Optimization of growth conditions and the inducer concentration for increasing spike protein expression in recombinant Lactococcus lactis and its kinetic modeling

Abstract

Lactococcus lactis bacterium can be genetically modified to transport the spike protein from SARS-CoV-2, making it a potential candidate for a COVID-19 mucosal vaccine. This study aimed to optimize the nisin concentration, pH, incubation time, and media composition to induce spike protein expression. The concentrations of nisin used in this study ranged from 0 to 40 ng/mL, the incubation period was 3 to 24 hr, and the pH of the growth media ranged from 4 to 8. The media was also supplemented with various yeast extract and sucrose concentrations. The highest protein band intensity was observed at a concentration of 40 ng/mL and an incubation period of 9 hr. Supplementation with 4% w/v yeast extract and 6% w/v sucrose significantly increased the expression of HCR spike protein. In silico simulation suggested a maximal protein band intensity of 70.95 arbitrary units, while the nisin concentration needed to produce half the maximal protein band intensity was estimated to be 9.599 ng/mL. No significant difference in spike protein expression was found between pH variations. The media composition, inducer, and incubation time strongly affect the spike protein expression.

INTRODUCTION

The coronavirus infectious disease 2019 (COVID-19) pandemic has highlighted the importance of vaccination in the prevention of infectious diseases [1]. Currently, the only vaccine available is one that is administered via an invasive intramuscular route. Vaccines administered via the mucosal route, such as intranasal and oral routes, are less invasive and can stimulate the formation of the mucosal immune system, which is critical as a defense against viral infections in the respiratory tract [2,3,4].

The immune system is a complex system comprised of organs, tissues, and special cells that collaborate to protect and defend the body. Vaccines generally contain specific antigens that enhance the body’s immune response because they induce memory cells to work faster in recognizing and protecting the body from future antigen attacks. A vaccine is the most effective way to prevent the spread of infectious diseases caused by bacteria and viruses [5].

The severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) genome encodes four major structural proteins: the spike (S), envelope (E), membrane (M), and nucleocapsid (N) protein. The coronavirus structure is shaped like a cube, with the S protein on the virus’s surface [6]. The S protein is split into two subunits, S1 and S2, responsible for invasion and attachment to host cells. S1 initiates viral attachment by attaching to angiotensin-converting enzyme 2 (ACE-2) on the host cell via the receptor-binding domain. This process causes a conformational change in S2, resulting in host-virus fusion and virus entry into the cell [7,8,9].

In silico research by Yurina identified nine spike protein sequences from various coronavirus strains, ranging from severe acute respiratory syndrome coronavirus (SARS-CoV) to SARS-CoV-2 [10]. Her research revealed the presence of a highly conserved region (HCR) at amino acid sequence 945-1100. HCR epitope prediction tests on B and T cells revealed several epitopes recognized by the two cell types. As a result, this discovery can serve as a starting point for further research into developing epitope-based coronavirus vaccine candidates found in HCR sequences. Our previous study suggests that such a vaccine is a promising candidate as a COVID-19 mucosal vaccine [10, 11].

Lactococcus lactis is a lactic acid bacterium with various applications, including being used as a fermented milk starter and a cheese ingredient. The entire genome of L. lactis has been sequenced and broadly studied. This bacterium multiplies at a temperature of 30°C, has a doubling time of 35 to 60 min, and has fermentative or respiratory metabolism [12]. Its potential as a host for the overexpression of homologous and heterologous proteins has also been investigated, even though it is primarily used in the food industry to produce fermented foods. L. lactis is simple and affordable to grow, and a wide range of genetic techniques and vector systems are accessible and well established for it. The bacterium can be genetically manipulated to produce and secrete a variety of proteins, and it has been extensively studied for its ability to deliver therapeutic proteins/antigens into mucosal tissues, mainly through intranasal, oral, or genital administration as a therapy or vaccine [13,14,15,16,17,18].

