Enhance nisin yield via improving acid-tolerant capability of Lactococcus lactis F44

Traditionally, nisin was produced industrially by using Lactococcus lactis in the neutral fermentation process. However, nisin showed higher activity in the acidic environment. How to balance the pH value for bacterial normal growth and nisin activity might be the key problem. In this study, 17 acid-tolerant genes and 6 lactic acid synthetic genes were introduced in L. lactis F44, respectively. Comparing to the 2810 IU/mL nisin yield of the original strain F44, the nisin titer of the engineered strains over-expressing hdeAB, ldh and murG, increased to 3850, 3979 and 4377 IU/mL, respectively. These engineered strains showed more stable intracellular pH value during the fermentation process. Improvement of lactate production could partly provide the extra energy for the expression of acid tolerance genes during growth. Co-overexpression of hdeAB, murG, and ldh(Z) in strain F44 resulted in the nisin titer of 4913 IU/mL. The engineered strain (ABGL) could grow on plates with pH 4.2, comparing to the surviving pH 4.6 of strain F44. The fed-batch fermentation showed nisin titer of the co-expression L. lactis strain could reach 5563 IU/mL with lower pH condition and longer cultivation time. This work provides a novel strategy of constructing robust strains for use in industry process.


Results
Activity of nisin and the growth of the original strain in acidic condition. Nisin, as a kind of small molecular peptide, showed different activity in diverse pH conditions. For analyzing the activity of nisin in fermentation process, 4000 IU/mL nisin standards were added to the fermentation medium at pH 2.0, 3.0, 4.0, 5.0, 6.0 and 7.0, respectively. Nisin titer was measured every 2 hours during the incubation. As it is shown in Fig. 1, the activity of nisin decreased significantly with the increase of pH value of medium. Nisin could remain stable for 4 h at pH 2.0, but when the pH value reached 5.0, nisin titer fell by 50% in 2 h.
Fermentation process was also greatly affected by the initial pH of broth. The fermentation medium was adjusted to pH 2.0, 3.0, 4.0, 5.0, 6.0 and 7.0, respectively, and the starter strain L. lactis F44 was incubated for Introduction of single acid tolerant gene to improve nisin production. To determine the effects of different acid tolerant proteins on nisin production, various acid tolerant genes with different mechanism were introduced respectively into L. lactis F44 and the effects of these genes on acid tolerance were analyzed. Firstly, genes with different acid tolerant mechanisms were transformed to L. lactis F44 strain individually. It is indicated that the capacity of nisin production increased differently due to the expression of different acid tolerant genes (Fig. 2). Among the engineered strains, the strain harboring murG gene, which is responsible for the transfers of the N-acetylglucosamine moieties onto the carrier lipid in cell wall biosynthesis 43 , performed best in nisin production. Nisin accumulation was up to 4377 IU/ml, which is 55.7% higher than that of the original strain. Meanwhile, the growth rate of the engineered strains was slightly lower than that of F44 (see Supplementary Table S3 ). At the late fermentation period, such as 8 h after starting fermentation, F44 stopped growing almost, while the engineered strains can still maintain good growth activity at least till 10 h.
Enhancing lactic acid synthesis pathway to improve nisin production. We also cloned two groups of pyk, pfy and ldh genes encoding pyruvate kinase Pyk, phosphofructokinase Pfk and lactate dehydrogenase Ldh, the key enzymes of lactic acid synthesis pathway, from L. lactis F44 and Lactobacillus casei Zhang, respectively (Fig. 2). Then six genes were introduced to L. lactis F44 separately. The results showed that nisin titer of the engineered strains harboring pyk, pfy and ldh all have been improved. Among the engineered strains, the nisin titer of F44 overexpressing ldh gene from L. casei Zhang was highest. The nisin titer was improved from 2810 IU/mL to 3979 IU/mL, increased by 41.6%. As for the cell growth, the six strains performed similarly with the strains harboring single acid tolerant gene, which showed better robustness than F44 in the stationary phase (see Supplementary Fig. S2a,b).
