Targeted and Repetitive Chromosomal Integration Enables High-Level Heterologous Gene Expression in Lactobacillus casei

Lactic acid bacteria (LAB), including Lactobacillus casei, have the potential for overexpression of heterologous proteins, such as bioactive molecules and enzymes. However, traditional genetic tools for expression of these proteins show genetic instability or low yields of the desired product. In this study, we provide a procedure for repetitive integration of genes at various chromosomal locations, achieving high-level and stable expression of proteins in Lb. casei without selective pressure. The protocol developed in this study provides an essential reference for chromosomal overexpression of proteins or bioactive molecules in LAB.

ABSTRACT

Lactobacillus casei is a potential cell factory for the production of enzymes and bioactive molecules using episomal plasmids, which suffer from genetic instability. While chromosomal integration strategies can provide genetic stability of recombinant proteins, low expression yields limit their application. To address this problem, we developed a two-step integration strategy in Lb. casei by combination of the LCABL_13040-50-60 recombineering system (comprised of LCABL_1340, LCABL_13050, and LCABL_13060) with the Cre/loxP site-specific recombination system, with an efficiency of ∼3.7 × 10³ CFU/µg DNA. A gfp gene was successfully integrated into six selected chromosomal sites, and the relative fluorescence intensities (RFUs) of the resulting integrants varied up to ∼3.7-fold depending on the integrated site, among which the LCABL_07270 site gfp integration showed the highest RFU. However, integrants with gfp gene(s) integrated into the LCABL_07270 site showed various RFUs, ranging from 993 ± 89 to 7,289 ± 564 and corresponding to 1 to 13.68 ± 1.08 copies of gfp gene integration. Moreover, the integrant with 13.68 ± 1.08 copies of the gfp gene had a more stable RFU after 63 generations compared to that of a plasmid-engineered strain. To investigate the feasibility of this system for bioactive molecules with high expression levels, the fimbrial adhesin gene, faeG, from Escherichia coli was tested and successfully integrated into the LCABL_07270 site with 5.51 ± 0.25 copies, and the integrated faeG achieved stable expression. All results demonstrate that this two-step integration system could achieve a high yield of heterologous gene expression by repetitive integration at a targeted chromosomal location in Lb. casei.

IMPORTANCE Lactic acid bacteria (LAB), including Lactobacillus casei, have the potential for overexpression of heterologous proteins, such as bioactive molecules and enzymes. However, traditional genetic tools for expression of these proteins show genetic instability or low yields of the desired product. In this study, we provide a procedure for repetitive integration of genes at various chromosomal locations, achieving high-level and stable expression of proteins in Lb. casei without selective pressure. The protocol developed in this study provides an essential reference for chromosomal overexpression of proteins or bioactive molecules in LAB.

INTRODUCTION

Lactic acid bacteria (LAB) are widely used in the fermentation of dairy products. Nowadays, they are gaining increasing attention because of their “generally recognized as safe” status from the United States Food and Drug Administration (FDA), and they are used for producing a range of heterologous proteins, such as bioactive molecules and enzymes (1). Strains of probiotic LAB, especially Lactobacillus casei strains, have been used to express plasmid-encoded protective antigens, human lactoferrin, and bifidobacterial phytases (2–5). However, the segregational and structural instability of plasmids, and the addition of antibiotics as a selective pressure, limit the use of these strains in food processes and clinical purposes (2–5). Moreover, plasmid-dependent genetic tools increase the economic cost because of the treatment of the waste medium, which contains antibiotics that were used to maintain the existence of the plasmid in the engineered bacteria (6, 7).

Integration of a gene of interest into a host chromosome is an alternative method to achieve stable gene expression without selective pressure from antibiotics and other molecules (8, 9). Several integration systems have been established for integration of genes into the chromosome in Lb. casei (10–14). Among them, conditional replication vectors, possibly with a counterselection marker, such as the upp gene encoding uracil phosphoribosyltransferase (UPRTase), have been used to achieve chromosomal insertions. This method is often used for insertions that are too laborious and time-consuming for the generation of integrants by prolonged subcultivation of single-crossover strains under replicative conditions to allow vector excision by a second homologous recombination event (10, 11). Another integration method is the transposon system P_junc-Tpase_IS1223, which is composed of plasmids pVI129, expressing IS1223 transposase, and pVI110, a suicide transposon plasmid carrying the P_junc sequence, the recognition site of the IS1223 transposase. Compared with conditional replication vector-based tools, although the integration efficiency is remarkably increased, the random recognition sites of the IS1223 transposase limit its application for targeted integration (12). A genetic system using the integrase and the attP sequence from phage ΦAT3 has also been constructed for heterologous DNA integration in Lb. casei (13), but the infrequency of attB sites on the chromosome limit heterologous gene insertion in the genome.