Nisin is a 34-amino acid antimicrobial peptide produced by lactic acid bacteria to prevent the growth of Clostridium botulinum spores in cheese. It is a food-grade compound that, at nanograms per milliliter, can induce the production of homologous and heterologous proteins, including intracellular, extracellular, and cell wall-bound enzymes, as well as toxic proteins and viral and bacterial antigens [19]. Gene expression conditions are greatly influenced by various factors, including incubation temperature, growth media pH, and the number of inducers. Nisin can also regulate the gene expression system in Gram-positive bacteria such as L. lactis, and this is known as the nisin-controlled gene expression (NICE) system [20]. This study aimed to determine the optimum growth conditions for spike protein expression in L. lactis.

Optimizing bacterial culture conditions to produce protein expression, such as in COVID-19 vaccine development research, requires adjusting the composition of the growth medium. The results of research by Mierau et al. [12] showed a strong correlation between media composition, culture cell density during nisin induction, and the amount of nisin added. Nutritional supplements of 7% lactose, 2.5% peptone, 2% yeast extract, and 0.01% sodium phosphate increased the production of lysostaphin, an antibacterial protein belonging to a major class of antimicrobial peptides and proteins that is also active in lactic acid bacteria [12]. Another study by Vijayakumar et al. showed that the addition of yeast extract (6% b/v) as a nitrogen source and sucrose (4% b/v) as a carbon source could increase nisin production in L. lactis [21].

Yeast extract, also known as yeast hydrolysate, is recognized as a Generally Recognized as Safe (GRAS) ingredient by the United States Food and Drug Administration (FDA). Its popularity is increasing due to its low production cost, abundant and affordable supply of raw materials, and economic efficiency. Yeast extract is composed of amino acids, lipids, vitamins, minerals, and other soluble components, making it an important component in the food industry as a flavoring, additive, vitamin supplement, and nutrient source for bacterial culture media [22]. L. lactis requires external nitrogen sources, such as yeast extract, for protein production, including nisin, which is an antimicrobial peptide. The choice of nitrogen source affects nisin production by L. lactis, with yeast extract and peptone being good choices [21, 23]. However, the molecular mechanisms in L. lactis are not fully understood, although proteolytic biomolecular schemes in Lactobacillus spp. provide an overview of peptide hydrolysis processes relevant to recombinant protein production.

Sucrose is one of the primary carbon sources for lactic acid bacteria and is hydrolyzed into glucose and fructose by β-d-fructosidase. Sucrose is considered a more effective carbon source than glucose in nisin production in L. lactis because it does not cause glucose repression, which reduces the rate of gene transcription [24]. The effect of different types of carbon nutrients on nisin production has been studied, and it has been shown that sucrose gives the best results for increasing nisin production [25].

MATERIALS AND METHODS

Bacterial culture

L. lactis NZ3900 bacteria (without plasmid; MoBiTec GmbH, Goettingen, Germany) served as a negative control, and L. lactis pNZ8149-HCR (L. lactis pNZ-HCR) bacteria were grown on M17 Broth (HiMedia, Thane West, India) with glucose supplementation (Santa Cruz Biotechnology, Dallas, TX, USA; G-M17B) and lactose supplementation (Santa Cruz Biotechnology, Dallas, TX, USA; L-M17B) at a final concentration of 0.5%. The bacteria-containing media were then incubated at 30°C overnight. The bacteria grown on the media were then inoculated with the fresh G-M17B and L-M17B media at 4% v/v. A UV-Vis spectrophotometer at a wavelength of 600 nm was then used to determine bacterial growth (OD600) per unit of time. In each treatment group, nisin (MoBiTec GmbH, Goettingen, Germany) was added at concentrations of 0, 10, 20, and 40 ng/mL to induce protein expression (Supplementary Table 1). Each medium was then incubated overnight at 30°C. The optimal nisin concentration based on the highest protein expression level was then used to optimize the incubation time using the same procedure with 3, 6, 9, and 24 hr of incubation (Supplementary Table 2). Furthermore, bacterial culture was carried out with pH variations of pH 6, 7, and 8 (Supplementary Table 3). Media supplementation was varied using the addition of sucrose and yeast extract (Supplementary Table 4). Bacteria were harvested by centrifuge at 3,000 rpm for 10 min at 4°C. The pellets were centrifuged again after being resuspended in phosphate buffered saline (PBS). The pellets were then resuspended in PBS with lysozyme and sonicated for five cycles of 30 sec each before being treated with a protein inhibitor cocktail. Cells were harvested by centrifugation at 12,000 rpm for 3 min.

