Introduction

Industrial-scale production of single-chain variable-domain antibody fragment (scFv) expression in Escherichia coli was made possible by genetic engineering. To achieve large-scale microbial biomass production in fed-batch cultures, high cell density cultivation and efficient exogenous protein expression are required (Zarschler et al. 2013). However, cell lysis frequently occurs during the fermentation process because of diverse reactor microenvironments and the expression of heterologous transcripts and proteins. At present, study of the cell autolysis mechanism has mainly focused on the shaking flask level; however, in order to gain more target protein, the culture level of engineering strain expressing foreign protein often needs to enlarge its culture bioreactor level, so it is of great significance to study the effect of the specific expression of foreign proteins at the bioreactor level on cell activity and cell lysis (Sandén et al. 2003). The unpredictable cell lysis phenomenon in the large-scale culture process of E. coli often causes significant difficulty in the downstream purification process and leads to degradation of the target protein, eventually resulting in a decline in the output and quality of the target protein (Andersson et al. 1996).

Overexpressing foreign proteins impose different levels of toxicity to cells, resulting in physical defects and lower cellular activity. Simultaneously, the additional metabolic burden can result in a reduction in the metabolic energy required for effective protein and nucleic acid biosynthesis (Chou 2007). To explore the negative impact of cell lysis on host cells in a biological reactor, it is needed to understand the mechanisms underlying the stress-induced cell lysis caused by the expression of recombinant, exogenous proteins in an E. coli host. Escherichia coli has formed a sophisticated stress response mechanism during its evolution, and it is regulated by complex protein–protein interactions and protein phosphorylation events that are governed by signal transduction pathways. These responses allow the bacteria to adapt to a variety of external and/or internal stimuli, including temperature, pH, osmolarity, nutrient concentrations, the production of secondary metabolites, and the accumulation of misfolded proteins in the periplasmic space (Sikdar et al. 2013). Furthermore, a recent study found that cell death and lysis in E. coli are regulated by programmed cell death, which is similar to apoptosis in eukaryotic cells. Based on a shaking flask culture, Murata et al. (2012) identified a σE-dependent cell lysis process, which is considered to be a survival strategy when cells suffer from adverse environmental stresses. When cells suffer from stresses during the early stationary phase of growth, most of them enter the viable but nonculturable (VBNC) state, and cell lysis is induced to remove these damaged cells, thereby maximizing the use of nutrients to maintain cell viability for prolonged periods.

However, current progress on understanding stress and cell lysis mechanisms is limited to the shake flask level, so a comprehensive understanding of the impact of exogenous protein expression on cell lysis in bioreactors is needed. To fill this knowledge gap, we have investigated the correlation between exogenous protein expression and cell lysis in a bioreactor, as well as the role of σE-directed stress response pathways on cell lysis, using an scFv-expressing E. coli W3110 system.