To address these problems, a targeted chromosome integration tool based on prophage-derived recombinases and the Cre/loxP system was established for construction of integrants in Lb. casei (14). This recombineering system consists of a presumptive 5′ to 3′ exonuclease, LCABL_13060; a single-stranded DNA (ssDNA) annealing protein, LCABL_13050; and a predicted host nuclease inhibitor, LCABL_13040. These are analogous to the lambda phage-derived proteins Beta, Gam, and Exo, collectively named the λ-Red system (15). The heterologous gene is inserted into the Lb. casei BL23 chromosome by homologous recombination between the linear double-stranded DNA (dsDNA) donor and the chromosome, mediated by LCABL_13040-50-60 (comprised of LCABL_1340, LCABL_13050, and LCABL_13060). Subsequently, the antibiotic-selectable marker flanked with two mutant loxP sites (lox66 and lox71) is excised by Cre recombinase. However, this cannot be used for repetitive integration in one round of genetic manipulation.

Fortunately, Cre recombinase also mediates site-specific recombination between a loxP site existing on a circular DNA and a loxP site on the chromosome, leading to integration of the heterologous DNA into the chromosome without antibiotic selective pressure (16, 17). Moreover, after one round of integration, an additional loxP site is generated in the chromosome. Therefore, it is possible to achieve repetitive integration of the gene of interest into the targeted site by the Cre/loxP system in the Lb. casei genome.

Lb. casei BL23, with anti-inflammatory properties, is a model strain widely used in genetic, physiological, and biochemical studies (18). Therefore, it is urgently needed to establish an efficient and stable chromosomal integration tool for high expression of a gene of interest in Lb. casei BL23. In this study, we combined the LCABL_13040-50-60 recombineering system and the Cre/loxP system to establish a two-step integration system in Lb. casei BL23. Using this strategy, we found that the gene integration location and copy number significantly affect the expression level of heterologous genes in Lb. casei BL23.

RESULTS

Construction of a two-step targeted genome integration system in Lb. casei.

Targeted genome integration was performed with the help of Cre recombinase, which catalyzes site-specific recombination between two loxP sites in circular DNA and chromosomal DNA. Thus, the first step of the system was to introduce a loxP site into the genome of Lb. casei BL23 by the LCABL_13040-50-60 recombineering system and the Cre/loxP site-specific recombination system. As shown in Fig. 1, the loxP-cat-loxP fragment was inserted into LCABL_07270 by homologous recombination between the linear donor cassette aHU-loxP-cat-loxP-aHD and chromosomal DNA, mediated by LCABL_13040-50-60 recombinases (14). The recombinant was selected on medium containing chloramphenicol, and the plasmid pMSP456, containing the LCABL_13040-50-60 gene cluster, was cured by culturing the recombinants without erythromycin selection. Then, the plasmid-free recombinant was transformed with pMSPcre and induced by nisin to express Cre recombinase to excise the chloramphenicol resistance gene cat from the chromosome, leaving a loxP site at the targeted location. Therefore, a loxP site was precisely introduced into the chromosome of the resulting Lb. casei MT1 strain harboring plasmid pMSPcre.

FIG 1.

Schematic diagram showing the process of the entire integration system. Prophage recombinases LCABL_13040-50-60 mediated homologous recombination, resulting in disruption of geneX and insertion of the cat marker, while Cre site-specific recombinase subsequently eliminated the marker and left a loxP site. The inserted loxP site was verified by sequencing analysis. Finally, the chromosomal integration of the cat gene carried by the integration construct pUC-lox-cat was mediated by Cre recombinase. 40-50-60, recombinases LCABL_13040-50-60; P_nisA, nisin-inducible promoter; HU and HD, up and down homology arms; cat, bla, and Erm indicate chloramphenicol, ampicillin, and erythromycin resistance genes, respectively.

