506 Vandœuvre-lès-Nancy Cedex, France.


    Search for more papers by this author



To construct and validate the recombinase-based in vivo expression technology (R-IVET) tool in Streptococcus thermophilus (ST).

Methods and Results

The R-IVET system we constructed in the LMD-9 strain includes the plasmid pULNcreB allowing transcriptional fusion with the gene of the site-specific recombinase Cre and the chromosomal cassette containing a spectinomycin resistance gene flanked by two loxP sites. When tested in M17 medium, promoters of the genes encoding the protease PrtS, the heat-shock protein Hsp16 and of the lactose operon triggered deletion of the cassette, indicating promoter activity in these conditions. The lactose operon promoter was also found to be activated during the transit in the murine gastrointestinal tract.


The R-IVET system developed in ST is relatively stable, functional, very sensitive and can be used to assay activity of promoters, which are specifically active in in vivo conditions.

Significance and Impact of the Study

This first adaptation of R-IVET to ST provides a highly valuable tool allowing an exploration of the physiological state of ST in the GIT of mammals, fermentation processes or dairy products.


Streptococcus thermophilus is a lactic acid bacterium widely used as a starter in the dairy industry. It is traditionally used in combination with other lactic acid bacteria such as Lactobacillus delbrueckii subsp. bulgaricus in yogurt or Lactobacillus helveticus in hard cheeses (Tamime and Deeth 1980; Fox 1993). Although Strep. thermophilus is phylogenetically related to pathogenic streptococcal species such as S. pneumonia, S. pyogenes or S. mutans and to the commensal species S. salivarius (Mitchell 2003; Tettelin 2004), ST has never been associated with pathogenic incidences. On the contrary, the very long history of its use in food industry led to the GRAS (generally recognized as safe) status assignment by the American Food and Drug Administration and the QPS (qualified presumption of safety) status by the European Food Safety Authority (EFSA). While health claims were attributed to live yoghurt cultures to improve lactose digestion in 2010 by the EFSA, the status of Strep. thermophilus as a probiotic (Guarner et al. 2005) has still to be proven. Streptococcus thermophilus is found in the faeces of humans after consumption of yogurts (Brigidi et al. 2003; Mater et al. 2005; Elli et al. 2006). Moreover, metagenomic sequencing performed on faecal specimen of 124 Europeans have shown the presence of ST DNA in the faeces of 90% of the individuals (Qin et al. 2010). However, before ST strains can be credited with potential probiotic allegations, further studies should be performed to know their physiology in the GIT and their aptitude to exert probiotic functions in situ.

The complete genome sequence of six strains of ST provides a rich source of information from which genes expressed in the GIT can be discovered. The use of postgenomics technologies combined with appropriate animal models and in vitro human systems ultimately allows the understanding of the complex interactions of the bacterial cells with the human host cells and with the rest of the microbiota (de Vos et al. 2004). A comprehensive investigation of the physiological state of ST when passing through the GIT of gnotobiotic rats has already been performed by proteomic analysis (Rul et al. 2011; Thomas et al. 2011), revealing a massive use of the glycolysis pathway. As a complementary tool to these protein data, we decided to use the recombinase-based in vivo expression technology (R-IVET) to explore transcriptional responses of ST when bacterial cells are present in complex environments such as the GIT. R-IVET is a promoter-trapping technology, which was devised to identify genes, which are specifically activated in particular complex environments such as infected tissues, GIT, fermented food or soils (Rediers et al. 2005). It is based on a transcriptional fusion between a promoterless recombinase gene and random genomic DNA fragments, which are tested for their promoter functionality (Camilli et al. 1994). R-IVET systems are generally composed of two different elements: the recombinase transcriptional fusion located on a plasmid and the chromosomal target module that contains one or several reporter genes which are deleted by the action of the recombinase produced upon the transcription of the cloned promoter. As a result of the promoter activity, the bacterial cell is irreversibly labelled and can be analysed at any later collection time. R-IVET systems, initially developed to reveal virulence functions induced when pathogenic bacteria infected host tissues (Angelichio and Camilli 2002), have been applied to lactic acid bacteria to detect genes which are specifically activated in the GIT of mice (Bron et al. 2004) or in fermented dairy environment (Bachmann et al. 2010). This system applied to Lactobacillus plantarum led to the identification of 72 genes specifically induced in the mouse GIT (Bron et al. 2004). Another valuable utilization of the R-IVET system is the transcription monitoring of a selected gene as a function of time in a specific environment (Angelichio and Camilli 2002; Rao et al. 2008).

In the present study, we describe the construction of the two genetic components of the R-IVET system in the Strep. thermophilus strain LMD-9. After testing the stability of the genetic tool, we investigated its functionality using three different Strep. thermophilus promoters (PprtS, Phsp and Plac) and analysed whether the chromosomal antibiotic cassette could be deleted under different growth conditions in M17 medium. As a proof of concept for a subsequent use in complex environments, the R-IVET system containing the promoter of the lac operon (Plac) was tested in the GIT of mice.