Methods

Strains, plasmids, and culture conditions

The strains used in this study are the E. coli W3110 wild-type strain and a W3110 strain carrying the vector pBZ001 (W3110/pBZ001). This vector contains the scFv gene (939 bp), the expression of which is driven by the T7A3 promoter, and an operator system consisting of two perfect palindromic sequences. The signal sequence used for the periplasmic production of scFv was derived from OmpA, and this sequence is cleaved when the recombinant protein is translocated across the cytoplasmic membrane, thereby allowing the mature protein to fold into a functional scFv. The strains were cultured overnight at 37 °C on Luria–Bertani (LB) plates. A single colony was picked and inoculated into 100 mL of Terrific Broth/Super Broth (TB/SB) (Table 1) seed medium in a 1-L Erlenmeyer flask and cultured at 37 °C with shaking at 230 rpm on an oscillatory shaker. When the optical density at 600 nm (OD600) reached 4, the seed culture was inoculated into 2 L of TB/SB fermentation medium in an ez-Control 5-L fermenter (Applikon, Delft, The Netherlands). The fermenter temperature was maintained at 30 °C ± 0.1 °C, and the pH was maintained at 7.0 ± 0.05 by the addition of 17 % phosphoric acid or 25 % aqueous ammonia. To ensure that the level of dissolved oxygen (DO) achieved 30 % oxygen saturation, the air flow rate was maintained at 2 volumes of air per volume of fermentation broth per minute (vvm); when necessary, 1 vvm of pure oxygen was provided. Before DO spiking, feeding was initiated using a precalibrated peristaltic pump at different flow rates (3.5 or 6 mL h−1l−1). Tetracycline was sterilized by filtration through a 0.2-μm cellulose acetate membrane filter and added to a final concentration of 15 μg mL−1. To induce the expression of exogenous scFv in E. coli W3110/pBZ001, a final isopropyl β-D-1-thiogalactopyranoside (IPTG) concentration of 100 μM was achieved by injecting IPTG into the fermenter 1 h after feeding was initiated, and the fermentation process was continued for approximately 38 h after induction.

Table 1 The medium used in the study
Full size table

Determination of VBNC cell number

Sample treatment

Culture broth samples (1 mL) were collected aseptically at different fermentation stages based on the OD600 and culture time, and they were serially diluted 10-fold to an appropriate concentration using 0.01 M phosphate-buffered saline (PBS) for counting. One hundred microliters of the diluent was used for colony-forming unit (CFU) determinations by plating it onto LB agar plates and counting the number of colonies after 24 h of incubation at 37 °C (viable and culturable cell number). Meanwhile, an appropriate dilution (100 μL) was centrifuged at 9000 × g at 4 °C for 3 min (FRESCOL-7, Thermo Fisher Scientific). The supernatant was discarded, and the pellet was washed twice with 0.01 M PBS.

CTC staining

5-Cyano-2,3-ditolyl tetrazolium chloride (CTC) fluorescent staining counting was conducted as described previously (Abe et al. 2007), with slight modifications. Viable bacteria cells transport electrons via respiration, thereby reducing CTC to CTC formazan (CTF), which results in the formation of a red fluorescent precipitate in the cell membrane. The washed pellets were resuspended in approximately 100 μL of a 3 mM CTC solution (dissolved in 0.01 M PBS) and incubated for 30 min at 25 °C with shaking at 200 rpm in the dark.

Fluorescence microscopy count

As Hobbie et al. (1977) described, stained samples were subsequently harvested by filtration through a black polycarbonate membrane filter (0.02 μm pore size, 13 mm diameter; Sartorius AG, Göttingen, Germany). The cell number was determined using a BX53F fluorescence microscope (Olympus, Tokyo, Japan) with a Fluotar objective (U Plan Semi-Apochromat objective 100×/1.3, oil) and a wide-field eyepiece (10×, focusable). An Olympus filter set (BP425-445 exc./FF593/BA628-640 em.) was used to detect CTF. At least 30 to 50 cells in ten randomly chosen areas on a filter were counted, and the average value and standard deviation were calculated. The cell number was calculated as follows:

$$ E=\frac{X\times {S}_1}{S_2}\times \frac{1}{V} $$

where E is the cell number in the sample, X is the average cell number in the microscopic fields, S 1 and S 2 are the areas of the membrane and microscopic field, respectively, and V is the sample volume.