The second step was to generate site-specific recombination between the loxP site on the integration construct pUC-lox-cat and the loxP site on the chromosome of Lb. casei MT1. The integration construct pUC-lox-cat was obtained by ligation of the loxP-cat cassette into a nonreplicate plasmid, pUC19, in Lb. casei BL23, to avoid the replication of the integration construct, which could give rise to false-positive clones. After 1 µg of pUC-lox-cat was electroporated into Lb. casei MT1 expressing Cre recombinase, about 3,700 recombinants were selected on medium containing only chloramphenicol, on which the erythromycin-resistant plasmid pMSPcre would be cured (Fig. 1). As a control, no resistant colonies were obtained when the equivalent pUC-lox-cat was electroporated into Lb. casei BL23 expressing Cre recombinase. These results indicated that pUC-lox-cat was successfully integrated into the LCABL_07270 site on the chromosome of Lb. casei MT1. To confirm the integration of pUC-lox-cat, twenty chloramphenicol-resistant recombinants were picked randomly and amplified by 5′ and 3′ junction PCR with the primer pairs aHUF/catR3 and catF3/aHDR. As shown in Fig. S1A and S1B, all of the recombinants harbored pUC-lox-cat. Next, the stability of a pMSPcre-cured integrant was detected, and the results showed that the chloramphenicol resistance phenotype of this integrant was stable without selective pressure after 99 generations, under both static and aerobic conditions (Fig. 2). In contrast, recombinant Lb. casei/pCD4033 was subsequently lost after selective pressures were removed.

FIG 2.

Stability of the chloramphenicol-resistant (Cm^r) phenotype of the integrant (IT) with pUC-lox-cat inserted into the chromosome of Lb. casei BL23 under static or aerobic conditions. The stability of Lb. casei BL23 transformed with a replicative vector pCD4033 (RT) was also performed under static or aerobic conditions for comparison. Each value is the mean ± standard deviation of three independent experiments.

Effects of the precise integration sites in the chromosome on gene expression.

Besides LCABL_07270 (named the “a” site), loxP was inserted into five other sites (LCABL_11510, LCABL_12890, LCABL_16340, LCABL_22090, and LCABL_28960, named “b,” “c,” “d,” “e,” and “f,” respectively) (Table S1) of Lb. casei BL23 that were uniformly located on the chromosome, and the integration events had no effect on the growth of the resulting Lb. casei MT2, MT3, MT4, MT5, and MT6 strains. Then, the integrated construct pUC-lox-gfp containing the P_ldh-gfp expression cassette was electroporated into Lb. casei strains MT1 to MT6, and recombinants with a single-copy insertion of pUC-lox-gfp were confirmed by PCR amplification (Fig. S2A and S2B). Relative fluorescence intensities (RFUs) (analysis of variance [ANOVA] test; P < 0.05) and the relative gene dosages (ANOVA test; P < 0.05) of these six loci were measured during the exponential growth phase of the recombinants. The highest RFU was observed when the gfp gene was integrated into the “a” site, and it was ∼3.7-fold higher than that of the d site (Fig. 3A). In line with the RFU, the relative dosage of the gfp gene at the a locus was ∼2.2-fold higher than that at the “d” locus (Fig. 3B). However, the relative gene dosage of the a locus was similar to those of the “b” and “f” loci (Student’s two-tailed t test; P = 0.09 and P = 0.31,), while the RFU of the a locus had significant differences from those of the “b” and “f” loci (Student’s two-tailed t test; P < 0.01 and P < 0.05).

FIG 3.

Effects of chromosome position on green fluorescent protein (GFP) expression. (A) RFUs are plotted relative to the chromosomal positions. Each value is the mean ± standard deviation of three independent experiments. The wild-type strain BL23 was used as a control. (B) Gene dose ratio of the gfp gene at the six sites, relative to the d site. Each value is the mean ± standard deviation of three independent experiments. *, P < 0.05; **, P < 0.01; N.S., no significant difference.

Repetitive integration of gfp in the “a” site.

As stated above, the a site exhibited the highest RFU among the six selected sites. Therefore, we chose the a site to investigate whether this system could be used for repetitive chromosomal integration of the gfp gene (Fig. 4A). As shown in Fig. 4B, the RFUs of fifteen integrants grown on MRS broth with chloramphenicol spanned from 993 ± 89 to 7,289 ± 564 (ANOVA test; P < 0.01), corresponding to 1 to 13.68 ± 1.08 (ANOVA test; P < 0.01) copy numbers of the gfp gene as determined by quantitative PCR (qPCR). This indicates that the expression level of heterologous genes could be improved by repetitive chromosomal integration of the genes in Lb. casei. Moreover, as shown in Fig. S3, cell growth was not affected in the 13.68 (±1.08)-copy strain compared with that of single-copy strains. Interestingly, the highest RFU obtained from this study was higher than the RFU of a recombinant strain harboring plasmid pCD4033-gfp grown on MRS medium with chloramphenicol (Fig. 4B).