Material and methods

Bacterial strains, media, growth and transformation conditions

The strains used in this work are listed in Table 1. The Strep. thermophilus strain used to implement the R-IVET system is LMD-9 from the American Type Culture Collection (Manassas, VA). Streptococcus thermophilus strains were grown in anaerobic conditions (AnaeroGen, Oxoid, Basingstoke, UK) at 42°C in M17 medium (Terzaghi and Sandine 1975) supplemented with 2% (w/v) lactose (LM17) or 0·5% (w/v) glucose (GM17). When needed, antibiotics purchased from Sigma (Saint Quentin Fallavier, France) were added at the concentration of 300 μg ml−1 for spectinomycin or 5 μg ml−1 for erythromycin. Growth kinetics of Strep. thermophilus strains were monitored by viable counts (CFU ml−1) on M17 agar or by measuring OD using the Bioscreen C Microbiology Reader (Thermo Fisher Scientific, Saint-Herblain, France). Bioscreen wells were loaded with 350 μl of culture adjusted to an OD600 nm value of 0·05 and covered with 50 μl of oil to reduce the oxygen tension. Values of OD600 nm recorded every 15 min for 6 h were used to determine generation time as log2 divided by the slope of the growth curve.

Table 1. Bacterial strains and plasmids used in the present study
 Relevant markers and characteristicsReference or source
  1. a

    SpecR CmR, ErmR: resistance to spectinomycin, chloramphenicol and erythromycin, respectively.

  2. b

    cre: gene encoding the Cre recombinase deprived of its own promoter.

Streptococcus thermophilus
 LMD-9Wild-type strain for which genome sequence is availableATCC BAA-491, Makarova et al. (2006)
 STUL5001SpecRa; LMD-9 derivative containing the loxP-specR-loxP cassette in the STER_0891 locusThis study
Escherichia coli
 TOP10One Shot® TOP10 chemically competent E. coli cellsInvitrogen
pSET4sSpecRa; replication function of pG+host3 and pUC19Takamatsu et al. (2001)
pNZ5520CmRa; pNZ18 derivative containing a linker with an BglII restriction site upstream of ‘creb (promoterless cre gene)Bachmann et al. (2008)
pG+host9TRErmRa; pWV01 derivative, with thermoresistant replication function [repA (Tr)]Maguin et al. (1996)
pULNcreAErmRa; pG+host9TR derivative containing ‘creb with the Lactococcus lactis las operon transcriptional terminator (Tlas) upstream of ‘cre, with thermoresistant replication functionThis study
pULNcreBpULNcreA with the transcriptional terminator of the Lactococcus lactis aminopeptidase N gene (TpepN) downstream of ‘crebFig. 2, this study
pULNcreB-PshsppULNcreB derivative containing the promoter of the gene encoding the heat-shock protein Hsp16 (Pshsp) cloned in the BglII site upstream of ‘crebThis study
pULNcreB-PlacpULNcreB derivative containing the promoter of the lactose operon (Plac) cloned in the BglII site upstream of ‘crebThis study
pULNcreB-PprtSpULNcreB derivative containing the promoter of the gene encoding the PrtS protease (PprtS) cloned in the BglII site upstream of crebThis study

Competent cells of Strep. thermophilus LMD-9 were prepared by growing them in a chemically defined medium (CDM) as described by Gardan et al. (2009). Cells were grown overnight at 42°C in CDM after which the culture was diluted in fresh CDM and further grown to a final value OD600 nm of 0·2. Competent cells were mixed with 250–500 ng of plasmid DNA or 20–50 ng of linear DNA per ml of culture and further incubated at 42°C for 1 h before plating onto solid M17 medium containing the appropriate antibiotic for selection. For long-term storage at −80°C, competent cells were centrifuged at 8000 g during 5 min, resuspended in a solution (1/10 volume of the initial culture) of CDM containing 14% glycerol and frozen in liquid nitrogen.

The Escherichia coli strain TOP10 (Invitrogen, Breda, the Netherlands) was used as an intermediate cloning host. Growth was performed in LB medium in aerobic conditions at 37°C with erythromycin added at the concentration of 400 μg ml−1 when needed.

DNA extraction, cloning and sequence analysis

Plasmid DNA was isolated from E. coli cells with a Miniprep Kit (Fermentas, Villebon sur Yvette, France) according to the manufacturer's instructions. Strep. thermophilus genomic DNA was extracted from 50 ml overnight LM17 culture according to Fischer and co-workers (Fischer et al. 1997) and was used as target DNA for polymerase chain reactions (PCRs). DNA fragments were separated by electrophoresis in agarose gels using 0·5× TBE buffer at 100V. 1-kb and 100-bp DNA ladders (Fermentas) were used as molecular weight markers. Gels were stained with ethidium bromide and imaged using a GelDoc-It Imaging System (Bio-Rad, Marne-la-Coquette, France). Restriction endonucleases (Fermentas), the T4 DNA ligase and the calf intestinal alkaline phosphatase (CIAP) (New England Biolabs, Leusden, the Netherlands) were used according to the manufacturer's recommendations. DNA fragments used for the ligation reactions were purified using the High Pure DNA purification Kit (Roche Molecular Biochemicals, Mannheim, Germany) before and after endonuclease restriction. Ligation reactions were performed in 10 μl at 16°C overnight with 50 ng of digested and CIAP-dephosphorylated vector and 15 ng of insert, corresponding to a ratio of five copies of insert for one copy of vector. Nucleotide sequences were determined from PCR products on both strands by the company Beckman Coulter Genomics (Takeley, UK). Sequence comparisons and alignments were performed with the BioEdit software ( using the ClustalW algorithm.