RNA extraction, cDNA synthesis, and RT-PCR

Bacteria (5 × 106–107 cells) in fermented broth were immediately centrifuged at 8000 × g (4 °C, 2 min), the supernatant was discarded, and the pellet was stored in liquid nitrogen for RNA extraction. Total RNA was isolated from E. coli W3110 cells using the RNAiso Plus Kit (TaKaRa, Shiga, Japan), according to the manufacturer’s instructions. The amount and purity of the RNA were reflected by the absorbance at 260 and 280 nm (A260 and A280, respectively), and the RNA purity was assessed by 2 % agarose gel electrophoresis. The primers used in our study were designed with Primer 5.0 software, and the sequences of the primers are given in Table 2. After removing the genomic DNA, total RNA was reverse-transcribed into cDNA, as described in the instructions of the PrimeScript™ RT Reagent Kit with gDNA Eraser (TaKaRa). Real-time polymerase chain reaction (RT-PCR) was conducted using an iCycler CFX96 quantitative PCR system (Bio-Rad, Hercules, CA, USA), SYBR® Premix Ex Taq™ II (Tli RNaseH Plus) fluorescence quantitative reagents (TaKaRa), and individual eight-tube PCR strips (Bio-Rad). Each reaction mixture (10 μL) consisted of 5 μL SYBR® Premix Ex Taq™ II (2×), 0.4 μL of forward primer (10 μM), 0.4 μL of reverse primer (10 μM), and appropriate volumes of cDNA template and sterile double-distilled H2O. The reaction mixture was incubated at 95 °C for 30 s for predenaturation, followed by 40 cycles of amplification, which consisted of a denaturation step (95 °C for 5 s), an annealing step (60 °C for 30 s), and a denaturation step (95 °C for 30 s). Gene expression was calculated relative to that of 16S rRNA expression using the 2−ΔΔCT method. Three independent RT-PCR replicates were performed. Relative gene expression was shown as the ratio of gene expression in E. coli W3110/pBZ001 and wild-type E. coli W3110 at the same culture time.

Table 2 The sequences of the primers used in real-time polymerase chain reaction (RT-PCR)
Full size table

The degree of cell lysis

We defined the degree of cell lysis as the ratio between the reduction of maximum biomass at time t and the maximum biomass, which occurred when E. coli W3110/pBZ001 and wild-type E. coli W3110 entered stationary phase after 15 h of incubation:

$$ \mathrm{Cell}\ \mathrm{lysis}\ \mathrm{degree}=\frac{{\mathrm{OD}}_{600}\left(15\ \mathrm{h}\right)\hbox{-} {\mathrm{OD}}_{600}\left(\mathrm{t}\right)}{{\mathrm{OD}}_{600}\left(15\ \mathrm{h}\right)} $$

Secreted and intracellular scFv levels as determined by an enzyme-linked immunosorbent assay

Ninety-six-well microtiter plates were coated with 100 μL of human serum albumin (HAS) (25 μg mL−1) in 50 mM carbonate buffer at pH 9.6 and incubated at 4 °C overnight. The plates were emptied and washed five times with PBS-Tween-20 (PBS-T) washing solution (0.15 M sodium phosphate buffer, pH 7.4, supplemented with 0.05 % Tween-20). One hundred and fifty microliters of 20 % (w/v) bovine serum albumen solution in PBS-T buffer was added and incubated for 1 h at 37 °C to block the unoccupied sites on the surfaces of the polystyrene wells. The plates were emptied and washed as described above. Antibody samples (the fermentation supernatant) and controls (scFv standard after protein A purification) were diluted in phosphate buffer, pH 7.4, and 50 μL of the antibody and control diluents were preincubated for 1 h at 37 °C. The plates were washed as before. Goat anti-mouse horseradish peroxidase-conjugated secondary antibody, diluted 1:1000, was added (50 μL per well) to the plates. The plates were incubated for 1 h at 37 °C, emptied, and washed. One hundred microliters of substrate and chromogenic reagent were mixed for 15 min at 37 °C in the dark, and 100 μL of 2 M H2SO4 was added to stop the reaction. The absorbance at 450 nm was read by a microplate reader.