FIG 4.

Effects of repetitive chromosomal integration of the gfp gene on GFP expression. (A) Schematic illustration of multiple gene copies integrated at the same chromosomal location. (B) Relative fluorescence units (RFU; ■) and gfp gene copy numbers (▲) of the 15 randomly selected integrants. “16” indicates the RFU of the recombinant harboring plasmid pCD4033-gfp. Each value is the mean ± standard deviation of three independent experiments. **, P < 0.01. (C) The stability of an integrant with the highest copy number of the gfp gene on the chromosome or recombinants harboring vector pCD4033-gfp on MRS without chloramphenicol. “% of RFU” indicates the RFU level of the strain aftern generations without the use of selective pressure divided by the RFU level of the same strain aftern generations with the use of selection pressure. Each value is the mean ± standard deviation of three independent experiments.

Next, the effects of stability of the gfp gene on expression were compared between the recombinant with pCD4033-gfp and the integrant with significantly more copies (13.68 ± 1.08) (Student’s two-tailed t test; P < 0.01) of the gfp gene on MRS medium without chloramphenicol. The results showed that green fluorescent protein (GFP) expression could not be detected after 18 generations without selective pressure from recombinant pCD4033-gfp (Fig. 4C). As shown in Fig. 4C, even after 63 generations without selective pressure, the % of RFU (RFU level of the strain after n generations without the use of selective pressure/RFU level of the same strain after n generations with the use of selection pressure) was still maintained at levels near 100%.

Feasibility of the integration system to achieve high expression of FaeG.

To investigate the feasibility of this two-step integration system, the extracellular fimbrial adhesin FaeG expression cassette (faeG from Escherichia coli CVCC200 fused with the promoter and signal peptide of S-layer protein from Lactobacillus crispatus K313) was integrated into the chromosome. The highest copy number was 5.51 ± 0.25 (Student’s two-tailed t test; P < 0.01) (Fig. 5A). As shown in Fig. 5B, the FaeG protein expression level of the 5.51 (±0.25)-copy strain (integrant 5) was higher than that of the one-copy strain (integrant 3). Densitometry analysis using 10 μl of protein sample indicated the gray value of the band from integrant 5 was ∼4.3-fold higher than that of the band from integrant 3 (Student’s two-tailed t test; P < 0.01) (Fig. 5B).

FIG 5.

Modulation of FaeG expression. Each value is the mean ± standard deviation of three independent experiments. (A) Integrated copy numbers of the faeG gene of integrants 1 to 16. (B) Western blot and densitometric analysis of FaeG expression by integrant 3 and integrant 5. (C) Integrated copy numbers of the faeG gene of integrant 5; 9, 18, and 27 indicate the number of generations. (D) Western blot and densitometric analysis of the stability of FaeG expression by integrant 5; 9, 18, and 27 indicate the number of generations. Lane M was the Fermentas PageRuler prestained protein ladder. **, P < 0.01; N.S., no significant difference.

To demonstrate the stability of the integrant, the expression levels of the FaeG protein labeled with His and the faeG gene copy number were both measured. As shown in Fig. 5C and D, the faeG copy number and the gray values by densitometry (using 10 μl of protein sample) were all stable after 27 generations without selective pressure (ANOVA test; P > 0.05).

DISCUSSION

Lb. casei possesses the largest genome of the lactic acid bacteria (LAB). It is indigenous in oral, vaginal, and intestinal tissue of many animals (19). However, traditional molecular tools for overexpression of recombinant proteins show genetic instability or insufficient yields. To achieve high levels of heterologous gene expression, we propose an efficient and stable strategy based on a two-step targeted gene integration tool in Lb. casei BL23. Using this strategy, we successfully integrated 13.68 ± 1.08 copies of the gfp gene and 5.51 ± 0.25 copies of the faeG gene into the selected a site of the Lb. casei BL23 chromosome and achieved high-level, stable expression. In our opinion, screening more colonies would likely obtain integrants with more than 13.68 ± 1.08 copies of the gfp gene and more than 5.51 ± 0.25 copies of the faeG gene inserted into the chromosome.