Combined expression of several genes to improve nisin production. To further improve the acid tolerance of the strain, the genes which helped to enhance the nisin yield of L. lactis F44 through single overexpression and involves in different acid tolerant mechanisms and energy metabolism were selected to be co-overexpressed, and we constructed the following overexpression strain: L. lactis F44 (ABG) (combined expression of triple genes hdeA, hdeB and murG), F44 (ABGL) (combined expression of tetrad genes, hdeA, hdeB, murG and ldh (Z)), F44 (MLL) (combined expression of murG and ldh (L)), F44 (MLZ) (combined expression of murG and ldh (Z)), F44 (PPL) (combined expression of triple genes, pfk (Z), pyk (Z) and ldh (Z)). The nisin titer of all these engineered strains increased during the fermentation process (Fig. 3a). However, at the exponential time, the growth rate of the engineered strains was significantly lower than F44 strain. This might result from Original strain F44, strains overexpressing acid tolerant genes and strains overexpressing lactic acid synthetic genes were grown in feed medium at 30 °C, and photographed every 2 h. Comparative nisin production in F44 and engineered strains. Error bars, SD from three replicate flasks. * p < 0.05, * * p < 0.01, t-test. the extra-consumption of energy in over-expression of acid tolerant genes, and shortage of energy used in cell growth. When entering the late exponential phase, strain F44 (ABGL) showed remarkable increase in the production of lactate, which could provide bacteria more energy (Fig. 3b). The extra energy supply might be used to maintain the stability of cells. Among the engineered strains, the highest nisin titer was observed for F44 (ABGL), which displayed 4913 IU/mL nisin titer, increased by 74.84% compared with F44 strain. In addition, the transcription levels of hdeA, hdeB, murG and ldh(Z) in the F44 (ABGL) were significantly higher than its corresponding genes in the wild type F44 (Fig. 4a,b). The efficiency of the PCR reached 90.13% (see Supplementary Fig. S4), which meant the results of qRT-PCR analysis showed that the P45 promoter of pLEB124 had high efficiency at different pHs and all the heterogenous genes in the engineered strains could be successfully transcribed to the corresponding mRNAs.
The expression levels of the nisin biosynthesis-related genes between wild type F44 and the engineered strain F44 (ABGL) were examined by qRT-PCR. The cells at mid-log phase (8 h) for the highest gene expression was harvested for qRT-PCR. As shown in Supplementary Fig. S6, the expression levels of most genes related to nisin biosynthesis of F44 (ABGL) were slightly higher than that of the original strain F44. And the change of expression levels between the two strains was less than 2 times, which demonstrated no significant differences were found. Moreover, the expression levels of other nisin related genes nisB and nisR of F44 (ABGL) were lower than that of F44.
At the same time, all engineered strains obtained higher survival rate in low pH condition (Fig. 5a,b). Strain F44 (ABGL) displayed the highest survival rate to low pH compared with the original strain F44 (Fig. 5b). The strong acid-tolerant capacity observed from the F44 (ABGL) should be the result of the overexpression of different proteins related to different mechanisms of acid tolerance. These data suggested that the higher nisin yield of the engineered strains than F44 is the result of a significant increase in acid tolerance capacity.
Higher expression of acid tolerant genes ensures the stability of the intracellular pH. Intracellular pH (pH i ) plays a major role in response to acid stress in lactic acid bacteria 44,45 . In order to further explore the influence of the acidic fermentation environment on the engineered strains and the original strain F44, the pH i  of these strains were measured during the fermentation process. Obvious differences in pH between F44 and the engineered strains were observed after 6, 8, and 10 h of fermentation. The engineered strains appeared to have the ability to maintain a higher pH i than the F44 (Fig. 6a,b). This indicated that the engineered strains could avoid the sharp decrease of pH i under acid stress effectively. Significant pH i differences were detected in the engineered strains and the original strain F44 in response to acidic environment, which proved the protection of the acid tolerant genes on the L. lactis strain under acid environment. The data above showed that the increase of nisin yield might be the result of the stabilization of the intracellular pH by the overexpression of some acid tolerant genes.
