Markerless Multiple-Gene-Deletion System for Streptococcus mutans

Abstract

Inactivation or selective modification is essential to elucidate the putative function of a gene. The present study describes an improved Cre-loxP-based method for markerless multiple gene deletion in Streptococcus mutans, the principal etiological agent of dental caries. This modified method uses two mutant loxP sites, which after recombination creates a double-mutant loxP site that is poorly recognized by Cre recombinase, facilitating multiple gene deletions in a single genetic background. The effectiveness of this modified strategy was demonstrated by the construction of both single and double gene deletions at the htrA and clpP loci on the chromosome of Streptococcus mutans. HtrA and ClpP play key roles in the processing and maturation of several important proteins, including many virulence factors. Deletion of these genes resulted in reducing the organism's ability to withstand exposure to low pH and oxidative agents. The method described here is simple and efficient and can be easily implemented for multiple gene deletions with S. mutans and other streptococci.

The availability of ever-increasing numbers of genome sequences from pathogenic microorganisms has greatly facilitated the genetic dissection of the factors that determine their fitness and pathogenic potential. Genetic analysis of many oral streptococci, including Streptococcus mutans, which is considered the principal etiological agent of dental caries (29), has thus far been limited due to a lack of advanced genetic tools. Typically, the function of a gene is assessed by either inactivation or selective modification. Classic strategies used to obtain gene deletions from S. mutans include single-crossover insertion duplication mutagenesis (45) or allelic exchange with antibiotic cassettes via double-crossover recombination (22, 27). In addition, various transposon (16, 23, 39) and random insertional mutagenesis (25, 46, 47) strategies are employed for obtaining gene deletions in S. mutans. Generally, these strategies result in the introduction of a suitable marker into the genome, facilitating the selection of mutants. However, when multiple gene deletions are required, the limited number of convenient, selectable markers makes the adoption of these strategies less feasible.

The use of the Cre-lox recombination system to remove selectable markers from the genome is an efficient way of circumventing this problem. The versatility of Cre recombinase makes it ideal for use in various gene manipulation strategies involving plants (15), Saccharomyces cerevisiae (40), mice (41, 49), human cell lines (32), and various microorganisms (12, 37, 44). The Cre recombinase of bacteriophage P1 is a 38-kDa protein that belongs to the integrase family of site-specific recombinases (42). It catalyzes cofactor-independent recombination between two of its recognition sites, known as loxP, which also originates from Escherichia coli phage P1 (1). The 34-bp consensus sequence for the loxP site consists of an asymmetrical core spacer of 8 bp, defining the orientation of the loxP site, and two 13-bp palindromic flanking sequences (19). A DNA sequence flanked by the loxP sites is excised when the loxP sites are convergently oriented, whereas the sequence is inverted when the loxP sites are divergently oriented (33). Cre recombinase is able to act on both inter- and intramolecular loxP sites, although recombination of the intramolecular lox sites is kinetically favorable (26). However, wild-type loxP sites may cause problems during multiple gene deletions. This is because the integration of multiple loxP sites into the genome can cause genomic instability, due to potential recombination between loxP sites from different cassettes.

In this study, we demonstrate the use of an improved Cre-loxP-based system for the simultaneous inactivation of multiple genes in S. mutans. HtrA and ClpP, two important serine proteases, were selected for analysis, since they are associated with the virulence of this pathogen and because of their uniquely detectable phenotypes. Using the modified Cre-loxP method, we successfully deleted htrA and clpP from S. mutans. Mutant strains lacking clpP and/or htrA exhibited a wide range of stress-sensitive phenotypes, such as reduced tolerance to low pH and to oxidative stress-inducing agents.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

E. coli strain DH5α was grown in Luria-Bertani medium supplemented, when necessary, with ampicillin (100 μg·ml⁻¹), kanamycin (Km; 50 μg·ml⁻¹), and spectinomycin (Sp; 100 μg·ml⁻¹). S. mutans strains were routinely grown in Todd-Hewitt medium (BBL; Becton Dickinson) supplemented with 0.2% yeast extract (THY) and, when necessary, Km (300 μg·ml⁻¹), Sp (300 μg·ml⁻¹), or erythromycin (Em; 10 μg·ml⁻¹). The strains used in this study are available upon request.