Results and discussion

scFv expression caused cell lysis in E. coli W3110/pBZ001

The prebatch cultivation results showed that DO levels always spiked after 11.5 h ±1 h of inoculation, which indicates that the glycerol in the culture was exhausted (data not shown). Therefore, feeding was started at 9 h. One hour after feeding was initiated, scFv expression was induced by 1 mM IPTG when the OD600 was greater than 40. Although E. coli W3110/pBZ001 grew slightly slower than the wild-type W3110 strain, they roughly entered stationary phase after 15 h of induction. Obvious cell lysis of E. coli W3110/pBZ001 started at 18 h, whereas lysis of the wild-type strain began at 12 h. At 24 h, the degree of cell lysis of E. coli W3110/pBZ001 reached 18.3 %, which was significantly higher than that of the wild-type strain (Fig. 1).

Fig. 1
figure 1

Comparison of cell lysis and single-chain variable-domain antibody fragment (scFv) expression between Escherichia coli W3110/pBZ001 and the wild-type strain

Full size image

The expression of scFv in the pBZ001vector was regulated by the T7A3 promoter. This expression system was characterized by weak basal activity, which was the reason why scFv was barely detected in cell supernatants and the fermentation broth prior to induction. Within 3.5 h after induction, scFv production was rapidly induced at the rate of 0.906 mg g−1 DCW h−1. Most of the products are intracellular accumulated and less than one-fifth of the total amount of scFv was secreted into the extracellular space; no scFv expression was detected in the wild-type strain (Fig. 1).

As shown in Fig. 1, introducing the plasmid containing the exogenous scFv gene slightly inhibited cell growth. Moreover, it was obvious that the high-level expression of scFv after induction led to early cell lysis. These problems in an E. coli host can be a major bottleneck that hampers heterologous protein production in bioreactors (Schlegel et al. 2013). However, the cell lysis phenomenon is not obvious in a shake flask. The stability culture stage can be maintained for a long time(>72 h) and the cell density decreases little in the decline phase (data not shown).

Cell lysis mainly occurs in VBNC cells in a bioreactor

The VBNC state of bacteria was proposed by Xu et al. (1982), and it represents cells that have lost the ability to form colonies on a medium when they suffer from stresses; however, their cell structure (e.g., cell membrane integrity), metabolic activity (such as respiration, biochemical substance absorption, etc.), DNA synthesis, and the expression of specific mRNA genes that are characteristic of living cells remain intact. Once a suitable environment is restored, VBNC cells can be transformed into the cultivable state.

As shown in Fig. 2, after scFv overexpression in W3110/pBZ001, the number of CFU immediately decreased exponentially, while a slight reduction in living cells occurred from 10 to 18 h, indicating the entrance of the vast majority of cells into the VBNC state. A significant reduction in living cells and a slight CFU decrease, which accompanied the cell lysis, was detected after 18 h, indicating that cell lysis mainly occurs in cells entering the VBNC state. A similar conclusion was drawn for the wild-type strain. Most VBNC cells accumulated in the wild-type strain after 15 h and substantially lysed at 24 h. This is consistent with the assumption proposed by Noor et al. (2009) that VBNC cells are formed during the stationary phase and, in turn, are lysed by a σE-dependent process.

Fig. 2
figure 2

The relationship between the occurrence of cell lysis and the reduction of the viable but nonculturable (VBNC) number in E. coli W3110/pBZ001 and the wild-type strain. The VBNC number is equal to the difference between the viable cell number and the colony-forming unit (CFU) number

Full size image

Lin (2000) expressed yeast α-glucosidase in E. coli as a model system to obtain a more comprehensive understanding of the cell physiological response to the overexpression of heterologous genes during glucose-limited, fed-batch cultures. They found that the synthesis of recombinant proteins inhibited the synthesis of cellular housekeeping proteins, as there was a competition for RNA polymerase, ribosomes, and translation factors. As a result, there was a decrease in different cellular metabolic reactions, such as replication, transcription, translation, glucose uptake, and respiration, and this competitive effect led to the inhibition of cell growth. Thus, the host cells became VBNC cells, which maintain metabolic activity, but irrevocably lose their ability to grow (Lin 2000). Therefore, we assume that the lysis of E. coli W3110/pBZ001 cells observed in our study occurred mainly in VBNC cells.