Fed-batch fermentation of the engineered strains helps to promote nisin yields. To further identify the performance of the artificial acid tolerant module, we performed fed-batch fermentation to optimize the fermentation processes. For F44, the consumption rate of sucrose was high in the initial 6 h, while the sucrose concentrations remained unchanged after 12 h 6 . The sucrose was added during fermentation process as an energy supplement for the cell growth and the nisin production. As shown in Fig. 7a,b, the growth rate of F44 was still higher than other engineered strains in the exponential phase. With the fermentation time extended, the biomass of acid tolerant strains were similar to that of F44. There were no significant differences on the biomass among F44 and other two strains, F44(ABGL) and F44 (MurG). Nisin production of F44 (ABGL) and F44 (MurG) was obviously higher than that of the original strain F44 (Fig. 7c). Compared with the nisin titer of F44, the nisin titer of F44 (MurG) and F44 (ABGL) were improved from 3454 IU/mL to 4774 IU/mL and 5167 IU/mL, increased by 38.2% and 49.6%, respectively. Although the growth conditions of the strains were similar, the pH i of engineered strains should be significantly higher than that of F44 due to the overexpression of the acid tolerant genes which provided the relatively mild intracellular environment for cell growth and nisin production.
To increase the nisin production and reduce the restriction of acid stress, the broth was maintained at pH 6.0 from 6 th hour by adding NaOH solution (10 M). All strains grew fast initially during the first 6 h. F44 grew faster than the engineered strains during the exponential phase, while the difference between F44 and other strains diminished by the fermentation extension (Fig. 8a). When pH was adjusted to 6.0, the nisin titer achieved the fairly high level and then increased slowly, finally dropped gradually. This indicates that the bacteria require an adaptive period when encountering environmental change. Then strains recovered rapidly, the cell growth continued and nisin was accumulated more. The nisin titer of strain F44 (ABGL) also reaches peak value of 5563 IU/mL  which was 1.48 fold that of the original strain, and was also higher than that of F44 (MurG) during the whole fermentation process (Fig. 8b).
Similarly, we also measured the OD 600 and nisin production through the fed-batch fermentation, of which the pH was adjusted to 5.5, 5.0, and 4.0, respectively (see Supplementary Fig. S7). In the culture at pH 5.5, the growth of all strains is only marginally affected by pH regulation, and the nisin yield had a quite similar trend at medium pH from 6.0 to 5.5 (see Supplementary Fig. S7a,b). The nisin titer of F44 (ABGL) was 4674 IU/mL, which increased by 24.2% than that of F44.
The low pH of broth has a great effect on the growth and production of nisin. The growth of all strains was strongly inhibited in acidic environment after adjusting the pH of fermentative broth (see Supplementary Fig.  S7c,e). F44 grew faster during the exponential phase, while the difference between F44 and the engineered strains diminished in the condition of pH 5.0 and 4.0, which showed that overexpression of acid tolerant genes indeed protected cells under acid stress.
The activity of nisin is better at low pH, but the growth of all strains was restricted which lead to weakening the nisin production (see Supplementary Fig. S7d). When pH of medium was adjusted to 5.0, the nisin titer decreased and nisin production (c) with F44, transformants with single expression of murG and combinatorial expression of hdeA, hdeB murG and ldh genes, respectively. The culture temperature was maintained at 30 °C. Concentrated sucrose (500 g/L) was fed at a constant rate of 2 mL/h between 6-12 h during fermentation process. Samples were taken at every 2 h. at first and then transitorily increased in an adaptive buffer stage, finally dropped gradually. The nisin titer of F44 (ABGL), which appeared a short-term bounce of 3671 IU/mL after pH adjustment, was still significantly higher than F44, but that of strain F44 did only reach the level of 2799 IU/mL before acidification of broth. The nisin titer of engineered strains was always higher than that of F44 during the whole fermentation process. For the culture at pH 4.0 by adding HCl at 6 h, the final biomass levels further decreased. The growth retardation of all strains was observed after the acidifying operation, which illustrated that improving acid tolerance of engineered strains have reached the limitation, which has a great impact on nisin production (see Supplementary Fig. S7f).

Discussion
Many metabolites produced in the industrial fermentation process could seriously inhibit the bacterial growth 46 , with the example of acidic metabolites, such as lactic acid and some carboxylic acids. These metabolites not only inhibit the cell growth but also reduce the yield of the target products. During the lactate fermentation process, the accumulation of lactic acid produced by lactic acid bacteria acidified the broth, therefore suppressed the cell growth of lactic acid bacteria whose optimal pH was about 6.3-6.9, and also limited the yield of lactic acid 47 . The traditional method to deal with this problem was to maintain the optimal pH of broth by adding alkali, which not only complicated the fermentation process but also might cause contamination [48][49][50] . It was reported that one kind of bacteria consuming lactic acid was co-cultured with L. lactis 51 , which could reduce the influence of lactic acid and pH on the growth of bacteria and nisin production. But the co-cultivation increased the difficulty of the industrial fermentation and the cost of the separation process. Introducing the acid tolerant genes to improve acid resistance of cells is an alternative method, by which, the strains can maintain normal growth rate in the late stationary stage of fermentation and therefore the accumulation of target metabolites can be promoted.