Nowadays, the most common and simple method is to choose a high-copy-number plasmid to overexpress heterologous proteins (2–5). Results in this study indicate that plasmid-engineered strains are not suitable for long-term gfp gene expression without the presence of antibiotics. Fortunately, genetic stability is a hallmark of the chromosomal integration system, and this was confirmed for GFP production in Lb. casei by cross-generational culturing without selective pressure. This observation suggests that the chromosomal expression of heterologous genes using the proposed system overcomes the genetic instability characteristics of plasmid-encoded expression of recombinant proteins in Lb. casei. Although in general, such highly repetitive regions will undergo homologous recombination during continued cultivation, the background homologous recombination level of Lb. casei BL23 was very low in our previous work, where the frequency of the initial single crossover integration was 2 × 10⁻⁶ (per CFU), and one integrant was cultivated for 400 generations without erythromycin. Only three out of 600 erythromycin-sensitive integrants displayed a second crossover homologous recombination (20). Therefore, homologous recombination may only have a small effect on the stability of multiple integrations of the same gene expression cassette in continued culture.

For this integration system, insertion of a loxP site into the chromosome of a wild-type strain would be the unique pretreatment for the use of the proposed integration system. A similar prerequisite is also needed for other site-specific integration systems, including the integrase of phage ΦAT3, isolated from Lb. casei ATCC 393 for mediated chromosomal integration, which needs a preexisting attP sequence on the chromosome (13). However, no efficient genome editing tool has been reported to introduce extra attP sequences into the chromosome, which limits the use of the integrase of phage ΦAT3. The advantage of the method proposed in this study is the ease of loxP site insertion at almost any locus in the genome with the assistance of the LCABL_13040-50-60 recombineering system (14). Thus, it is possible to choose different sites in order to detect the effects of chromosomal position on heterologous gene expression. Moreover, Cre recombinase has been demonstrated to be functional in several lactic acid bacteria (21, 22). Therefore, this research provides a viable and novel integration method for testing in other LAB strains.

Previous attempts to improve the expression of heterologous genes from the chromosome include constructing gene expression cassettes with different strengths of promoters (23, 24), cascading promoters (25), ribosome-binding sites (26), or terminators (27). However, a gfp gene expression cassette with different strengths of promoters only exhibited a 2.8-fold activity increase in Lactococcus lactis (28). Moreover, the screening of such promoter libraries is tedious and labor-intensive (28). This study provides a new chromosomal integration tool for enhancing gene(s) expression levels through the selection of chromosomal positions and gene copy numbers. Here, a high variation (∼3.7-fold) in the expression level of gfp genes randomly distributed within the Lb. casei chromosome was observed. Similar results have been reported in other bacteria, such as E. coli, Lc. lactis, and Bacillus subtilis (29–31). Previously, an effect of the genomic integration position on the level of protein expression was reported in E. coli (32). The common explanation is that DNA replication originates at the replication origin on the genome, and new rounds of replication are initiated before the previous rounds have been completed. Therefore, the copy numbers of integrated genes at different positions will fluctuate, resulting in rising gene dosage with increasing distance from the replication origin (32). Here, gene dosage can only account for an ∼2.2-fold difference in position-dependent variation of gfp gene expression, while position-dependent gfp gene expression levels can vary by ∼3.7-fold, indicating that chromosomal position effects are not solely due to gene dosage in Lb. casei. Several other processes that are involved in chromosome structuring and organization, such as the expression of neighboring genes and the 3D structure of the genome, may also impact gene expression (29, 33).

So far, to our knowledge, no report has focused on the chromosomal repetitive integration of a gene(s) in one round of integration in Lb. casei. For Lc. lactis, the lactococcal group II intron has been used for multicopy integration of heterologous genes (34). However, the specific bacterial insertion sequences required on the chromosome limit its application for targeted integration, which could be well addressed by the method described in this study (34). Using this system, targeted repetitive integration of the gfp or faeG gene into the selected site was obtained and the expression levels of both genes were significantly improved. These results indicate that repetitive integration of genes into the chromosome facilitates a high yield of gene expression, which agrees with the results reported by Gu et al. (35).

In this study, we developed a two-step chromosomal integration strategy, which combines the LCABL_13040-50-60 recombineering system with the Cre/loxP site-specific recombination system for targeted integration of the DNA fragment into the chromosome in Lb. casei BL23. Based on this strategy, high-level expression of the gfp and faeG genes on the chromosome was successfully achieved, and the effect of the chromosomal position and the gene copy number on heterologous gene expression was demonstrated. This strategy has significant potential for clinical and industrial applications because of its high efficiency and stable expression of recombinant proteins.