Construction of Cre-loxP-mediated gene replacement mutants.

Based on an analysis of the S. mutans strain UA159 genome (accession no. AE015037) (3), 1.9- and 1.7-kb DNA fragments containing htrA and clpP, respectively, were PCR amplified using the primer pairs Smu-HtrA-F1/Smu-HtrA-R1 and Smu-ClpP-F1/Smu-ClpP-R1 (Table 1). The PCR products were cloned into the pGEM-T-Easy TA cloning vector (Promega) (pIB102 and pIB143 for htrA and clpP, respectively) and were confirmed by using restriction digestion. A Km resistance cassette amplified from pUC4ΩKm2 (34), using two different sets of primers (loxP-Km-F/loxP-Km-R and lox71-Km-F/lox66-Km-R), was cloned into BaeI-PflMI-digested/T4 DNA polymerase-blunted htrA in pIB102, yielding pIB501, containing wild-type loxP sites (htrA::loxP-Km-loxP), and pIB506, containing mutant loxP sites (htrA::lox71-Km-lox66); both products were verified by restriction digestion. An Sp resistance cassette was amplified from pUC-Spec (20), using the primers lox71-Sp-F and lox66-Sp-R (Table 1), and inserted into the BbsI-ClaI-digested and blunt-ended clpP gene in pIB143, yielding pIB507. The plasmids pIB501, pIB506, and pIB507 were linearized with NotI and used for the transformation of S. mutans UA159 as previously described (9), with transformants selected on THY plates containing Km or Sp. For the simultaneous inactivation of htrA and clpP, S. mutans UA159 was cotransformed with NotI-linearized pIB506 and pIB507, and transformants were selected on THY plates containing Km and Sp. Deletion of the htrA (IBS501 and IBS508) and clpP (IBS509) genes by double crossover was verified by PCR using the primer pairs Smu-HtrA-F2/Smu-HtrA-R2 and Smu-ClpP-F2/Smu-ClpP-R2, respectively.

TABLE 1.

List of oligonucleotides^a

Primer name	Sequence (5′ to 3′)	Purpose
Smu-HtrA-F1	gactagcattatttggaattttctcatcgg	Amplification of htrA and construction of pIB101
Smu-HtrA-R1	gaggtgagatatgaagacaacttcgttgac
Smu-HtrA-F2	aataacgaaggtcaaatcagaa	Verification of the htrA gene deletion
Smu-HtrA-R2	atatctttagtattaagatatt
Smu-ClpP-F1	gttccaggtcttggggatgcaggcgatcgcttg	Amplification of clpP and construction of pIB143
Smu-ClpP-R1	ggtggtttatgcggtgccagcggtaaacccg
Smu-ClpP-F2	ctgaagatattgaccaacgtca	Verification of the clpP gene deletion
Smu-ClpP-R2	cgtatcaatataggatggaaga
loxP-Km-F	CGATAACTTCGTATAATGTATGCTATAGCAAGTTATgaggatgaagaggatgaggaggcag	Amplification of Kmr marker with flanking native loxP sequences
loxP-Km-R	CGATAACTTTGTATAGCATACATTATACGAAGTTATgctttttagacatctaaatctagg
lox71-Km-F	CGTACCGTTCGTATAGCATACATTATACGAAGTTATgaggatgaagaggatgaggaggcag	Amplification of Kmr marker with flanking mutant loxP sequences
lox66-Km-R	CGTACCGTTCGTATAATGTATGCTATACGAAGTTATgctttttagacatctaaatctagg
lox71-Sp-F	CGTACCGTTCGTATAGCATACATTATACGAAGTTATctagataaaaaatttagaagcccaatg	Amplification of Spr marker with flanking mutant loxP sequences
lox66-Sp-R	CGTACCGTTCGTATAATGTATGCTATACGAAGTTATctgttatttaaatagtttatagttaaatttac

^aloxP sites are presented in boldface, and mutations created to generate lox66 and lox71 sites are underlined. DNA sequences homologous to template DNA are presented in lowercase letters.