The pH value of broth can decrease to 4.5 due to the accumulation of acidic metabolites (lactic acid) during fermentation, which limits the normal growth of L. lactis, whose optimal pH is about 7.2, and then also affects the production of nisin. At the same time, the fluctuations of intracellular physiological environment under acid stress also lead to the reduction of physiological activity. The improvement of acid tolerance capacity is beneficial to prolong the fermentation period, and improve the growth activity of bacteria, which is of great importance to ferment process producing acidic metabolites. In response to acid stress, a large number of acid tolerant proteins were expressed to protect L. lactis cells. In recent years, the development of genome sequencing and various omics approaches have assisted to illuminate many acid tolerant mechanisms. Many genes were found significantly upregulated under acid stress by transcriptional analysis 52 . Some of these genes were further proved to be involved in various mechanisms under stress conditions 53,54 . Acid tolerant genes expression in the engineered strains could improve the acid tolerance capacity and maintain good cell growth, and therefore might improve nisin yield. We attempted to overexpress 17 acid-tolerant genes and 6 lactic acid synthetic genes in L. lactis F44 (Table 1). It was found that acid tolerance and nisin yield of the engineered strains could be improved in varying degree. Among these engineered strains, overexpression of murG, hdeAB, ldh makes better performance on the robustness and nisin production of cells.
Encoding a kind of glycosyltransferase, murG is involved in the peptidoglycan synthesis of bacterial cell walls by catalyzing the successive transfers of the N-acetylglucosamine moieties onto the carrier lipid 43 . The cell wall-associated protein MurG might be abundant and accumulated under acidic conditions. The expression of murG lead to higher survival, and the enzyme might act as a key role of fixing cell walls under acid stress. HdeA and its structural homologue HdeB could maintain the optimal chaperones activity at pH 2.0 and pH 4.0, respectively, therefore could be used as acid-resistance proteins in bacteria 55 . These two chaperones can help bacteria to survive in acidic protein-unfolding conditions. It is interesting that the chaperones HdeA and HdeB from E. coli could take effect in L. lactis which belongs to gram-positive bacteria without periplasmic space.
Besides, during the production of lactic acid, as one end product of the glucose metabolism in the lactic acid bacteria, energy can be produced in this process. As the key enzymes involved in the synthesis of lactic acid, pyk, pfy and ldh were confirmed to a higher expression level in lactate fermentation by proteomic analysis 52,56 . Most acid tolerant mechanisms need to consume energy. Under low pH conditions, the energy from sugar metabolism is mainly used to withstand harsh environment, therefore the cell growth and the production capacity are affected by the energy supply. So enhancing the lactic acid production and energy generation by overexpressing these genes could effectively improve the cell growth, acid tolerance capacity and nisin yield. It is interesting that the optical density of F44 was higher than that of engineered strains which may result from the faster growth rate of F44 in the exponential phase (see Supplementary Table S3). Indeed, the replication and expression of acid tolerant genes from the vector might cause metabolic burdens in the engineered strains and therefore delayed their growth. After 10 h fermentation, the cell growth and the nisin production of all strains were in stagnation stage with pH close to about 4.5 (see Supplementary Table S3). At the same time, the nisin accumulation dropped greatly, and nisin titer of all strains decreased in a same trend, which illustrated that improving acid tolerance of engineered strains has reached the limitation (see Supplementary Fig. S2c and Table  S3). These results showed it was difficult to maintain physiological activities of cells as well as normal metabolic process under severe acid stress. However, as expected, nisin yields of the engineered strains were higher than that of the original strain F44, especially in late fermentation stage, which meant these cells could be resistant to acidic environment and remained active and stable (Fig. 2). Although the nisin yields of engineered strains and F44 exist