The Influence of Nutrient Medium Composition on Escherichia coli Biofilm Development and Heterologous Protein Expression

Abstract

In the present study, the effects of different nutrient media on the development of Escherichia coli biofilms and the production of a heterologous protein were examined. E. coli JM109(DE3) cells transformed with pFM23 plasmid carrying the gene for enhanced green fluorescent protein (eGFP) expression were used. Cells were grown in two different culture media, Lysogenic Broth (LB) and M9ZB, in a flow cell system for 10 days. Epifluorescence microscopy, fluorimetry, and a high-performance liquid chromatography (HPLC) method based on hydrophobic interaction chromatography (HIC) were used to assess bacterial growth, plasmid copy number (PCN), and eGFP production in both planktonic and biofilm cells. The results showed that biofilm development was favored in M9ZB medium when compared with LB. However, the number of eGFP-expressing cells was higher in LB for both planktonic and sessile states (two-fold and seven-fold, respectively). In addition, the PCN in biofilm cells was slightly higher when using LB medium (on average, 29 plasmids per cell versus 20 plasmids per cell in M9ZB), and higher plasmid stability was observed in biofilms formed in LB compared to their planktonic counterparts. Hence, E. coli biofilms grown in LB enhanced both plasmid stability and capacity to produce the model heterologous protein when compared to M9ZB.

1. Introduction

Escherichia coli has been one of the most used organisms for heterologous protein production due to its inexpensive and fast high-cell density cultivation, well-known genetics, and large number of cloning vectors [1,2,3]. In addition to the host strain, factors such as the expression vector, cultivation conditions (medium composition, temperature, pH, and induction conditions), and purification strategies are essential for the successful production of heterologous proteins [4].

E. coli plasmids have been the preferred expression vectors because of the high gene dosage that can be attained and the capability to control the levels of gene expression through inducible promoters [5]. Although a high plasmid copy number (PCN) may ensure a high gene dosage, this does not always mean a high level of target heterologous protein [5,6]. Studies in suspension cell cultures have demonstrated that high plasmid copy number imposes a metabolic burden on the host cell, which reduces the bacterial growth rate and favors plasmid instability, thereby decreasing the overall protein yield [7,8].

The use of biofilm cells in recombinant protein production has been scarcely studied, but some promising results have been obtained by our research group in the last few years [6,9,10,11]. The biological organization of biofilms provides them with many advantages over cells grown in suspension cell cultures, including high cell density, protection against hostile conditions, and operation stability, particularly when biofilms are used in biotechnological processes for the production of value-added compounds [12]. In addition, the presence of plasmids has been shown to stimulate biofilm formation and lead to a higher production of recombinant proteins compared to planktonic cells [10,11,12,13,14,15]. Our group has demonstrated that cells expressing eGFP (enhanced green fluorescent protein) formed more biofilm than non-expressing cells [10]. Moreover, E. coli biofilms were able to produce eGFP at a much higher level (30-fold) than their planktonic counterparts [10]. Furthermore, it has been shown that biofilm cultures have a beneficial effect on the maintenance of high and medium copy number plasmids [6,16]. O’Connell et al. [16] reported that continuous biofilm cultures of E. coli ATCC 33456 containing the pEGFP plasmid (a pUC family vector) maintained a high PCN over extended periods, while in chemostat a significant plasmid loss was observed. More recently, we showed that although the PCN was higher in planktonic cells, a significant PCN reduction (around 89%) was observed in planktonic cells after 5 days of flow cell operation, while no plasmid loss was observed in E. coli JM109(DE3) biofilms [6].