Cre-mediated mutant locus resolution.

The pCrePA plasmid (kindly provided by S. H. Leppla, NIH) contains the cre gene under the promoter of Bacillus anthracis protective antigen (pagA) gene isolated from pAE5 (35, 36), an Em^r gene as a selectable marker, and a temperature-sensitive replicon (pWV01) from pHY304 (30, 38). To excise the loxP/P*-Km^r (where P* indicates the mutant loxP sites) resistance cassette or the loxP*-Sp^r resistance cassette from the chromosome, IBS501, IBS508, IBS509, and IBS510 were transformed with pCrePA. Transformants were grown at 30°C on THY-Em plates, and selected colonies were then grown at 30°C in THY-Em broth and plated on THY-Em plates. Approximately 500 colonies from THY-Em plates were patched onto THY plates containing either Km or Sp. Colonies that were Em resistant and Km sensitive (Em^r Km^s), Em resistant and Sp sensitive (Em^r Sp^s), or Em resistant and Km and Sp sensitive (Em^r Km^s Sp^s) were transferred into antibiotic-free THY plates by patching. The plates were incubated overnight at 37°C to cure pCrePA (30, 38), generating Em^s cells, which were selected for further analysis. Deletions of the htrA and clpP genes and the proper resolution of loxP (both native and mutant) sites were verified by PCR analysis and DNA sequencing.

Evaluation of stress-sensitive phenotypes.

Cultures grown overnight were washed and resuspended in 0.85% saline to an optical density at 600 nm of 5.0. The cultures were serially diluted 10-fold in 0.85% saline, and 7.5 μl of each dilution was spotted onto THY plates containing puromycin (4.0 to 6.0 μg/ml; Sigma-Aldrich) or methyl viologen (MV; 5.0 to 10.0 mM; Sigma-Aldrich). Plates were incubated at 37°C under microaerophilic conditions, and the bacterial growth was evaluated as described previously (8).

For experiments involving pH changes of the growth medium, the initial pH of the THY broth was adjusted to pH 5.5 or pH 7.0 with HCl prior to sterilization. A citrate-phosphate buffer (50 mM) with the desired pH was then added to the medium after sterilization. The buffered THY medium was inoculated with S. mutans cultures (3% [vol/vol]) grown overnight and incubated at 37°C. Growth of the various cultures was monitored as described previously (8).

RESULTS

A strategy for markerless gene deletion using the Cre-loxP system.

The genetic events involved in Cre-loxP-based gene deletion are presented in Fig. 1A. Although this method has been used with many microorganisms, it has never been applied to streptococci. Using this technique, we attempted to delete the htrA gene in S. mutans. The htrA gene was first cloned into a plasmid vector (pIB102) and then inactivated via the insertion of a Km^r cassette flanked by wild-type loxP sequences (loxP-Km^r-loxP, pIB501). Following the transformation of pIB501 in S. mutans, double-crossover events were selected on the basis of Km resistance. The double-crossover recombination frequency was found to be 2.3 × 10⁻³ with respect to the number of viable cells, and the allelic exchange event was verified by PCR (Fig. 2). One such mutant, designated IBS501, was transformed with pCrePA. The chromosomally integrated loxP-Km^r-loxP cassette was then excised by the transient expression of Cre recombinase. After overnight growth at 30°C, approximately 13% of the colonies were Km^s, indicating the loss of the Km^r marker, which was also confirmed by PCR (Fig. 2). Longer periods of incubation at 30°C in THY-Em increased the yield of colonies with the successful excision of the loxP-Km^r-loxP cassette. Em^r Km^s colonies were selected and grown in THY broth at 37°C. As pCrePA has a temperature-sensitive replicon (30, 38), the elevation of the growth temperature to 37°C without the selection pressure (Em) resulted in rapid curing of the plasmid. After colonies were incubated overnight, 99.9% of the S. mutans colonies (500 colonies tested) were cured of pCrePA, resulting in a mutant strain containing a deletion of the htrA gene (IBS502). This result is similar to the gene deletion results with B. anthracis, where a single passage at 37°C completely cured the cells of pCrePA (37). These results clearly demonstrate that the Cre-loxP system can be efficiently employed for gene deletions in S. mutans; the entire protocol takes less than a week to obtain a particular mutant.