The stress response and σE-dependent cell lysis in W3110/pBZ001

The heat shock response

Figure 3 compares the expression of heat shock response genes between W3110/pBZ001 and wild-type strain in an identical culture system. From 3 to 15 h after induction, the expression of rpoH was notably higher in W3110/pBZ001 than in the wild-type strain, reaching a maximum 12 h after induction and undergoing a dramatic decline after 15 h when cell lysis began. The expression of dnaK, dnaJ, groES, and groEL, which are controlled by rpoH, showed the same tendency. The expression of all genes in the two strains tended to converge to the same level at the end of the fermentation (48 h), indicating that their metabolic activity stagnated after cell lysis.

Fig. 3
figure 3

The relative expression of heat stress response genes in E. coli. The relative expression is the ratio of gene expression in E. coli W3110/pBZ001 and wild-type E. coli at the same time point

Full size image

The heat shock response in E. coli is self-regulated under adverse conditions to enable survival in various microenvironments (Arsène et al. 2000). The rpoH gene, encoding the σ32 regulator, is responsible for degrading misfolded proteins, preventing protein misfolding, or refolding misfolded proteins by regulating the DnaK/J–GrpE and GroEL–GroES chaperone pairs (Hoffmann and Rinas 2004; Kumar and Sourjik 2012). By binding to hydrophobic regions of unfolded polypeptides, the role of heat shock proteins, such as DnaK/DnaJ in prokaryotes, is to maintain protein solubility and prevent aggregation, while GrpE functions as a trigger to separate DnaK/DnaJ/polypeptide complexes via nucleotide exchange (Hartl et al. 2011; Noor 2015). After IPTG induction in a fermentation culture, the expression of rpoH, dnaK, dnaJ, grpE, groEL, and groES in E. coli W3110/pBZ001 significantly increased compared with that in the wild-type strain at each time point, indicating that the heat shock response of W3110/pBZ001 cells was intensely triggered by scFv overexpression.

The acid stress response

Figure 4 compares the dynamics of the expression of H-NS, rpoS, gadE, and gadX in the acid stress response pathway. The mRNA level of H-NS, the senior negative regulator in the acid resistance (AR) regulating system (Krin et al. 2010), substantially decreased 2 h after the induction of scFv expression in E. coli W3110/pBZ001, while that of rpoS and gadX, the general regulators of the stress response (Talukder et al. 1996) and the retrograde transport gene in the AR system, increased by approximately 6- and 8-fold, respectively, compared with the wild-type strain. It is noteworthy that the transcriptional level of gadE, encoding GadE, the central activation element of the glutamate-dependent AR system, increased by more than 200-fold dramatically before the induction and remained at a high level 2 h later and then declined 5 h eventually.

Fig. 4
figure 4

The relative expression of acid stress response genes in E. coli. The relative expression is the ratio of gene expression in E. coli W3110/pBZ001 and wild-type E. coli at the same time point

Full size image

H-NS is believed to promote the degradation of rpoS mRNA and suppress the expressions of certain regulators, such as gadX, adiY, and cadC in amino acid-dependent AR systems (Krin et al. 2010; Talukder et al. 1996). During the phase of rapid scFv synthesis (10–13.5 h of induction), the expression of H-NS decreased and removed the inhibition of rpoS expression, directly or indirectly, and promoted the expression of the gadX and gadE genes in the glutamate-dependent AR system to cope with changes in the intracellular pH caused by scFv overexpression.