Enzyme-linked immunosorbent assay (ELISA)

Standard SARS-CoV-2 spike protein (GenScript Biotech, Piscataway, NJ, USA) and samples were coated on ELISA plates and incubated overnight at 4°C. The wells were subjected to three 5-min washes with Tris-Buffered Saline Tween^TM 20 (TBST) before being treated with 0.5% bovine serum albumin (BSA)/TBST. As before, the plates were incubated overnight. Wells were washed with TBS-T before adding 0.5 g/mL anti-SARS-CoV-2 spike protein (GenScript Biotech, Piscataway, NJ, USA) primary antibody and incubating the plates for 60 min at room temperature. After washing with TBST, a secondary antibody biotinylated rabbit anti-human antibody (1:100.000; Sigma, St. Louis, MO, USA) was added, and the plates were incubated for 60 min at room temperature. The wells were rewashed with TBST, and streptavidin-horseradish peroxidase (HRP) was added. The substrate was then washed with TBST and added to each 3,3′,5,5′-tetramethylbenzidine (TMB) well, and the absorbance was measured at 450 nm.

Western blot

Sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) was used to separate lysate samples from each group. The proteins were then transferred to a nitrocellulose membrane using Semi Dry Trans Blot (Bio-Rad, Hercules, CA, USA) and blocked overnight with 0.5% BSA/TBST. The membrane was subjected to three 5-min washes with TBST before being incubated for 60 min at room temperature with anti-SARS-CoV-2 spike protein primary antibody (K009416P, Solarbio, Beijing, China; 1:5.000). The membrane was then washed with TBST before being incubated for 60 min at room temperature with a secondary antibody biotinylated rabbit anti-human antibody (Cat no. 31820, Thermo Fisher Scientific, Waltham, MA, USA; 1:200.000). The membrane was subsequently washed again with TBST before being incubated with streptavidin-HRP (Cat no. OR03L-200UG, EMD Millipore, Billerica, MA, USA). After another 40 min of incubation, the membrane was washed with TBS-T and added as a substrate to each TMB (Product No. T0565-100ML, Sigma–Aldrich, St. Louis, MO, USA) membrane well. The reaction was stopped with sterile distilled water. ImageJ was used to examine band intensity.

Data analysis

HCR spike protein concentration data from ELISA samples were statistically analyzed with the SPSS 16.0 software (SPSS Inc., Chicago, IL, USA) and a significance level of 0.05. The analysis was carried out by first measuring normality with the Shapiro–Wilk test, then assessing homogeneity with Levene’s test, and finally analyzing the data with one-way ANOVA and post hoc Tukey’s HSD.

HCR spike protein band intensity data from the Western blot analysis were used to model the effect of the nisin concentration on band intensity using the Michaelis–Menten type equation:

$$\mathrm{B}=\mathrm{B\_}0+\left(\mathrm{B\_max}–\mathrm{B\_}0\right)·\mathrm{C}/\left(\mathrm{EC}50+\mathrm{C}\right), \left(1\right)$$

where B is the band intensity; B_0 and B_max are the lowest and highest band intensities, respectively; C is the nisin concentration; and EC50 is the concentration necessary to produce half the maximum band intensity. To fit the model to the data, all quadruplicate band intensity measurements were used in the non-weighted least-squares minimization process. The simulations were performed in the GraphPad Prism 9.5.1 software (GraphPad Software, Boston, MA, USA).

RESULTS

ELISA and Western blot were chosen due to the main principles of the two methods, which both use antigen-antibody binding, their high specificity and sensitivity, their ease of application, and their widespread use [26, 27]. The combination of these two methods can quantify the protein in a sample compared with a standard and confirm the identity of the analyzed protein based on molecular weight [28]. The amount of HCR protein produced by L. lactis pNZ-HCR after 40 ng/mL nisin induction was significantly higher than that produced by the control (p<0.05), as shown in Table 1 and Fig. 1.