FIG. 1.

(A) Schematic diagram of the Cre-loxP* system for gene deletion with S. mutans. The wild-type (WT) gene targeted for inactivation is cloned into a vector and interrupted by the insertion of a lox66-Ab-lox71 cassette. (i) The construct is transformed into S. mutans, and the successful allelic exchange event (IN, integrated product) is selected using appropriate antibiotic-containing plates. (ii) Removal of the lox66-Ab-lox71 cassette from the chromosome is achieved by Cre-mediated excision (EP, excised product) after the strain is transformed with pCrePA and grown at permissive temperature (30°C). (iii) Elevation of the growth temperature to 37°C results in the loss of the temperature-sensitive (Ts) pCrePA plasmid, resulting in a mutant strain carrying a deletion in the target gene without a selectable marker. (B) Schematic representation of the loxP* sites. The loxP site consists of two 13-bp inverted repeats surrounding an 8-bp asymmetric core sequence (spacer). lox71 and lox66 have 5 bp mutated in the left and right 13-bp repeats, respectively. Mutated sequences are in boldface letters and underlined. Cre-mediated recombination between lox66 and lox71 results in the formation of lox72, which has mutations in both 13-bp repeat sequences.

FIG. 2.

PCR analysis of the modified S. mutans strains. The htrA gene amplified with the Smu-HtrA-F2/Smu-HtrA-R2 primers from UA159 (lane 1), strain IBS501 (lane 2), strain IBS502 (lane 3), strain IBS508 (lane 4), strain IBS511 (lane 5), strain IBS510 (lane 6), and strain IBS511 (lane 7). The clpP gene amplified with the Smu-ClpP-F2/Smu-ClpP-R2 primers from UA159 (lane 8), strain IBS509 (lane 9), strain IBS512 (lane 10), strain IBS510 (lane 11), and strain IBS513 (lane 12).

Single and multiple deletions using the Cre-loxP mutant.

As described above, the use of wild-type loxP sites is not ideal for multiple gene deletions due to potential Cre-mediated recombination between two loxP sites from different loxP-Ab^r-loxP cassettes. This limitation can be circumvented by employing loxP* sites, as described by Araki et al. (5). Two mutant sites, lox71 and lox66, were selected based on the following properties. The loxP site is composed of an 8-bp spacer flanked by 13-bp inverted repeats (19). In lox71, the left 13-bp repeat element has five mutated bases, whereas in lox66, the right 13-bp repeat element has five mutated bases. Recombination between the lox71 and the lox66 mutant creates the loxP double mutant site lox72, which contains mutations in both of the repeats (Fig. 1B) and therefore has a dramatically reduced affinity for Cre recombinase (4). To demonstrate the successful deletions of multiple genes by using the loxP*-Ab^r cassettes, the htrA and clpP loci were selected for analysis; both of these genes were successfully inactivated in S. mutans, either individually or simultaneously. The recombination frequencies were found to be 1.7 × 10⁻³, 0.4 × 10⁻³, and 0.2 × 10⁻³ for htrA, clpP, and the simultaneous inactivation of both loci, respectively. In each case, single colonies were chosen and designated IBS508, IBS509, and IBS510, respectively (Table 2). To excise the lox66-Ab^r-lox71 cassette from the mutant locus, the double-crossover mutants IBS508, IBS509, and IBS510 were transformed with pCrePA. The excision efficiencies of the Ab^r cassettes from the htrA, clpP, and simultaneously knocked-out loci were 31, 14, and 11% (500 colonies were tested), respectively. As described before, pCrePA was cured from these mutant strains after cultures were incubated at 37°C. Single-colony isolates, designated IBS511 (ΔhtrA), IBS512 (ΔclpP), and IBS513 (ΔhtrA ΔclpP), were selected. Deletion of the target loci was verified by PCR and by DNA sequencing, to confirm excision of the loxP*-Ab^r cassettes (Fig. 2).