PCR amplifications

Polymerase chain reactions were carried out in a Mastercycler pro thermocycler (Eppendorf, Hambourg, Germany). Oligonucleotides used as primers were purchased from Eurogentec (Seraing, Belgium), and their sequences are described in Table 2.

Table 2. Oligonucleotides used in the present study with their relevant characteristics
#Sequence 5′→3′Characteristics


  3. a

    Underlined sequences indicate restriction sites used in cloning procedures. Sequences shown in bold and italics correspond to loxP sites.

1ATCCGTAAAACCATCAAATCTTAATTConstruction of the floxed specR cassette, 3616-bp amplicon with #4 for the undeleted cassette or 2382-bp for the deleted cassette
2TGACTCCCCGTCGTGTAGATAACTAGAmplification of the spectinomycin resistance gene from pSET4S with #3 (1201-bp amplicon)
5ATAATAACTAGTGAGAGGCCGCCACGGCGATGTGCaSpeI, construction of pULNcreA, 1457-bp amplicon with #6
7ATATATGGTACCTCGCTTTGATTGTTCTATCGaKpnI, construction of pULNcreB, 159-bp amplicon with #8
9CCATTTTGAACGATGACCTCConfirmation of creB presence, 2041-bp amplicon with #10
11ATAATAAGATCTTCATGAATGTCTTGCTCACTCCaBglII, construction of pULNcreB-PprtS, 211-bp amplicon with #12
13ATAATAAGATCTGGACGAACTAAAATAGTGACGaBglII, construction of pULNcreB-Pshsp, 171-bp amplicon with #14
15ATAATAAGATCTACGATGGATGATATCGAAGaBglII, construction of pULNcreB-Plac, 324-bp amplicon with #16
17ACACAAAGTCGCTGTTCTCGSequencing of the R-IVET cassette
18TGCAAGTAAAATTGCACCTGTTSequencing of the R-IVET cassette
19ATCTTCCGCTTTAGTTTTTGGCSequencing of the R-IVET cassette
20CCATGGCAATTATTTGGTTACGSequencing of the R-IVET cassette
21TGGTACCGTGGAATCATCCTConfirmation of specR presence, 361-bp amplicon with #22

Analytical PCRs were carried out in 20 μl reaction mixture containing 0·5 U of Taq DNA polymerase (Fermentas), 200 μmol l−1 of each dNTP, 25 μmol l−1 of each primer, 2 μl of the 10× concentrated Fermentas buffer, 1·5 mmol l−1 MgCl2 and with 2 μl of colony lysate obtained by heating colony cells for 3 min in a microwave oven at maximum power or with 50 ng of purified DNA. Temperature cycle conditions were as follows: 3 min of denaturation at 95°C; 30 cycles of three steps of 30 s each (denaturation step at 95°C, annealing step at 54°C and extension step at 72°C); and a final extension at 72°C for 10 min.

Preparative PCRs were performed to produce DNA fragments intended for subsequent cloning or sequencing reactions. PCR was performed in 50 μl reaction mixture with 2·5 U of AccuTaq polymerase (Sigma-Aldrich, St Louis, MO, USA) using the analytical PCR cycle conditions but with extension steps carried out at 68°C.

Construction of the floxed spectinomycin resistance fragment for chromosomal integration

The strategy used to construct the Strep. thermophilus strain containing the floxed specR gene in the chromosome is presented in Fig. 1. In the first series of PCR (Fig. 1a), the three fragments UP-loxP, specR and loxP-DN were amplified individually with primer couples #1#24, #2#3 and #23#4, respectively. Each reaction was performed in a final volume of 50 μl reaction mixture with 1 U of Phusion high-fidelity DNA polymerase, 10 μl of Phusion 5× buffer (Finnzymes, Thermo Scientific, Espoo, Finland), 200 μmol l−1 of each dNTP, 0·5 μmol l−1 of each primer and 100 ng of template DNA. Genomic DNA of the strain LMD-9 was used as template to amplify the UP-loxP and the loxP-DN fragments. The plasmid pSET4S (Table 1, Takamatsu et al. 2001) was used to amplify the specR gene encoding a spectinomycin adenyltransferase. For the three reactions, cycling conditions were 1 min of initial denaturation at 98°C; 30 cycles of three steps (20 s denaturation at 98°C, 30 s annealing and 30 s extension at 72°C) and a final extension at 72°C for 10 min. Specific annealing temperatures were set at 41·4°C, 60°C and 46·8°C for the amplifications of the three fragments UP-loxP, specR and loxP-DN, respectively.