Different strategies have been explored to achieve maximum yields of heterologous proteins in suspended cell reactors, including by (1) changing the expression vector and/or the host, (2) optimizing the culture parameters (medium, pH, and temperature) and induction conditions (inducer concentration and temperature) of the recombinant strain, (3) co-expressing other genes such as chaperones, and (4) adjusting the synthesis of stress-related sigma factors and/or deleting specific gene sequences such as proteases, which may promote both the quantity and quality of the desired protein [2,3,5,17,18,19]. Nonetheless, the investigation of productive biofilms remains scarce. One of the parameters that can affect both biofilm formation and the level of heterologous protein expression is the culture medium composition. Several reports have demonstrated that the nutritional medium content affects biofilm formation [11,20,21,22]. Some authors have shown that an increase in nutrient levels promotes biofilm formation [12,13]. However, contradictory results showing that high nutrient concentration inhibits biofilm formation were also reported [14,15]. Regarding the effect of nutrient medium composition on heterologous protein production, most of the studies were performed in planktonic conditions [23,24,25]. In these conditions, it has been demonstrated that the increased concentration of some salts, peptone, and yeast extract can increase the production of the desired protein [26,27]. Furthermore, Matsui et al. [28] showed that the use of tryptone and yeast extract can increase plasmid stability. However, it is known that the high-level expression of heterologous protein in suspended cells often promotes a metabolic burden in the host cell that can affect plasmid stability and consequently the protein yield [29].

The main aim of this study was to determine the impact of different nutrient media compositions on biofilm development, plasmid stability, and eGFP expression in E. coli biofilm and planktonic cells. For that, E. coli JM109(DE3) cells transformed with the pFM23 plasmid for eGFP expression were used. Biofilm formation in two growth media, Lysogenic Broth (LB) and M9ZB, was carried out in a flow cell system for 10 days. LB medium is a rich medium that is frequently used for E. coli growth and expression of heterologous proteins [30]. M9ZB medium is the combination of M9 and ZB media, which are a minimal growth medium used for bacterial cultures and a medium recommended for the preparation of competent E. coli cells, respectively. M9ZB is also used for the cultivation of E. coli recombinant strains [31]. As far as we know, this is the first time that the M9ZB medium has been tested for biofilm formation and heterologous protein expression in sessile cells. On the other hand, this work enriches the group’s previous studies [6,11] with the assessment of the population of planktonic and biofilm cells that produce the protein of interest by a single-cell analysis, and the evaluation of plasmid stability by determining the PCN in both types of cells by a chromatographic method.

2. Materials and Methods

2.1. Bacterial Strain

The pFM23 plasmid (created from pET28A, Novagen, Madison, WI, USA) containing the eGFP gene was transformed by heat shock in E. coli JM109(DE3) (Promega, Madison, WI, USA) for the intracellular production of eGFP [32]. Transformants were selected on agar medium supplemented with 20 μg mL⁻¹ kanamycin (Eurobio, Orsay, France).

2.2. Flow System

A flow cell system (Figure 1) containing a recirculating tank (planktonic cells) and a flow cell composed of removable coupons with polyvinyl chloride (PVC) slides for biofilm formation was operated as described by Gomes et al. [9]. Briefly, E. coli cells were recirculated at 30 °C for 10 days under turbulent flow conditions (Reynolds number = 4600) [9], which are the hydrodynamic conditions previously used by the group [9,10,22,33] and completely characterized by computational fluid dynamics [34]. The tank was aerated and continuously fed with fresh medium at a flow rate of 0.025 L h^–1 [9].

To determine the effect of culture medium on biofilm formation and eGFP expression, two different media were tested: LB-Miller (Sigma, St. Louis, MO, USA) and M9ZB medium. According to the supplier, LB-Miller is made of 10 g L⁻¹ tryptone, 5 g L⁻¹ yeast extract, and 10 g L⁻¹ NaCl. The M9ZB medium was prepared as indicated by the Novagen pET System manual and Studier et al. [30,35]. The mixture of 10 g L⁻¹ N-Z-amine A and 5 g L⁻¹ NaCl was first prepared. Then, 100 mL of 10× M9 salts solution (10 g L⁻¹ NH₄Cl, 30 g L⁻¹ KH₂PO₄, 60 g L⁻¹ Na₂HPO₄·7H₂O), 1 mL of 1 M MgSO₄ solution, and 10 mL of 400 g L⁻¹ glucose solution were added. All solutions were prepared and autoclaved separately. All compounds were purchased from Merck, Algés, Portugal. The comparison of carbon and nitrogen sources and amounts between media are presented in Table S1 in Supplementary Material. Both liquid media were supplemented with 20 μg mL^–1 kanamycin.