TABLE 2.

S. mutans strains used in this study

Strain	Genotype	Source or reference
UA159	Wild type, serotype c	R. A. Burne, University of Florida
IBS501	UA159 htrA::loxP-Km-loxP	This study
IBS502	UA159 ΔhtrA::loxP	This study
IBS508	UA159 htrA::lox71-Km-lox66	This study
IBS509	UA159 clpP::lox71-Sp-lox66	This study
IBS510	UA159 htrA::lox71-Km-lox66, clpP::lox71-Sp-lox66	This study
IBS511	UA159 ΔhtrA::lox72	This study
IBS512	UA159 ΔclpP::lox72	This study
IBS513	UA159 ΔhtrA ΔclpP::lox72	This study

Successful generation of multiple gene deletions via Cre-lox recombination depends on the correct resolution events in the presence of the resident integrated lox72 sites. Therefore, we attempted to delete clpP from a ΔhtrA strain (IBS511), where the resident lox72 site and the newly introduced lox66/lox71 sites were well separated from each other; thus, the possibility of a recombination event between lox72 and the lox66/lox71 sites was remote. However, studies with Lactobacillus plantarum demonstrate that when lox66 or lox71 sites are in close proximity to a lox72 site, Cre-mediated recombination between lox66 and lox71 sites occurs correctly in 99.4% of cases (24). Preferred Cre-mediated resolution between lox66 and lox71 sites occurs almost exclusively, whether the sites are distant or within close proximity to a lox72 site, underscoring the selectivity of the Cre enzyme and the advantage of this system relative to methods that employ native loxP sites.

Stress-sensitive phenotype of the mutant strains.

To verify the biological effects of the mutations using the Cre-loxP* method, the phenotypes arising from the deletions of htrA and/or clpP were analyzed. The serine proteases HtrA and ClpP are involved in regulating the growth of S. mutans under various stress conditions (2, 11, 13, 28). The mutant strains were tested for their abilities to withstand superoxide stress generated by MV and thermal stress generated by puromycin (14, 48). We observed that IBS511 (ΔhtrA) was highly sensitive to MV, while IBS512 (ΔclpP) was more sensitive to puromycin treatment (Fig. 3).

FIG. 3.

Deletions of htrA and clpP affect the stress tolerance of S. mutans. Freshly grown overnight cultures of UA159 wild type (WT) and its derivatives were serially diluted with 0.85% NaCl. Portions of each dilution (7.5 μl) were spotted onto THY agar plates with no additions (control, THY), methyl viologen (10 mM, THY + MV), or puromycin (6 μg ml⁻¹, THY + Pu). The plates were incubated at 37°C under microaerophilic conditions. Experiments were repeated more than three times, and relevant areas of representative plates are shown.