Figure 1.

Construction of the Streptococcus thermophilus mutant STUL5001 containing the floxed specR cassette. The first set of PCRs are presented in (a) and the final overlapping PCR in (b). The resulting chromosomal region containing the floxed specR cassette obtained after transformation with the final overlapping PCR fragment is shown in (c). Numerical values indicated in italics in genes correspond to locus tags as described in the annotated sequence of the LMD-9 chromosome (NC_008532·1). In the STUL5001 mutant, the locus STER_0891 is truncated (t0891). Oligonucleotides used as amplification primers are indicated as black arrows and labelled with their # numbers as detailed in Table 2. UP and DN regions correspond to upstream and downstream sequences flanking the floxed specR cassette. specR corresponds to the gene which encodes for a spectinomycin adenyltransferase conferring resistance to spectinomycin. The back squares containing the grey arrow correspond to the loxP sequences.

In the second amplification phase (Fig. 1b), an overlapping PCR was carried out with primers #1#4 using the three fragments UP-loxP, specR and loxP-DN mixed together in equimolar concentration. Cycling conditions were 1 min of initial denaturation at 98°C; 50 cycles of three steps (20 s of denaturation at 98°C, 30 s of annealing and 2 min of extension at 72°C); and a final extension at 72°C for 10 min. A touchdown PCR was applied for the first 20 cycles with a 1°C decrease per cycle from 60°C to 40°C followed by a constant 42°C annealing temperature for the last 30 cycles. Thirty ng of the resulting amplified fragments were used for the transformation of LMD-9 competent cells. A spectinomycin-resistant clone confirmed to have the cassette at the STER_0891 locus by colony PCR with primers #1#4 was chosen and named STUL5001. Its floxed specR cassette and junction regions with the chromosome were sequenced to check that no mutation had occurred during the construction of the mutant.

Construction of the transcriptional fusion plasmid pULNcreB

To construct the cre transcriptional fusion vector suitable for Strep. thermophilus, the promoterless gene of the site-specific recombinase Cre was cloned into pG+host9TR, a thermoresistant replication derivative of the pG+host9 plasmid (Maguin et al. 1996; Table 1) carrying an erythromycin resistance gene as the selection marker. The fragment bearing, in this order, the las operon transcriptional terminator Tlas, a multiple cloning site and the promoterless cre gene was amplified from the plasmid pNZ5520 (Bachmann et al. 2008; Table 1) using primers #5#6 (Table 2), incorporating a SpeI restriction site at the 5′ end of the amplicon and a SalI site at its 3′ end. The PCR fragment was SpeI-SalI double-digested and cloned into the dephosphorylated and similarly digested pG+host9TR giving pULNcreA (Table 1). The terminator TpepN amplified from pNZ5520 (Bachmann et al. 2008; Table 1) using primers #7#8 was digested with both KpnI and SalI enzymes and cloned into the dephosphorylated and similarly digested pULNcreA leading to pULNcreB. The plasmid was completely sequenced on both strands, and a restriction map is given in Fig. 2.

Figure 2.

Restriction and genetic map of pULNcreB. Only restriction sites used for the construction of the plasmid are indicated on the circle. Six restriction sites are present in the multiple cloning site (MCS) located upstream of the promoterless ‘cre gene, (*) BglII was used to clone promoter sequences. Squares (□) represent the transcriptional terminators of the Lactococcus lactis lactose operon (Tlas) and of the aminopeptidase N gene (TpepN). The wild-type sequence of the replication gene repA [repA (Tr)] allows replication of the plasmid at temperatures above 37°C.

Assessment of the segregational stability of pULNcreB and the floxed specR cassette

The segregational stability of pULNcreB in the transformant strain STUL5001 was determined in a maintained exponential growing culture performed in LM17 broth without erythromycin. Samples were taken at different culture times corresponding to a defined number of generations and plated on LM17 agar without erythromycin. For each sampling time, 50–100 colonies were repicked on LM17 agar containing erythromycin. The segregational stability of pULNcreB was determined as the percentage of colonies resistant to erythromycin compared to the total number of colonies. For confirmation, 10 EryR colonies randomly chosen for each sampling time were analysed by PCR with primers #9#10 (Table 2) to ensure that the EryR phenotype resulted from the presence of the plasmid sequence and not from a spontaneous chromosomal EryR mutation.

Cloning of promoter sequences into the R-IVET vector pULNcreB

The promoter region of the Strep. thermophilus genes prtS, hsp16 and that of the lac operon were amplified from the genomic DNA of LMD-9 using the primer sets #11#12, #13#14 and #15#16, respectively (Table 2). All six primers were designed to contain a BglII restriction site at their 5′ extremity to clone the PCR fragments into the BglII cloning site of pULNcreB located upstream of the promoterless cre gene. PCR products of 211 bp, 171 bp and 324 bp were digested with BglII and ligated to the dephosphorylated and BglII-digested pULNcreB. First step cloning was performed in E. coli TOP10 and recombinant plasmids (Table 1) displaying the correct sequence were transferred to Strep. thermophilus by natural transformation where selection was carried out in M17 supplemented with 5 μg ml−1 erythromycin.