2.3. Biofilm and Planktonic Analysis

The system was stopped every day to allow biofilm sampling and restarted using the same flow rate. The biofilm thickness was first determined by using a digital micrometer [22]. Then, it was resuspended in 25 mL of 8.5 g L⁻¹ NaCl solution for total and culturable cell quantification, and eGFP and plasmid analysis. For biofilm culturability, samples were spread on solid growth medium (PCA; Merck, Algés, Portugal) supplemented with 20 μg mL⁻¹ kanamycin, and colony enumeration was performed after incubation at 30 °C. Biofilm total cells were determined with the Live/Dead^® BacLight™ bacterial viability kit (Invitrogen Life Technologies, Alfagene, Carcavelos, Portugal) as described in Gomes et al. [9]. The final values of biofilm total and culturable cells were expressed as log cells cm⁻² and log CFU cm⁻², respectively.

For planktonic fraction, a 10 mL sample was taken from the recirculating tank daily, and total and culturable cell numbers were assessed by the same methodology as biofilm cells. The final values of planktonic total and culturable cells were presented as log cells mL⁻¹ and log CFU mL⁻¹, respectively.

2.3.1. Quantification of eGFP-Expressing Cells and Specific Protein Concentration

The number of eGFP-expressing cells in planktonic and biofilm cells was assessed by epifluorescence microscopy as indicated by Gomes et al. [9]. The percentage of eGFP-expressing cells was calculated by dividing the number of expressing cells by the total cell number.

The specific eGFP production was analyzed by fluorimetry as fully described by Mergulhão and Monteiro [32]. The eGFP fluorescence was measured using a microplate reader (SpectraMax M2E, Molecular Devices, Inc., Wokingham, UK) with an excitation filter of 488 nm and an emission filter of 507 nm. Then, the eGFP concentration was calculated through a standard curve obtained from purified eGFP serially diluted (range of 0 to 0.31 µg µL⁻¹) and the final values were presented as specific eGFP production (fg cell⁻¹).

2.3.2. Plasmid Extraction and Quantification

Plasmid DNA was extracted by alkaline lysis as described by Alves et al. [36]. Briefly, bacterial pellets were resuspended in 100 µL of Solution I (50 mM glucose, 25 mM Tris–HCl, 10 mM EDTA, pH 8). Alkaline lysis was performed by adding 100 µL of Solution II (0.2 M NaOH, 1% (w/v) SDS) and incubating the mixture for 10 min at room temperature. Cellular debris, genomic DNA and proteins were precipitated by adding 100 µL of Solution III (5 M acetate/3 M potassium, pH 5) and incubating for 10 min in ice. The precipitate was removed by centrifuging twice at 20,000× g for 30 min at 4 °C (Centrifuge 5424, Eppendorf, Hamburg, Germany).

Plasmid quantification in biofilm and planktonic samples was performed according to the analytical high-performance liquid chromatography (HPLC) method based on hydrophobic interaction chromatography (HIC) described by Diogo et al. [37]. A 4.6/100 mm HIC source 15 PHE PE column (Amersham Bioscience, Uppsala, Sweden) was connected to an HPLC system (Shimadzu, Nexera-i LC-2040C 3D, Duisburg, Germany) and equilibrated with 1.5 M (NH₄)₂SO₄ in 10 mM Tris–HCl buffer, pH 8. Then, 50 µL of each sample was injected and eluted at 1 mL min⁻¹. After 1 min, the elution buffer was changed to 10 mM Tris–HCl without ammonium sulfate and maintained for 0.8 min in order to elute all bounded species. Finally, the elution buffer was changed to the equilibration buffer and maintained for 8.2 min to re-equilibrate the column for the next run. The absorbance of the eluate was recorded at 260 nm. The plasmid concentration of the samples was calculated using calibration curves constructed with serial dilutions of plasmid DNA standards (ranging from 0.04 to 20 µg mL⁻¹). The calibration curves were established among the peak area appearing between 0.67 and 0.7 mL of elution volume versus the concentration of plasmid DNA standard [37]. Then, the plasmid copy number (PCN) was calculated using Equation (1) [38]:

$$\mathrm{PCN}=\frac{6.02\times {10}^{23} \left({\mathrm{copy} \mathrm{mol}}^{-1}\right)\times \mathrm{DNA} \mathrm{amount} \left(\mathrm{g}\right)}{\mathrm{DNA} \mathrm{length} \left(\mathrm{bp}\right)\times 660 \left({\mathrm{g} \mathrm{mol}}^{-1}{ \mathrm{bp}}^{-1}\right)}$$

(1)

where the size of pFM23 plasmid is 6053 bp.

2.4. Statistical Analysis

The results originated from averages of independent experiments for each culture condition (LB and M9ZB). All reported data are presented as mean ± standard deviation (SD) from at least three experiments with triplicates.

Paired t-test analysis was conducted to determine the differences between culture media on each experimental day. Statistically significant differences were reported for p values < 0.1 (corresponding to a confidence level of 90%) and <0.05 (corresponding to a confidence level of 95%).

3. Results

In this study, two different nutrient media were used with the aim of comparing eGFP production and plasmid stability in biofilm and planktonic populations growing in a flow cell system.

The planktonic fraction was monitored by quantifying the total and culturable cells (Figure 2A,C, respectively). Analysis of the planktonic total cells (Figure 2A) indicated that there were very few statistical differences in E. coli growth in both media (only one of the eight experimental points has p < 0.05). Despite the few statistical differences, from day 5 onwards, it seems to be a higher cell number of bacteria when growing in LB medium, with on average 29% more cells being present in LB than in M9ZB until the end of the experiment. Regarding the number of planktonic culturable cells (Figure 2C), similar behavior was observed for both media, but slightly higher CFU mL⁻¹ values were obtained in M9ZB in all experimental points (on average 62%) when compared to LB.

The impact of nutrient medium composition on biofilm development was assessed by quantifying total and culturable cells and determining the biofilm thickness (Figure 2B,D,E, respectively). Globally, the amount of biofilm total cells (Figure 2B) was higher for M9ZB (on average 8.25 log cells cm⁻²) than for LB (on average 8.0 log cells cm⁻²) with statistically significant differences for five of the eight experimental points (p < 0.1). When comparing biofilm culturable cells (Figure 2D), higher values were also obtained for M9ZB (on average 65%) with statistical differences in most time points (p < 0.1 for days 6 and 10, and p < 0.05 for days 3, 5, 7 and 9). Biofilm thickness (Figure 2E) was similar between growth media until day 5, but thereafter, the biofilms developed in M9ZB were thicker than biofilms formed in LB (176 µm versus 161 µm, respectively), which is most likely a result of their higher cell density (Figure 2B,D).

Looking at the results of heterologous protein expression (Figure 3), the number of planktonic cells expressing eGFP relative to the number of total cells (Figure 3A) was slightly higher for LB medium in most experimental points. The number of eGFP-expressing cells in both media increased until day 6 when on average 27% of the total cells produced eGFP. Regarding the percentage of eGFP-expressing cells in biofilm (Figure 3B), it can be seen that in M9ZB, this parameter remained almost constant throughout the experiment (around 7%, with the maximum value of 13% on day 6), while in LB, a strong increase was observed until day 6 when 66% of the total cells expressed eGFP (p < 0.05). From day 6 onwards, the amount of expressing cells decreased about 20% and was nearly similar until the end of the experiment. Furthermore, in the LB medium, the number of sessile cells expressing eGFP was about 54% higher than in E. coli suspensions, while in M9ZB, the number of expressing cells in the biofilm was similar to its planktonic counterparts (Figure 3A,B).