S. mutans rapidly adapts to an acidic environment by escalating a strong acid tolerance response (7, 17, 18). To determine the role of HtrA or ClpP in the acid tolerance response, mutant strains were grown in buffered THY broth at pHs 7.0 and 5.5. At pH 7.0, the ΔclpP and ΔhtrA ΔclpP mutants grew more slowly than the wild type or the ΔhtrA strain (Fig. 4A), but during incubation at pH 5.5, the growth rates of the ΔhtrA and ΔhtrA ΔclpP mutants were severely impaired (Fig. 4B). This suggests that HtrA and ClpP are essential for the survival of S. mutans under a variety of stress conditions, including high temperature, oxidative stress, and acid tolerance.

FIG. 4.

Growth characteristics of the protease mutants. Growth of the wild-type and the S. mutans mutant strains, in buffered THY of pH 7.0 (A) and pH 5.5 (B), at 37°C were monitored with a Klett-Summerson colorimeter. The strains used for analysis include UA159 (▪, WT), IBS511 (▴, ΔhtrA), IBS512 (•, ΔclpP), and IBS513 (⧫, ΔhtrA ΔclpP). The experiments were repeated twice, and representative growth curves are shown.

DISCUSSION

Several markerless-gene-deletion mutagenesis methods have been developed for the construction of mutants of different bacterial species. In streptococci, markerless in-frame deletions have been constructed previously by using overlapping extension PCR (21, 27) or by employing temperature-sensitive suicide vectors, such as pG+host (6, 10). In both approaches, mutants have been reported to occur at very low frequencies (1 to 2.5%). This is attributed to the low frequency of the recombination events that lead to excision. Moreover, there is an equal possibility for the formation of an in-frame deletion mutation or a reversion to the wild-type genotype (6, 10). Because of this, counter-selection strategies are integrated with these approaches to facilitate the screening process. One such method, which involves a galactokinase (galK)-based negative selection system, resulted in a markerless in-frame deletion of the lacG gene from S. mutans (31). In this case, 50% of the screened isolates contained the desired deletion; however, the plasmid excision occurred at a very low frequency (1/3,000). The main limitation of this method is the nature of the starting strain, which should be devoid of the galactose utilization operon; therefore, this method is very difficult to adapt to different S. mutans strains or to other bacteria. Two other methods, the cotransformation strategy with a thermosensitive plasmid (9) and the Janus cassette allele-exchange method (rpsL) (43), allow for markerless deletions and point mutations to be constructed. However, application of the rpsL method is limited to S. pneumoniae, which develops higher competency than S. mutans. In addition, a low rate of excision of the integrated plasmid would hinder the adaptation of this method for markerless gene deletion with S. mutans (6). With the cotransformation strategy, the efficiency of gene deletion was found to be approximately 10% (9), and therefore it is cumbersome to apply this method for multiple gene deletions.

To our knowledge, this is the first report of the Cre-loxP system being used for gene deletion analysis in a streptococcal species. The Cre-loxP* system described here offers several advantages compared to the methods described above. The Cre-loxP-based method does not require any host cofactors or accessory proteins, exhibits high recombination efficiency, and is independent of the length of DNA located between the two loxP sites. The temperature-sensitive replicon in pCrePA is derived from the broad-host-range plasmid pWV01, which can replicate efficiently in streptococci and other gram-positive bacteria (30). It is a very powerful tool that can be used for functional genomics study. The efficiency of the process can be further improved if Cre is expressed from, or within, a cassette flanked by loxP* sites, conferring antibiotic resistance. Recombination between the loxP* sites in that case would lead to the removal of the antibiotic resistance marker and cre, such that the extra step of pCrePA removal could be avoided.

The efficiency of the Cre-loxP* recombination system for multiple gene deletions was demonstrated by the deletion of htrA and clpP, which encode serine proteases essential to the virulence of S. mutans. Deletion of these genes resulted in the appearance of stress-dependent phenotypes when the mutant strains were exposed to stress induced by extremes of temperature and pH, as well as oxidative stress. In conclusion, Cre-loxP* facilitates rapid genetic analysis of S. mutans in order to elucidate the molecular mechanisms of virulence in this pathogen and can be extended toward the study of other streptococci.