Animal experiments and preparation of ST cells for oral gavage

Mouse experiments were performed according to the guidelines N°86/609/CEE of the French Government. Mice had free access to standard mouse chow and water. We used conventional mice possessing their natural intestinal microbiota. In order to be able to specifically recover Strep. thermophilus cells containing pULNcreB or the R-IVET cassette from faeces, mice were selected in advance for their faecal flora sensitive to erythromycin and/or spectinomycin. This was performed by plating dilutions of fresh faeces onto LM17 agar supplemented by 5 μg ml−1 erythromycin or 300 μg ml−1 spectinomycin. To assess the stability of the R-IVET tool components in vivo, we used one six-week-old C57BL/6J mouse possessing a faecal flora sensitive to both erythromycin and spectinomycin. For the experiment testing Plac activity in ‘in vivo’ conditions, three 18-week-old Swiss mice were selected for their erythromycin-sensitive faecal flora and were given water supplemented with lactose at the concentration of 5% (w/v) during 20 h before the Strep. thermophilus administration and until the end of faeces collection. This was performed to increase Strep. thermophilus survival and activity of the lactose operon in the mouse GIT (Mater et al. 2006). For the mouse oral administrations, 200 μl of freshly prepared Strep. thermophilus (approximately 109 CFU) were introduced in the stomach with a bulb-tipped gastric gavage needle. For the Plac assay in mice, cells of STUL5001 pULNcreB-Plac were grown overnight at 37°C in GM17 supplemented with spectinomycin at 300 μg ml−1. When culture reached the OD600 nm of 0·6, a fraction of 20 μl corresponding to 109 CFU was collected and completed to 200 μl with cold M17 buffer without sugar and antibiotic. Cell suspensions were kept on ice until the mouse gavage performed within the following hour. For all faecal analyses, only fresh samples were collected and immediately homogenized in phosphate buffer. Appropriate serial dilutions were spread on LM17 or GM17 agar supplemented either with erythromycin or spectinomycin.


Construction of the loxP-specR-loxP reporter strain STUL5001 and assessment of its growth capacities and stability

The Strep. thermophilus strain LMD-9 was selected to implement the R-IVET system because its genome sequence is available (Makarova et al. 2006), and it exhibits natural transformation capacities (Blomqvist et al. 2006; Gardan et al. 2009; Fontaine et al. 2010).

An overlapping PCR was performed to generate a mosaic PCR fragment consisting of an internal loxP-specR-loxP region flanked by sequences corresponding to the upstream (UP) and downstream (DN) of STER_0891, the locus chosen to integrate the R-IVET cassette (Fig. 1). This gene, although encoding a putative glucose permease, was used as the target locus because it had already been disrupted in another study without growth disturbance (Fontaine et al. 2010). After transformation of LMD-9 with the mosaic PCR fragment, a spectinomycin-resistant clone was chosen (STUL5001) and was confirmed by sequencing to have the cassette correctly integrated (Fig. 1).

To assess the impact of the partial replacement of the putative glucose permease locus by the loxP-specR-loxP cassette, we monitored growth kinetics of both the wild-type LMD-9 strain and the STUL5001 mutant in M17 medium containing either lactose (LM17) or glucose (GM17) as the main source of carbon (Fig. 3). When cultured in LM17, generation times of LMD-9 (31·9 min) and STUL5001 (33·0 min) were not found significantly different (P < 0·62). When strains were grown in GM17, growth was slower than in M17 medium containing lactose, but no significant difference (P < 0·34) was observed between STUL5001 (41·2 min) and LMD-9 (43·6 min). The faster growth of Strep. thermophilus in lactose compared to glucose was already known (Thomas et al. 2011). In Strep. thermophilus, lactose does not elicit carbon catabolite regulation (van den Bogaard et al. 2000) supporting preference for lactose over glucose as the main source of carbon and energy. More importantly, comparable growth rates in glucose medium for LMD-9 and STUL5001 indicated that the R-IVET cassette present in the putative glucose permease locus does not affect growth in a complex and rich environment such as the M17 medium.

Figure 3.

Growth curves of the LMD-9 strain (circles) and its derivative STUL5001 strain (squares) in M17 medium containing lactose (empty symbols) or glucose (filled symbols). Growth expressed as the optical density at 600 nm was monitored at 15 min intervals with the Bioscreen CTM. Data points represent average values and standard deviations of measurements from three independent cultures. Generation times expressed in min were 31·9 ± 2·1 for LMD-9 in lactose, 33·0 ± 2·6 for STUL5001 in lactose, 43·6 ± 3·8 for LMD-9 in glucose and 41·2 ± 1·1 for STUL5001 in glucose. Using one-way analysis of variance, no significant differences were found between LMD-9 and STUL5001 in lactose (P < 0·62) and in glucose (P < 0·34).