The specific eGFP values (Figure 3C,D) showed that for both planktonic and biofilm states, the cells grown in LB synthesized more protein than cells grown in M9ZB. On average, the eGFP production in LB medium was eight-fold and three-fold higher for planktonic and biofilm cells, respectively, than in M9ZB. Additionally, in M9ZB, the eGFP levels remained practically constant over time for both planktonic and biofilm states (around 0.31 fg cell⁻¹ and 1.51 fg cell⁻¹, respectively), whereas in LB, the specific eGFP concentration increased until days 5 and 7 for planktonic and biofilm environments, respectively, reaching a maximum of 4.31 fg cell⁻¹ and 15.96 fg cell⁻¹, and then it started to decrease, reaching the steady state.

In general, the PCN (Figure 3E,F) was higher in planktonic cells for both culture media (on average three-fold for M9ZB and two-fold for LB). In planktonic cells grown in LB (Figure 3E), a strong reduction in PCN (of around 89%) was observed until day 5, and from then on, the PCN levels were kept between 4 and 51 plasmids per cell. On the other hand, in planktonic cells grown in M9ZB medium, following an initial decrease by 75% until day 6, the PCN increased until day 8 to the values registered at the start of the experiment, decreasing to about 2 plasmids per cell in the next two days. In biofilm cells (Figure 3F), the PCN reached its maximum at days 5 and 7 in LB and M9ZB, respectively. Then, sessile cells in LB lost about 59% of plasmid content, maintaining values around 29 plasmids per cell until the end of the assay, while in M9ZB, this reduction was about 75% with cells keeping on average 14 plasmids inside them.

4. Discussion

Heterologous protein production using E. coli biofilms has shown to be advantageous for many reasons, including higher plasmid copy number [16] and a higher level of protein production compared to their planktonic counterparts [10,39]. However, the most important goal of heterologous protein production is to achieve high protein production at low cost. Several parameters, such as nutrient medium composition, temperature, hydrodynamic conditions, and surface properties can affect biofilm formation [11,22,40] and consequently influence the final protein yield.

In a previous study [11], our research group compared the effect of two culture media—a diluted medium (DM) and LB—and concluded that biofilms grown in LB produced more eGFP than those grown in DM. In this preliminary work from 2017, primacy was given to culture media typically used to promote biofilm growth such as DM [22,33], while here, there was a concern in testing media recommended for the high expression of recombinant proteins in E. coli such as the M9ZB.

Planktonic results showed a higher number of culturable cells in M9ZB compared to LB medium. This could be related to the presence of glucose in M9ZB medium, which is a compound that is absent in the LB medium used. Previously, Tripathi et al. [41] also found that a culture of E. coli DH5α cells harboring the recombinant plasmid pMAL-c2X achieved higher cell densities in media with a high concentration of glucose, such as Tryptone Yeast (TY) medium, Terrific Broth (TB), and Super Broth (SB).

Biofilm formation was also favored in the M9ZB medium, as indicated by the higher values of both culturable and total cells when compared to LB. Several studies have shown that the nutritional content of the medium, namely carbon and energy substrates, affects biofilm formation [42,43]. Dewanti and Wong [44] observed that using a minimal salts medium (MSM) supplemented with glucose resulted in a stable biofilm with shorter bacterial cells and a thicker extracellular matrix. Buhler et al. [13] have also reported that increasing the glucose concentration of culture medium can promote the growth of E. coli nutrient-depleted biofilms.

Concerning heterologous protein production, a higher number of cells expressing eGFP and higher specific eGFP concentrations were found in LB for both planktonic and sessile cells. Complex media, such as LB, contain yeast extract and tryptone that are known to provide biosynthetic precursors, such as vitamins and amino acids, that support protein expression in E. coli [26,27]. Furthermore, yeast extract can also be responsible for reducing the cellular stress (protease production) during the synthesis of recombinant proteins [45]. Similar to what was observed in the present work with LB, Goyal et al. [26] found that a synthetic medium containing yeast extract and tryptone resulted in higher Streptokinase protein concentration in E. coli, without increasing cell density [26].