MATERIALS AND METHODS

Bacterial strains, plasmids and growth conditions.

Bacterial strains and plasmids used in this study are listed in Table 1. Here, as hosts for plasmid construction, E. coli DH5α cells were cultured aerobically in Luria-Bertani (LB) medium at 37°C. Unless otherwise mentioned, lactobacilli and their derivatives were grown statically in MRS broth (Oxoid) at 37°C. If necessary, antibiotics were supplemented as follows: 100 µg/ml ampicillin, 30 µg/ml kanamycin, and 10 µg/ml chloramphenicol for E. coli DH5α, and 5 µg/ml erythromycin or chloramphenicol for lactobacilli.

TABLE 1.

Strains and plasmids in this study

Strain or plasmid	Characteristic(s)	Source or reference
Strains
Escherichia coli DH5α	F− supE44 ΔlacU169 Ф80lacZ ΔM15 hsdR17 recA1 endA1 gyrA96 thi-1 relA1	Novagen
Lactobacillus casei BL23	Derivative of Lb. casei ATCC 393 (pLZ15−)	41
Lb. casei MT1	Derivative of Lb. casei BL23 pflB::loxP	This study
Lb. casei/pCD4033	Lb. casei BL23 harboring plasmid pCD4033	This study
Plasmids
pUC19	Ampr; cloning vector	This study
pET-28a	Kanr; cloning vector	This study
pMSP456	Expression LCABL_13040-50-60 under PnisA control	14
pMSPcre	Expression Cre under PnisA control	14
pUCgalK	Source of fragment loxP-cat-loxP	14
pCD4033-gfp	Gene gfp as a reporter in the downstream of Pldh	37
pCD4033	Cmr	37
pALRb-faeG	Source of FaeG expression cassette	20

Plasmid construction.

All primers used in this study are listed in Table 2. The linear donor cassette (aHU-loxP-cat-loxP-aHD) for disruption of LCABL_07270, named the a site (Table S1), was prepared according to our previous study (36). The loxP-cat-loxP fragment was obtained from the pUCgalK vector (14) using the catF1 and catR1 primers. The upstream and downstream homologies (aHU and aHD) were PCR amplified from the chromosome of Lb. casei BL23 (GenBank accession number FM177140.1) using the primer pairs aHUF/aHUR and aHDF/aHDR and spliced by an overlap extension PCR using primers aHUF and aHDR. The XhoI restriction site responsible for ligation to loxP-cat-loxP was introduced by the aHDF primer. The fused product, aHU-SalI-aHD, was digested with BamHI and HindIII and ligated into the corresponding pUC19 sites. The yielding vector paHUD and loxP-cat-loxP were digested with XhoI and ligated to create paHUD-cat. Finally, the linear donor cassette, aHU-loxP-cat-loxP-aHD, was generated by PCR from the paHUD-cat vector using primers aHUF and aHDR. The same method was used to obtain the other linear donor cassettes for disruption of LCABL_11510, LCABL_12890, LCABL_16340, LCABL_22090, and LCABL_28960 (Table S1), named the b, c, d, e, and f sites, respectively, except for the skeleton plasmid pET-28a or pUC19.

TABLE 2.