The chromosome stability of the loxP-specR-loxP cassette in STUL5001 was assessed after 30 generations of growth in nonselective medium. Replica plating of 100 individual colonies, amplification of the specR gene in 20 randomly selected colonies and resequencing over the integration locus of one clone confirmed the stability of the STUL5001 strain phenotype and genotype.

Construction of the cre expression vector pULNcreB and stability assessment of the R-IVET tool in Streptococcus thermophilus

The second component of the R-IVET system is the plasmid carrying the promoterless cre gene, which constitutes the basis to construct a library containing random genomic DNAs or a clone containing a specific promoter to be tested. The choice of a replicon able to be maintained in Strep. thermophilus was a critical point as only a few segregationally stable exogenous plasmids exist for Strep. thermophilus (Girard and Moineau 2007; Fontaine et al. 2010). Moreover, the LMD-9 strain used to construct the R-IVET system is known to contain two indigenous plasmids which may represent a supplementary constraint with respect to the feasibility to introduce an additional plasmid. The vector pG+host9TR, a thermoresistant replication derivative of pG+host9 (Maguin et al. 1996), was used as a replicon base as it was shown to generate high transformation rates in Strep. thermophilus (Fontaine et al. 2010). The R-IVET vector pULNcreA was constructed by cloning into pG+host9TR the functional region from pNZ5520 (Bachmann et al. 2008), which contains in the order (i) the las operon transcriptional terminator, (ii) a multiple cloning site including SmaI and BglII restriction sites and (iii) the promoterless cre gene. Because pULNcreA did not show sufficient segregational stability in Strep. thermophilus (data not shown), the sequence of the transcriptional terminator of pepN gene was integrated downstream of the gene cre. This was intended to prevent any read-through transcription that might be responsible for the synthesis of an antisense RNA interfering with the expression of the replication genes (Bron et al. 2004).

The segregational stability of the vector pULNcreB in the recombinant strain was assessed in vitro by growing the strain in LM17 medium without erythromycin and was measured by comparing the number of CFU recovered on medium with and without erythromycin at different sampling points. Erythromycin resistance was found in 90% (72/80) of the clones after one generation, 73% (42/57) after 34 generations and a maintained rate of 38% (62/165) for the period extending from 44 to 52 generations. The presence of pULNcreB in the EryR clones was confirmed by colony PCR performed on 10 EryR colonies for each sampling time.

Keeping in mind that the R-IVET system developed in this study was intended to be used to detect genes, which are specifically active in the gastrointestinal tract of mammals, the presence of both chromosomal and plasmid constructions in STUL5001 harbouring pULNcreB was assessed in faeces after the passage in the GIT of a mouse. One hundred randomly picked colonies showed EryR and SpecR phenotype indicating that they carried both components of the R-IVET system. This was confirmed by a multiplex-PCR analysis performed on genomic DNA extracted from 5 clones with the two sets of primers specific for the floxed specR cassette and for pULNcreB (Fig. 4). These results justified the suitability of pULNcreB as a vector for the implementation of the R-IVET technology in Strep. thermophilus.

Figure 4.

Multiplex amplification of the two R-IVET components in Strep. thermophilus recovered from mouse faeces after GIT transit. Cell lysates of five colonies (lanes 1–5) recovered from the faeces were used as template for amplification with primers #9#10 (gene cre, upper band) and with the cassette primers #21#22 (spec, lower band). Fragments were separated on 1% (w/v) agarose gel in 0·5× TBE at 100 V for 1 h. The 1-kb DNA ladder (lane 1 kb) was used as molecular weight marker.

Assessment of prtS, hsp16 and lac operon promoter activity in in vitro conditions

To evaluate the functionality and the sensitivity of the developed R-IVET tool, we selected 3 Strep. thermophilus genes (i.e. protease prtS, heat-shock protein 16 hsp and lactose operon) known to be expressed in M17 medium. Promoters of these loci were cloned into pULNcreB upstream of the promoterless cre gene. Transformants of STUL5001 obtained with the recombinant plasmids were selected on M17 medium supplemented with erythromycin containing either lactose or glucose as the carbohydrate. Replica plating of colonies onto spectinomycin agar medium showed that all (100/100) clones tested were spectinomycin sensitive except for Plac transformants selected in glucose which displayed 39% (182/300) sensitive clones. The spectinomycin sensitivity of the clones resulted from the deletion of specR floxed cassette by site-specific recombination leaving a single chromosomal loxP site as demonstrated by PCR and sequencing. These results show that both promoters PprtS and Phsp had been activated in all Strep. thermophilus clones during their overnight growth as colonies on the rich and complex M17 medium containing indifferently lactose or glucose. As expected, the lactose operon was found activated in all clones selected on lactose but only in a fraction (39%) of Strep. thermophilus cells when glucose was provided as the sugar. Studying promoter activation in transformant colony assays as performed in this manner generates results which are rather imprecise. Indeed, it delivers a cumulative end point result of what has happened in the cells during the development of the colony. If a promoter is active in these conditions (constitutive or induced), as it is the case of PprtS and Pshsp, all cells of the colony will have had their cassette deleted at the time of test when transferred on spectinomycin. To analyse more precisely the activity of Plac, kinetics of cassette deletion was monitored in cultures performed in liquid M17 medium containing either lactose or glucose (Fig. 5). When cells were grown in M17 containing lactose, the R-IVET cassette was found to be absent in a high majority of cells (84%) after 4 h growth, while it was absent in 31% of cells grown in the presence of glucose. After 7 h growth in lactose, 99·9% of the cells had lost the cassette, whereas glucose led to the cassette deletion in 90% of the cells.