The lower eGFP expression in M9ZB medium could be related to the presence of glucose in the medium by repressing the lac promoter [30,46]. In the present work, a pET system was used, where the eGFP gene is transcribed by T7 RNA polymerase, which is regulated by the lacUV5 promoter [4]. It has been shown that glucose reduced the basal expression by reducing cyclic adenosine monophosphate (cAMP) levels, resulting in a low transcription from lacUV5 promoter [46]. Jevševar et al. [47] confirmed that glucose supplementation repressed the expression of recombinant human granulocyte colony-stimulating factor (hG-CSF). On the other hand, it was revealed that the addition of yeast extract to the culture media increased the background expression from the lac promoter [48,49].

Regardless of the culture medium used, there was an increase in the percentage of eGFP-expressing cells in both planktonic and biofilm between days 5 and 7. This phenomenon may be associated with the higher number of culturable cells also detected in the same period. It is expected that metabolically active cells such as the culturable cells are more likely to be eGFP-producing cells, as they have more biosynthetic precursors, energy, and other cellular components available for the synthesis of the foreign protein. In the particular case of biofilms grown in LB medium, the higher levels of expressing cells on days 5–7 may also be related to the simultaneous decrease in biofilm thickness, since the transfer of oxygen for the maturation of the protein under study is facilitated in thinner biofilms [9].

Plasmid stability is an important factor in heterologous protein production [5]. In this work, high specific eGFP concentration was obtained in biofilm cells, although on average higher copies of plasmid were found in planktonic cells. These results suggest that the gene dosage was not responsible for the expression of eGFP. The same phenomenon was observed by Gomes et al. [6], where biofilm cells had more eGFP concentration even though they presented seven-fold lower PCN than planktonic cells. In the present work, the low eGFP expression levels in planktonic cells for both media may be caused by the high metabolic burden linked to plasmid replication. This added metabolic burden usually causes structural and segregational plasmid instability as well as several metabolic changes in the host cell, which may reduce the yield and activity of the heterologous protein [29,50].

Although a small number of plasmids were found in biofilm cells in LB, the biofilm environment seems to be advantageous for plasmid retention, since the PCN decrease was more accentuated in suspended cells than in the biofilms formed in the flow cell system (89% versus 59% reduction). Similar to our results, O’Connell et al. [16] demonstrated that E. coli cultivation as a biofilm has a positive effect on high copy number plasmid maintenance compared to chemostats. The reason for the higher plasmid maintenance in biofilms is that sessile cells tend to grow more slowly than planktonic cells [51], leading to fewer cell divisions and consequently less plasmid partitioning. Furthermore, it was reported that the use of yeast extract and tryptone as nitrogen sources may result in higher plasmid stability [28].

The results obtained in this optimization work are essential to define the operating conditions to be used in further studies focused on the synthesis of other heterologous proteins of different sizes and involved in synthetic pathways for the production of metabolites of high market value.

5. Conclusions

Biofilm cells have shown potential to be used as microbial cell factories, although the influence of environmental factors in the production of high added-value compounds has barely been explored. This study confirmed that heterologous protein production in E. coli biofilms is affected by nutrient conditions. Even though the biofilm development was favored in M9ZB medium when compared with LB, the number of eGFP-expressing cells and the eGFP yield were higher in LB, which is the medium with higher nitrogen content (yeast extract and tryptone). In addition, the PCN in biofilm cells was slightly higher when using LB. On the other hand, the use of M9ZB medium, which has glucose as carbon source, had an inhibitory effect on eGFP expression.

This work is a contribution to the study of operational conditions affecting heterologous protein production in biofilms systems in order to maximize its production.