Table 1. Highly conserved region (HCR) protein concentration according to nisin concentration as assayed by ELISA.

Nisin concentration (ng/mL)	Mean HCR spike protein concentration ± SD (µg/mL)
0	63,071 ± 6,950
10	44,063 ± 1,711
20	55,837 ± 1,281
40	76,830 ± 3,382*

*p<0.05.

Fig. 1.

Expression levels of highly conserved region (HCR) spike with various nisin concentrations as the inducer.

The bar chart shows quantification of the protein levels under each condition. Error bars show the standard deviation. *p<0.05.

The protein bands that appeared were estimated to have molecular weights of 23.46 kDa based on the Western blot results, with the highest color intensity shown by the recombinant bacteria L. lactis pNZ-HCR with 40 ng/mL nisin induction. The appearance of protein bands in this area corresponds to the appearance of protein bands in Western blot.

As shown in Table 1, the protein bands produced by the bacterial cell lysates of L. lactis pNZ-HCR that were not induced by nisin showed protein expression similar to the HCR protein expression. The NICE system demonstrated basal expression activity even without the addition of nisin, as protein bands can be produced by L. lactis pNZ-HCR bacterial cell lysates [29].

An HCR protein band with a molecular weight of 23.07 kDa was also observed and appeared to be nearly identical to the value in previous tests on optimizing nisin concentrations. Furthermore, a protein band with a molecular weight of 16 kDa was found in the negative control tested (Fig. 2, lane e). This demonstrates that the negative control, L. lactis NZ3900, did not express HCR protein like the L. lactis pNZ-HCR recombinant bacteria. However, the presence of a band in the negative control demonstrated a cross-reaction between the antibody and an endogenous protein of L. lactis NZ3900. Other conditions that affect Lactococcus bacteria include the pH of the media. However, our result suggests that pH variation in the media did not significantly increase the expression of HCR protein (Fig. 3). Aside from optimizing the nisin concentration and incubation time applied to L. lactis pNZ-HCR, more testing is required for other variables that affect gene expression systems.

Fig. 2.

Representative results of Western blot for the various incubation durations.

There is no significant difference in highly conserved region (HCR) protein expression according to the various incubation durations. A: M=marker; a–d=HCR spike protein expression, harvesting after incubation for 3, 6, 9, and 24 hr; e=bands from the L. lactis control, harvesting after 3 hr of incubation. B: band intensity analyzed using ImageJ.

Fig. 3.

Band intensity of highly conserved region (HCR) protein after incubation in media at various pH levels.

After incubation in media at various pH levels, the expression of HCR is not significantly different.

Optimization of the nutritional supplements in the M17 media (pH 4, without adjustment) with yeast extract as a nitrogen source and sucrose as a carbon source is shown in Figs. 4 and 5. The highest color intensity protein bands were shown by the recombinant bacteria L. lactis pNZ-HCR when supplemented with 4% w/v yeast extract and 6% w/v sucrose (Fig. 6). Supplementation with both yeast extract and sucrose showed the highest protein expression compared with supplementation with yeast or sucrose only.

Fig. 4.

Representative results of Western blot for the variations of yeast extract supplementation.

There is a significant difference in highly conserved region (HCR) protein expression in the various combinations of yeast extract (A) and sucrose (B) supplementation. N=negative control of L. lactis NZ3900; M=marker; A=HCR spike protein expression in media without nisin induction; B=HCR spike protein expression with 40 n/mL nisin in media without nutrition variation; C–E=HCR spike protein expression in media variations with 4%, 6%, and 8% w/v yeast extract; F–H=HCR spike protein expression in media variations with 2%, 4%, and 6% w/v sucrose.

Fig. 5.

Expression levels of highly conserved region (HCR) spike for the variations of yeast extract (A) and sucrose (B) nutrient supplementation.

The bar chart shows the quantification of protein levels under each condition. Error bars show the standard deviation. *p<0.05.

Fig. 6.