Figure 5.

Monitoring of Strep. thermophilus growth and operon lactose activation in M17 medium containing glucose or lactose. A colony of STUL5001 transformed with pULNcreB-Plac was inoculated into M17 medium supplemented with erythromycin containing either lactose (filled symbols) or glucose (empty symbols). At the indicated sample times, appropriate dilutions of the cultures were plated onto spectinomycin and spectinomycin-free agar plates. Growth curves (circles) are expressed as CFU per ml of culture in a log scale. The percentage of cells where the spectinomycin R-IVET cassette had been deleted (bars) was calculated by removing to 100 the percentage of spectinomycin-resistant CFU number out of the number of CFU obtained on antibiotic-free plates. The assay was repeated twice independently, and only one representative experiment is presented.

Assessment of the lac operon promoter induction in the GIT

Activity of the lactose promoter was evaluated in vivo with Strep. thermophilus cells passing through the GIT of mice drinking water supplemented with lactose. Mice were given cells of STUL5001 pULNcreB-Plac grown in GM17 medium containing spectinomycin to select clones with an undeleted R-IVET cassette. Dilutions of faecal samples collected 3 h, 6 h and 24 h following the intragastric feeding were plated on GM17 containing erythromycin to specifically select R-IVET Strep. thermophilus strains. Colony PCR was performed to test whether the floxed specR cassette was intact or deleted (Fig. 6). At 3 h, clones displayed the full-size 3616-bp amplicon indicative of the intact cassette, whereas clones collected at 6 h and 24 h showed the 2382-bp PCR fragment corresponding to the deleted cassette. Our results show that Plac was active in Strep. thermophilus cells while transiting the mouse digestive tract, indicating that our R-IVET system was functional and could be used in these conditions.

Figure 6.

Monitoring of the cassette resolution in Strep. thermophilus after GIT transit. PCR products obtained from cell lysates of colonies obtained from mouse faeces at 3 h (a), 6 h (b) and 24 h (c) following the intragastric administration. Each cell lysate corresponding to a lane was prepared from 10 colonies. PCR was carried out with the primers #1#4 to detect the presence (3616 bp) or the deletion (2382 bp) of the floxed specR cassette. PCR products were separated on 0·8% (w/v) agarose gel in 0·5× TBE at 100 V during 1 h. The 1-kb DNA ladder (lane M) was used as molecular weight marker.


In this study, we developed a R-IVET tool specifically designed and functional for Strep. thermophilus. The available R-IVET vectors already constructed for Lactobacillus plantarum and Lactococcus lactis (Bron et al. 2004; Bachmann et al. 2008) could not be directly used in Strep. thermophilus LMD-9, which does not support maintenance of the plasmids containing the replication cassette derived from pIL252. We therefore constructed a R-IVET vector with the replication cassette derived from pG+host9 which shows a good segregational stability in Strep. thermophilus (Fontaine et al. 2010). Moreover, although the dual cre-lux reporter system developed in L. lactis allows a rapid monitoring of promoter activities in a quantitative and high-throughput way (Bachmann et al. 2008), the luciferase reporter genes luxAB were not included in our R-IVET plasmid to keep a minimal plasmid size allowing a higher stability. We showed that both components of our R-IVET system, which are the chromosomal floxed cassette and the R-IVET plasmid, were stable in Strep. thermophilus cells during their transit through the GIT of mice, which is a prerequisite to use this tool for identifying genes induced in the GIT.

Monitoring the promoter activity of selected genes in vitro conditions with our R-IVET system was an interesting way to test its functionality and gauge its sensitivity. We used prtS encoding the cell wall protease, hsp16 encoding a low molecular weight stress protein and the lactose operon, which were known to be expressed in M17 at various degrees and different stages of growth. We found that both promoters PprtS and Phsp were highly active in M17 irrespective of the sugar added. Activity of PprtS in M17 had already been observed in a study using a transcriptional fusion with the β-glucuronidase reporter gene (Liu et al. 2009). This is also consistent with the high protease activity of CNRZ385 found in M17 at the end of the exponential phase (Shahbal et al. 1993) and with the fact that PrtS protease was initially purified from Strep. thermophilus grown in LM17 (Fernandez-Espla et al. 2000). Similar to PprtS, Phsp was found to be highly active in M17 which is in agreement with a basal expression level in M17 medium increased in response to an acid stress as it is at the end of exponential phase (González-Márquez et al. 1997; Petrova and Gouliamova 2006). The results obtained with PprtS and Phsp show that the R-IVET system is very sensitive, but is unable to provide a precise promoter activity quantification due to a very low threshold. This is a limitation of the R-IVET technology when it comes to analyse the activity of promoters which are strongly active in the ‘in vitro’ conditions used to select the transformant colonies. It would be interesting to select transformants obtained with PprtS and Phsp in another ‘in vitro’ medium where components do not trigger activity of the promoter and then analyse the promoter activity in a second medium containing potential inducers.