Primers used in this study

Primer	Sequence (5′–3′)a	Restriction site or reference no.
aHUF	CGGGATCCAATAGCTCAGATTTTTAACAACA	BamHI
aHUR	TAATTTATCAAAAACCTTATTCATG
aHDF	ATAAGGTTTTTGATAAATTACTCGAGAAGCCATCGCTGCTAACAAA	XhoI
aHDR	CCCAAGCTTGTACATCATACTGTTCATGCC	HindIII
bHUF	GAACTCGAGTACCTGGCTGTATAATTTAACCTTT	XhoI
bHUR	GATAATCACTCCCTTTGATGTCA
bHDF	CATCAAAGGGAGTGATTATCGTCGACTATCTATGCACTCAGAACGC	SalI
bHDR	ACATGCATGCTCGTCATCAATCCTTACATCC	SphI
cHUF	AAAGATCTAATCCTGCCTCACGCATTAA	BglII
cHUR	CAGAACGCAAATCTCTTCTA
cHDF	TAGAAGAGATTTGCGTTCTGGTCGACTTACTTAATGCTATTCATTA	SalI
cHDR	AACTGCAGACATCGAGTTCAGCAAGCTAA	PstI
dHUF	GAACTCGAGTTTCTCGTGATGGCTACGTTAA	XhoI
dHUR	AGTTGTTTGCCTCCTAAAGTGG
dHDF	CCACTTTAGGAGGCAAACAACTAGATCTGCAGCTTAACTCAAGACAGGAA	BglII
dHDR	CAAAAGCTTGTGTCTTTGAGGAAAAAATGCG	HindIII
eHUF	GAACTCGAGTCTTCGGTTGGCATACAGTCAACCA	XhoI
eHUR	CAAGCCACCCTTTCTGTTAATTATG
eHDF	TTAACAGAAAGGGTGGCTTGGTCGACGCTTCATCATGGTATTAAAC	SalI
eHDR	ACATGCATGCATTGTTACTATAATTTGCTTGACCG	SphI
fHUF	GAACTCGAGACAGTTGAGCATGTCTTTTCGGATC	XhoI
fHUR	TCCTTAAGCTTCTGGATATTTCTTG
fHDF	AATATCCAGAAGCTTAAGGAGGATCCGTTGCCAATACTAACGATTT	BamHI
fHDR	GAAGATCTAGCACCGACACCATAATTGTAATAC	BglII
catF1	CCGCTCGAGATAACTTCGTATAATGTATGCTATA	XhoI
catR1	CCGCTCGAGATAACTTCGTATAGCATACATTATA	XhoI
catF2	CGAGATCTATAACTTCGTATAATGTATGCTATA	BglII
catR2	CGAGATCTATAACTTCGTATAGCATACATTATA	BglII
catF3	GTCGACGGCAATAGTTACCCTT
catR3	GGAGATCTCTGTAATATAAAAACCTTCTTC	BglII
gfpF	ATCCTCTAGAGTCGATTATAGGCAAGGGCCGGCTC
gfpR	AACTATTGCCGTCGAAATTCTTAGTAGAGCTCATC
faeGF	TTATACGAAAGTCGATGCAGCTAAGCAAGACTAAT
faeGR	AACTATTGCCGTCGAGAAGATTGCCGAAAATATGC
RT-gfpF	TTCTGTCAGTGGAGAGGGT	28
RT-gfpR	GGATAACGGGAAAAGCATT	28
RT-pyrGF	AATTGCGCTTTTCACTGATG	38
RT-pyrGR	CGAAATGATCGACCACAATC	38
RT-faeGF	AGTAACTGGTGGTGTAGATGG
RT-faeGR	AACCACCATAAAAGATAGAGC

^aThe restriction sites in the primer sequences are underlined.

The cat-loxP cassette was amplified from the pUCgalK vector (14) with primers catF1 and catR3 and digested with XhoI and BglII. The resulting fragment was ligated to the SalI and BamHI sites of pUC19, yielding the empty integration construct pUC-lox-cat.

The P_ldh-gfp cassette was obtained by PCR from the pCD4033-gfp vector (37) using the primer pair gfpF and gfpR with 15 bp complementary to the linear pUC-lox-cat generated by SalI. These two resulting fragments were fused using a 1:3 (vector:insert) molar ratio with the 5×In-Fusion HD cloning kit (TaKaRa), yielding the GFP expression integration construct plox-cat-gfp. A similar method was used for preparing the FaeG expression integration construct containing a His tag for Western blot analysis, except for the template plasmid pALRb-faeG (20), with primer pair faeGF/faeGR.

Targeted insertion of a loxP site into the chromosome.

The recombineering procedure was carried out according to our previous report (14). The chloramphenicol-resistant recombinants were picked randomly and verified by PCR with primer pairs aHUF/aHDR. Procedures for excising the cat gene and introducing a loxP site into the chromosome were also carried out according to the previously described method (14). The inserted loxP site on the a site was tested by PCR and sequencing, yielding Lb. casei MT1. The same method was used for insertion of the loxP site into the other five sites.

Integration of the heterologous genes into the chromosome.