obvious differences, there were no significant differences in the expression levels of nisin gene cluster between them according to the results of qRT-PCR. Thus, it is concluded that improvement of nisin yield is resulted from that overexpression of acid tolerant genes makes the strains more robust and adaptive under acid stress rather than increases the expression of genes related to nisin biosynthesis. The results of combined overexpression of hdeAB, murG, and ldh(Z) in F44 showed nisin yield of the engineered strain was improved further, and better effects were also achieved by analysis of fed-batch process (Figs 7b and 8b). Although the nisin yield at pH 5.5 did not reach a higher level, the value of 5.5 seemed to be the limit culture pH value enabling growth to proceed, and the inhibition of low pH (pH 5.0 and 4.0) on the growth and the ability of nisin production was obvious (see Supplementary Fig. S7). Neither the Δ pH between the cytoplasm and the culture medium nor the pH i were maintained constant when the culture pH value was lower than 5.0, and the efficiency of biomass synthesis relying on the energy supply also decreased 57 . The lower the pH was, the less the cell concentration was obtained when the strains reached the stationary phase, and the nisin production dropped greatly. The nisin titer of all strains decreased in a general trend, which showed it was difficult to maintain physiological activities of cells as well as normal metabolic process due to acidic damage during the cultures performed at medium pH 5.0 or lower. Although the nisin titer of the engineered strains harboring several acid tolerant genes was higher than the single gene expression strains, the overexpression of more acid tolerance genes would consume more energy, therefore, the overload of acid tolerant genes overexpression might affect the higher improvement of the nisin titer. In addition, from the results of acid stress assay, compared with the original strain, the engineered strains constructed in this paper had more survival rates in the environment of lower pH, and also had higher nisin yield, which confirmed our hypothesis that overexpressing these genes would increase the production of nisin (Fig. 5a,b). Maintaining relatively stable intracellular pH is important for microorganisms, which can ensure the normal physiological behavior of cells and guarantee the activity of enzymes 58,59 . Intracellular pH should be affected by the large changes of environmental pH. Exceeding certain range of environmental pH fluctuation would influence the dynamic balance of the intracellular physiological activities and even affect the survival of cells. The engineered strains showed higher pH i value than that of the original strain (Fig. 6a,b). These engineered strains could maintain a relatively stable pH i and enhance the adaptability of the bacteria in certain acidic environment.
According to what we learnt, it is the first report that the nisin production was improved by over-expressing acid-tolerant genes, which would provide some clues for the construction of other acid-tolerant microorganism. And in the future study, we want to optimize the promoters to regulate the expression level of the genes and reduce the side effects resulted from the excessive expression of heterogenous proteins. In addition, we also plan to construct an acid-tolerant network which can regulate the engineered strains in multi-level and cross joint manner.

Methods
Bacteria strains and growth conditions. The bacterial strains and plasmids used in the study are listed in the Supplementary Table S1. The L. lactis strain F44 was used for the phenotypic examination throughout this study. The E. coli TG1 was used for plasmid preparation. All the E. coli strains were grown at 37 °C, with shaking at 180 rpm in the Luria-Bertani (LB) medium. All L. lactis strains were preserved and cultured in seed medium (wt/vol) containing peptone (1.5%), yeast extract (1.5%), sucrose (1.5%), KH 2 PO 4 (2.0%), NaCl (0.15%), corn steep liquor (0.3%), cysteine (0.26%), and MgSO 4 ·7 H 2 O (0.015%). The pH value was adjusted to 7.2 with 10 M NaOH before autoclaving at 121 °C for 20 min. The antibiotics erythromycin (Em r ) (100 μ g/mL for E.coli TG1, 5 μ g/mL for L. lactis F44) was used for selection. The tolerance agar plates were supplemented with 1.5% agar after autoclaving. Micrococcus flavus ATCC 10240, preserved in the laboratory, was used as an indicator strain for the bioassay of nisin.