A. Representative results of Western blot for the variations of combined yeast extract and sucrose supplementation.

There is a significant difference in highly conserved region (HCR) protein expression among the variations of nutritional supplementation. A: N=negative control of L. lactis NZ3900; M=marker; A=HCR spike protein expression in media without nisin induction; B=HCR spike protein expression with 40 n/mL nisin in media without variation in nutrition; I=HCR spike protein expression in the media variation with 4% yeast extract; J=HCR spike protein expression in the media variation with 6% sucrose; K=HCR spike protein expression in the media variation with the combination of 4% w/v yeast extract and 6% w/v sucrose. B. The bar chart shows the quantification of protein levels under each condition. Error bars show the standard deviation. *p<0.05.

DISCUSSION

A series of analyses revealed that L. lactis pNZ-HCR with 40 ng/mL nisin induction for 9 hr gave the highest yield of HCR protein (Fig. 1 and Table 1). When administered to bacteria via the NICE system, nisin binds to the N-terminus of NisK, causing it to autophosphorylate. NisK then transfers its phosphate group to NisR and becomes a PnisA transcriptional activator [30]. The NICE system has been extensively used for over-expression and subsequent functional and structural studies of various proteins because it has been well characterized and is incredibly flexible. The expression of various proteins at multiple points during the growth cycle is now possible thanks to a recent combination of the NICE system and ZIREX system [30].

Moreover, the NICE system has also been applied to other Gram+ bacteria, including Leuconostoc lactis, Lactobacillus helveticus, Lactiplantibacillus plantarum, Streptococcus pyogenes, Streptococcus agalactiae, Streptococcus pneumoniae, Streptococcus zooepidemicus, Enterococcus faecalis, and Bacillus subtilis. Although regulated gene expression has been established in many instances, the introduction of a unique dual plasmid system, the varying nisin sensitivities of strains, the RNA polymerase sequence, and other factors that affect protein expression have all been shown to slow the growth of several species [20].

At various concentrations, yeast extract resulted in a significant increase in Spike HCR protein expression compared with the negative control. The optimal concentration was found to be 4% (w/v) yeast extract, with protein expression reaching the highest expression compared with the other groups (p<0.05). Yeast extract provides essential nutrients for bacterial growth and protein biosynthesis, such as amino acids as basic components for protein synthesis, vitamins and minerals for enzymatic activation and metabolic processes, and nucleotides for DNA/RNA synthesis and support of plasmid replication and gene transcription [22]. Nutritional components, especially amino acids resulting from the enzymatic processing of the peptides, may directly support the synthesis of spike HCR proteins, thereby increasing translational efficiency and allowing cells to produce more target proteins. However, the specific regulatory mechanism of amino acids involved in inducing spike HCR proteins still requires further research.

The addition of sucrose also had positive effects on protein expression, with the optimal concentration being 6% (w/v). At this concentration, there was a significant increase in the amount of protein expressed compared with the negative control in the M17 media (p<0.05). Sucrose is transported into cells via the sucrose phosphotransferase system (PTS), in which sucrose is converted into sucrose-6-phosphate and broken down into glucose-6-phosphate and fructose. These metabolic products are fed into the glycolysis pathway to produce ATP and other metabolic precursors essential for cellular activity [31]. This additional energy supports protein synthesis, including mRNA translation into protein. Therefore, the addition of sucrose can increase protein expression [32, 33]. However, the specific regulatory mechanism of sucrose involved in inducing Spike HCR proteins still requires further research.

Based on the results for the optimal yeast extract and sucrose concentrations, we continued testing the effect of a combination of yeast extract and sucrose on spike HCR protein expression compared with either supplement individually. The highest color intensity protein bands were shown by the recombinant bacteria L. lactis pNZ-HCR when supplemented with 4% w/v yeast extract and 6% w/v sucrose in combination, as shown in Fig. 6.

The combination of 4% yeast extract and 6% sucrose resulted in a more significant increase in protein expression than using yeast extract or sucrose alone. This suggests a synergistic effect of the combination of the two supplements. Although the use of yeast extract or sucrose only increased protein expression, combining the two proved more effective. There was a significant difference between L. lactis NZ3900 with the combination of 4% yeast extract and 6% sucrose compared with L. lactis NZ3900 with 4% yeast extract (p<0.05) or 6% sucrose (p<0.05) only. This suggests that complex media with such nutrient combinations provide abundant nutrient availability for protein expression [34, 35].