The mutant Lb. casei MT1 harboring pMSPcre was used for integration of pUC-lox-cat into the chromosome. The induction of Cre recombinase expression was carried out at an optical density at 600 nm (OD₆₀₀) of 0.25 to 0.30 for 2 h with 10 ng/ml nisin, followed by the preparation of electrocompetent cells. After the mixture of 1 µg pUC-lox-cat and competent cells were cooled for 10 min, electroporation (2,000 V, 25 µF, 400 Ω) was performed and recovered at 37°C for 2 h. Subsequently, the integrants were checked by PCR using primers catF1 and aHDR. The same method was used for the other integration experiments.

Gene dosage measurements.

Genomic DNA was extracted with a TIANamp Bacteria DNA kit (Tiangen Biotech, Beijing, China) in accordance with the manufacturer’s protocols. Real-time PCR was performed with the SYBR Premix Ex Taq II (TaKaRa, Japan), using the protocol for a real-time PCR detection system (Bio-Rad, USA). Gene dosage was measured as described previously (29). Briefly, relative quantities of the gfp target gene were determined using the pyrG reference gene (38). Changes in threshold cycle (△C_T) values for the integrant “d” reactions were used as calibrators (△△C_T = △C_T target − △C_T “d”) for analysis of results by the relative quantification (2^−△△CT) method (39). Quantities of the gfp target gene relative to that at the “d” locus are represented as gene dosage. Three technical replicates of the gene dosage analysis were performed for each experiment.

qPCR analysis of the gene copy number on the chromosome.

The integrated gene copy number at the “a” locus was detected as described previously (35). qPCR was performed with SYBR Premix Ex Taq II (TaKaRa, Japan, following the protocol of the a real-time PCR detection system (Bio-Rad, USA). The relative gfp gene copy numbers of randomly selected integrants were determined by normalizing reaction threshold cycle (C_T) values to that of the pyrG reference gene (38).

Western blot analysis of FaeG expression from the chromosome.

Overnight cultures of strains containing the faeG expression cassette were incubated in MRS medium, then the supernatant was harvested at 15 h by centrifugation at 8,000 × g for 10 min and filtered by 0.22-µm pore filters, precipitated by 10% (mass/vol) trichloroacetic acid, and dissolved in a 1/200 volume of PBS buffer. Western blot analysis was done in accordance with our previous work (20). Briefly, the membrane with FaeG protein was blocked by 5% (mass/vol) skim milk for 1 h and then incubated overnight with anti-His monoclonal antibody (ZSGB Bio) at a 1:1,000 dilution. After washing three times, the membrane was incubated with peroxidase-conjugated AffiniPure goat anti-mouse IgG (H+L) (ZSGB Bio) at a dilution of 1:5,000 for 1 h. Finally, the bands were visualized by enhanced chemiluminescence horseradish peroxidase (HRP) substrate (Immobilon Western, Millipore, Billerica, MA). Densitometric analysis was performed using ImageJ (40). Relative protein contents were presented as the intensity of the gray band on the membrane.

Stability test.

The stabilities of integrants with a cat gene on the chromosome or recombinants harboring vector pCD4033 (37) were determined without chloramphenicol selection. The overnight cultures grown with chloramphenicol at 37°C were diluted to 10⁻¹ in 5 ml of MRS broth without chloramphenicol added (this was designated generation 0) and grown statically or aerobically at 37°C for 12 h. The generation 0 cultures were 10-fold serially diluted, and 100 µl of diluted solution was spread onto MRS plates with or without chloramphenicol. Aliquots of these grown cultures were serially transferred and spread after 12 h of incubation into MRS medium to count the number of CFU with or without chloramphenicol. The stability was determined as (CFU with chloramphenicol/CFU without chloramphenicol) × 100%. The number of generations was determined by dividing the total incubation time from generation 0 by the duplicated time for the integrant counted.

The stabilities of the integrants with the highest copy numbers of the gfp and faeG gene on the chromosome or recombinants harboring vector pCD4033-gfp (37) were detected in accordance with the protocol described above. Aliquots of these grown cultures were serially transferred after 12 h of incubation into MRS medium without selective pressure. The relative fluorescence intensities were measured according to our previous work (14), and the stability was estimated as (RFU without chloramphenicol/RFU with chloramphenicol) × 100%. The stability of the faeG gene expression was estimated as the intensity of the gray band on the membrane and the relative copy number.

Statistical analysis.

All results are presented as mean ± standard deviation [SD] (n = 3 parallel samples). Statistical analysis was performed using ANOVA for multiple-group comparison, Student’s two-tailed t test was used to calculate the P values between groups. P values of <0.05 were considered statistically significant. P values of <0.01 were considered statistically highly significant.