Nisin activity assay. A stock solution of nisin was prepared by mixing 0.1 g of nisin standards (Sigma, USA) in 10 mL of 0.02 M HCl (10 6 IU/mL) and boiling for 5 min. The stock solution was diluted using 0.02 M HCl to standard nisin solutions. Nisin standards were added into the autoclaved initial fermentation medium at pH 2.0, 3.0, 4.0, 5.0, 6.0 and 7.0, respectively, and boiling for 5 min. All the fermentation media containing 4000 U/mL nisin (1 g of nisin in 100 mL media) were incubated at 30 °C. The fermentation broth was sampled every 2 h for nisin titer analysis. 500 μ L of fermentation broth was mixed with the same volume of 0.02 M HCl. The mixture was boiled for 5 min and then appropriately diluted with 0.02 M HCl. 1.5% (vol/vol) tween 80 (JiangTian, Tianjin, China) (tween 80 enhances nisin diffusion into the agar medium) and 1% (vol/vol) and 1% (vol/vol) indicator strain M. flavus ATCC 10240 buffer (the final concentration was 10 7 cfu/mL) was poured into the 26-mL assay medium which was cooled to about 45 °C after autoclaving. The medium was poured into a sterile plate for solidification and precultivation. Standard nisin solutions and test solutions were infused into individual wells (80 μ L per well) which had been bored into the assay plate (8 wells per plate) using a 7-mm-diameter metal tube, and then the plates were incubated at 37 °C for 24 h. Zones of inhibition were measured and the regression equation was calculated. Each assay of standard sample or the broth sample was performed in triplicate.
Fermentation performance at acidic condition. The seed medium was inoculated with bacterial colonies and incubated overnight. 5% of the overnight culture broth was used to inoculate fermentation medium for 12 h at 30 °C. The initial pH of fermentation medium were adjusted to 2.0, 3.0, 4.0, 5.0, 6.0 and 7.0, respectively. Cell density (OD 600 ), pH values, and nisin titer of broth samples were measured every 2 h. DNA manipulations and genetic construction. The primers of acid tolerant genes used in the study which were designed by primer premier 5 (Premier, Canada) are listed in Supplementary Table S2. These genes were directly amplified from Lactobacillus casei Zhang, L. lactis subsp. lactis F44 or E. coli DH5α via polymerase chain reaction (PCR). The restriction enzyme cutting sites were simultaneously inserted into the amplified gene. Combinations of two, three or more genes were constructed via overlap extension PCR. The resulting fragments were digested with BamHI and HindIII (or SmaI), and then ligated into plasmid pLEB124, cut with BamHI and HindIII (or SmaI) to generate the resulting plasmids. The resulting plasmids were transformed into E.coli TG1 by heat shock transformation for enrichment 60 . After antibiotics selection, the plasmids were extracted with TIANprep Mini Plasmid Kit (TIANGEN, China), and then were transformed into the L. lactis F44 by electroporation transformation 61 .
Nisin titer assay. The nisin titer assay was described previously 6 . Briefly, the nisin standards (Sigma, USA) and the fermentation broth which have removed the cells by centrifugation at 8000 rpm for 5 min, were boiled for 5 min. Then these samples were serially diluted with 0.02 M HCl. After autoclaving, the 26-mL cooled assay medium was inoculated with 1% (vol/vol) M. flavus buffer, with the concentration of 10 7 cfu/mL. Then the medium was added with 1.5% (vol/vol) tween 80 (JiangTian, Tianjin, China) to enhance nisin diffusion. The mixed medium was quickly poured into sterile plates. After solidification and pre-cultivation, we used a 7-mm-diameter metal tube to drill into the assay plate (8 wells per plate) and removed the agar from the well. Standard nisin solutions and broth solutions were injected into the respective wells (80 μ L per well) and the plates were incubated at 37 °C for 24 h. Inhibition zones were measured by calipers. A regression equation was calculated from the measured data. Each assay of the samples including standards and broth was performed in triplicate.