To obtain more quantitative information about the strength of L. lactis induction with nisin, we fitted our measured data to the Michaelis–Menten equation (1). This equation was developed by Michaelis and Menten [36] for an enzymatic system of two reactions in which a substrate is reversibly attached to the enzyme to form an enzyme-substrate complex and then irreversibly reacts to form a product and rebuild the original enzyme [37, 38]. While it is among the most well-known and useful modeling approaches for enzyme kinetics, it is also popular for many other biochemical reactions involving molecule binding, like antigen-antibody binding, protein-protein interaction, DNA-DNA hybridization, and gene induction with transcriptional factors. In our situation, it predicts the influence of the nisin concentration on the spike protein band intensity.

Figure 7 displays the best least-squares fit (solid curve, R=0.7967) of our data to the Michaelis–Menten equation (1). The estimated parameter values are given in Table 2; they are all statistically significantly different from zero. The predicted minimal protein band intensity, which lies roughly in the range of 10 to 30 arbitrary units, confirms our earlier observation that even with vanishing nisin concentrations, some induction takes place spontaneously. The predicted maximal protein band intensity of around 70 arbitrary units seems to correspond roughly to the optimal nisin concentration of 40 ng/mL. The nisin concentration necessary to produce half the maximal protein band intensity is estimated to be 9.599 ng/mL.

Fig. 7.

Michaelis–Menten prediction of the band intensity of highly conserved region (HCR) protein depending on the nisin concentration.

Black dots represent means of quadruple measurements, and whiskers represent the standard deviation. The dotted lines represent the 95% confidence interval.

Table 2. Estimated values of the parameters in the Michaelis–Menten equation (1).

Parameter	Estimated value	95% confidence interval
B_0	20.34	[11.18, 29.43]
B_max	70.95	[60.41, 85.69]
EC50	9.599	[2.959, 24.64]

The ability of nisin to induce HCR protein gene expression in the recombinant bacteria L. lactis pNZ-HCR was confirmed by detection of a protein band size of 23 kDa by Western blot. In comparison with the other groups, induction of L. lactis pNZ-HCR with 40 ng/mL nisin and incubation for 9 hr produced the optimal HCR protein expression. Michaelis–Menten kinetics-based simulations predicted that the highest HCR protein band intensity would be achieved at a nisin concentration of around 40 ng/mL as well. Supplementation of M17 media with yeast extract and sucrose significantly enhances the expression of Spike HCR protein in L. lactis NZ3900/pNZ8149-HCR induced with 40 ng/mL nisin. Nutritional supplementation with yeast extract optimally enhances Spike HCR protein expression at a concentration of 4% w/v, while sucrose shows optimal enhancement at 6% w/v under the same induction conditions. Furthermore, combined supplementation with 4% yeast extract and 6% sucrose significantly increases the expression of Spike HCR protein, demonstrating a higher effect compared with the use of either 4% yeast extract or 6% sucrose only.

DATA AVAILABILITY

The data underlying this article will be shared on reasonable request to the corresponding author.

FUNDING

This research was partly supported by the National Agency for Research and Innovation (Badan Riset dan Inovasi Nasional [BRIN]) through a Research and Innovation for Advanced Indonesia (RIIM) grant (No. 82/II.7/HK/2022) and the Indonesia Endowment Fund for Education (Lembaga Pengelola Dana Pendidikan). The work of Jurjen Duintjer Tebbens was supported by UNCE no. UNCE/24/MED/008 of Charles University (https://cuni.cz/UKEN-1.html) and by long-term strategic development financing of the Institute of Computer Science (RVO: 67985807) of the Czech Academy of Sciences.

AUTHOR CONTRIBUTIONS

Termidzi Husni Mubarak: investigation, writing-original draft; Silvia Maulita: investigation, writing-original draft; Oktavia Rahayu Adianingsih: validation, formal analysis; Jurjen Duintjer Tebbens: methodology, software, writing-review and editing; Takeshi Shimosato: methodology, writing-review and editing; Valentina Yurina: conceptualization, supervision, writing-review and editing.

CONFLICT OF INTEREST

There is no conflict of interest declared.