When the activity of Plac in M17 was analysed directly in transformant colonies, we found a full activity in lactose, while activity rate was estimated to be 39% when selected in glucose. However, this value was found to be rather variable depending on plate incubation time and other parameters (data not shown). The Plac activity was therefore analysed in a more precise manner by monitoring the spectinomycin resistance monitored in exponentially growing cells cultured in liquid M17 medium containing either lactose or glucose. We observed that Plac was constantly more active in LM17 than in GM17, with a ratio of 3 at the beginning of the exponential phase. The higher activity of Plac in lactose compared to glucose had already been reported elsewhere to be twice in Northern blot experiments (van den Bogaard et al. 2000) and up to the maximum of 9-fold in a kinetic study using a transcriptional fusion to the luciferase genes (Drouault et al. 2002). Variability in lactose glucose ratio values observed between these three studies is caused by the use of different sugar concentration in M17 medium, different sampling times and most probably by the different way of measuring promoter activity. If we compare luciferase and recombinase Cre as reporters of promoter activity, luciferase has the advantage of being very sensitive and producing luminescence which can be proportionally correlated with the frequency of transcription events. This is true as long as substrate and cofactor necessary for the luciferase enzymatic activity are present in excess inside the tested cells. The main advantage of R-IVET system is that the substrate of the recombinase (i.e. the floxed cassette) is intrinsically provided within the detection system as it is present in the chromosomal DNA of the tested cells. However, enzymatic activity the recombinase leads to the irreversible deletion of the floxed cassette. As a result of this, the promoter activity can only be expressed as the cumulative rate of clones exhibiting a deleted cassette at the time of sampling. Moreover, if numerous transcription events occur at the beginning of the assay, the number of cassettes available for activity measurement becomes very scarce for the rest of the experiment. The detection system can become rapidly saturated with a highly overwhelming proportion of deleted cells, which renders further quantification difficult to achieve. To ensure informative and accurate comparison of promoter activity between different media or environments, quantification using the R-IVET system should be expressed as the time needed to deplete 50% of the cassette pool.

When tested in the murine GIT, we showed that our R-IVET system was functional with Plac found active in Strep. thermophilus cells recovered from faeces after transits of 6 h and 24 h. This is in agreement with earlier studies showing expression of the lactose operon in GIT of mice or rats (Drouault et al. 2002; Mater et al. 2006; Thomas et al. 2011). Interestingly, Strep. thermophilus cells recovered after a 3-h transit period did not show any activity of the lactose operon. This shows that the counter selection performed with spectinomycin during growth in M17 eliminated Strep. thermophilus cells with a deleted cassette and guaranteed that all cells inoculated to the mice contained the R-IVET cassette. This cannot be explained by the absence of the lactose inducer in the GIT as lactose had been continuously provided in water to the mice before and during the cell transit. We may also suspect that Strep. thermophilus were still repressed by traces of spectinomycin coming from the in vitro preparation. However, concentration of the antibiotic present in the GIT was far below the minimal inhibitory concentration (MIC) and should not have been high enough to select for cells containing the R-IVET cassette. As an alternative explanation, we believe that during the three first hours following gavage, Strep. thermophilus cells were in lag phase undertaking adaptation to the harsh conditions encountered in the gastrointestinal environment. Most probably, metabolism of this first period was mainly focused in functions such as cell survival or responses to acid or bile stress. It is reasonable to consider that Strep. thermophilus started to consume lactose only after the initial physiological adjustment to the GIT conditions.

In this study, we constructed a R-IVET tool functional in Strep. thermophilus. We showed that it could be used to monitor promoter activity of genes which are not highly active in medium used for the selection of transformants. Using the Plac promoter, we validated its functionality by showing a partial activity in vitro in medium containing glucose and in vivo during the transit of murine GIT. Our R-IVET tool is now available to identify promoters of Strep. thermophilus which are specifically active in complex environments such as in vivo conditions (i.e. GIT of animals) or in dairy matrix of fermented products. This will reveal important markers of the physiology of Strep. thermophilus in its natural ecosystems.


We would like to thank the ‘Mission scientifique Syndifrais-CNIEL’ to have supported this work. M.J. was financially supported by HEC (Higher Education Commission, Pakistan) and by the University of Lorraine.

Conflict of Interest

No conflict of interest declared.