Scientific RepoRts | 6:27973 | DOI: 10.1038/srep27973 RNA isolation and transcriptional analysis by quantitative real-time PCR. L. lactis strain F44 genome was isolated with the TIANamp Bacteria DNA Kit (TIANGEN). The genome was diluted to concentrations of 10 −5 , 10 −4 , 10 −3 , 10 −2 , 10 −1 . After acid shock at pH 7.0, 6.0, 5.0 or 4.0 respectively, total RNAs were isolated with ZR RNA MiniPrep (The Epigenetics company). For qRT-PCR, 0.1~0.5 μ g of total RNA were reverse transcribed together with 1 μ L corresponding primers in a total reaction volume of 12 μ L using TIANScript RT Kit (TIANGEN). The reaction system was harvested by 3000 rpm for 30 s. Each reaction was incubated at 65 °C for 5 min, followed by incubating on ice. Then the reactions were run at 40 °C for 60 min, followed by 70 °C for 5 min to terminate the complementary DNAs synthesis. The generated cDNAs were stored at − 70 °C for further qRT-PCR. qRT-PCR was performed with Power SYBR Green PCR Master Mix (Applied Biosystems). Briefly, a 50 μ L-reaction solution containing 10~100 ng of cDNA, 200 nM each primer (see Supplementary Table S2), 25 μ L 2 × Ultra SYBR Mixture (with ROX) and sterile water was analyzed on a LightCycler 480 Real-Time PCR System (Roche, Switzerland) according to the manufacturer's instructions. Reactions were run in triplicate in three independent experiments for each condition. qRT-PCR conditions were as follows: 1 cycle at 95 °C for 10 min, 40 cycles of denaturation at 95 °C for 10 s, annealing at 55 °C for 10 s and extension at 72 °C for 20 s. The 16S rRNA gene was used as an internal control to normalize cycle threshold (C T ) values. Difference in the relative expression levels were calculated with 2^− (Δ Δ C T ) method. We performed a 10-fold dilution series experiment using the target assay to establish the standard curve, and the slope of the standard curve can be translated into an efficiency value: PCR efficiency = 10 −1/slope -1 62 . Acid tolerance capacity assay. To investigate the acid tolerance, the wild type strain and the recombination strain after activation culture overnight at 30 °C in the seed medium were harvested in mid-exponential growth phase. Aliquots of cells suspension were diluted to concentrations of 10 −6 , 10 −5 , 10 −4 , 10 −3 and 10 −2 . Cell survival numbers were estimated, where 100 μ L of serially diluted samples were spread in triplicate on seed media agar plates with different pH values 4.0, 4.2, 4.4, 4.6, 4.8 and 5.0, and incubated at 30 °C for 48 h. Plates with colonies in the range of 30 to 300 were then used to calculate the average number of CFU/ml.

Fermentation in flasks.
The original strain F44 and the engineered strains expressing the acid tolerant genes were inoculated in the seed medium for overnight after the cultivation on seed plates at 30 °C for 48 h. The flask fermentation experiments were carried out in 250-mL Erlenmeyer flasks containing 100 mL of fermentation medium with peptone (1.5%), yeast extract (1.5%), sucrose (1.5%), KH 2 PO 4 (2.0%), NaCl (0.15%), corn steep liquor (0.3%), cysteine (0.26%), and MgSO 4 ·7 H 2 O (0.015%). Five milliliters of the overnight cultures were inoculated in triplicate into the static flasks, and incubated for 14 h at 30 °C. Samples were withdrawn every 2 h for cell density analysis, fermentation broth pH, intracellular pH (pH i ), lactic acid concentration, and nisin production.
Fed-batch fermentation. The fed-batch fermentation experiment was conducted at 30 °C for 24 h. Initially the fermentation medium was adjusted to pH 7.2 with 10 M NaOH as described above. The medium was inoculated with 5% of the seed culture. After the fermentation of 6 h, the pH of fermentation broth was controlled at 6.0 by the addition of 10 M NaOH. 3 mL sucrose solution (500 g/L) was added to the fermentation broth at 6 h, 8 h, 10 h and 12 h. The fermentation broth was sampled every 2 h for cell density analysis, lactic acid concentration, residual sugar concentration, and nisin production.

Statistical analysis.
Optical density (OD) was measured at 600 nm with TU-1810 spectrophotometer to monitor the L. lactis cell growth. The pH values of the fermentation broth were measured with FE20 benchtop pH meter (Mettler Toledo, Swiss). To measure the lactic acid concentration and sucrose concentration during the fermentation process, samples were periodically collected for estimating lactate and sucrose concentration in supernatants after removing the cells by centrifugation (8000 rpm, 5 min, 4 °C). The concentration of the residual sugars in the broth was assayed by the dinitrosalicylic acid reagent (DNS) method 64 . The lactate concentration was quantified by a biosensor SBA-90 (Biology Institute of Shandong Academy of Sciences, China) 65 . To assess the statistical significance of differences in resistance to acid stress, student's t tests were used. For the significance of gene expression differences, the statistical t-test was used to identify genes differentially expressed between the wild type strain and the engineered strain. Statistical significance of nisin titer and pH i was estimated using t-test. Statistical analyses of the data were performed using SPSS software version 19.0 (IBM, USA).