The oxygen stress response

Figure 5 compares the dynamics of two major oxygen stress response genes between the W3110/pBZ001 and wild-type strains. sodA and katE, encoding superoxide dismutase and catalase HPII, respectively, can cope with protein carbonylation damage caused by sudden oxygen stresses (Dukan and Nyström 1999). Two hours after the induction of scFv expression in the W3110/pBZ001 strain, the mRNA levels of sodA and katE increased by approximately 10- and 6-fold, respectively. However, the expression of sodA in the W3110/pBZ001 strain declined 3 h later to an extremely low level compared with that of the wild-type strain, and reached its minimum level at 24 h, while the expression of katE in the W3110/pBZ001 strain also decreased by approximately 3-fold at 24 h (Fig. 5).

Fig. 5
figure 5

The relative expression of oxygen stress response genes in E. coli. The relative expression is the ratio of gene expression in E. coli W3110/pBZ001 and wild-type E. coli at the same time point

Full size image

The σE-dependent cell lysis and repair pathway

Figure 6 shows the expression patterns of the rpoE, ompA, ompC, ompW, and ompX genes during σE-dependent cell lysis, which was discovered by Murata et al. (2012), in the W3110/pBZ001 and wild-type strains. Prior to IPTG induction, the mRNA level of rpoE, which encodes the σE transcription factor, was lower in the W3110/pBZ001 strain than in the wild-type strain, and it increased by approximately 8-fold 2 h after induction. The enhanced mRNA rpoE levels eventually decreased to a significantly low level 14 h after induction. The mRNAs of ompA and ompC, which encode the two main outer membrane proteins in E. coli, remained at high levels in the W3110/pBZ001 strain up to 12 h of induction, and then decreased to extremely low levels. Prior to induction, the mRNA levels of ompW and ompX in the W3110/pBZ001 strain were approximately 18-fold lower after 9 h of incubation and 2.5-fold lower after 7.5 h of incubation, respectively. However, the expression of ompW and ompX significantly increased after induction, but decreased after 5 h, compared with that in the wild-type strain.

Fig. 6
figure 6

The relative expression of cell lysis genes governed by σE in E. coli. The relative expression is the ratio of gene expression in E. coli W3110/pBZ001 and wild-type E. coli at the same time point

Full size image

The σE transcription factor was originally defined in a study of the heat shock factor σH and σE regulates genes, which are required to repair abnormally folded periplasmic and membrane proteins that occur in response to thermal stress or other external stimuli (Ades 2008). Murata et al. (2012) presented a complete σE-dependent cell lysis pathway. Because of the intolerable accumulations of unfolded proteins in the periplasmic space or the cell membrane induced by sudden stresses, the σE protein in the cytoplasm is activated to upregulate the transcription of the small RNA MicA and rybB, which reduce the expression of genes encoding outer membrane proteins such as ompA, ompC, and ompW, resulting in defects in the outer membrane and ultimately leading to cell lysis. This σE-dependent cell lysis proceeds in a cascade following rpoE overexpression during the logarithmic and stationary phases in shake flask cultures. However, the sustained high level of expression of rpoE does not occur during dramatic cell lysis in bioreactor cultures, indicating that there are alternative ways of governing cell lysis in E. coli. OmpA is an important outer membrane structural protein that maintains cell morphology and transfers hydrophilic compounds across the cell membrane (Danoff and Fleming 2011), and OmpC is the main cation-selective porin that enables the entry of small molecules into the bacteria (Baslé et al. 2006; Joseph Sahaya Rajan et al. 2015). In Gram-negative bacteria, OmpW encodes a small β-sheet membrane protein that allows the bacteria to adapt to environmental changes (Brambilla et al. 2014), while the mechanism that controls OmpX expression is not clear. The levels of the mRNAs that encode these four proteins increased with the enhanced expression of rpoE during the early stage of induction (about 10 to 15 h) in the fermentation culture of W3110/pBZ001. When cells are exposed to extracytoplasmic stresses caused by the rapid accumulation of exogenous proteins, σE induces its own expression in response to RseB or RseA degradation by the DegS and YaeL proteinases, which leads to the repair of the outer membrane (Missiakas et al. 1997). Five hours after induction, the expression of genes encoding outer membrane proteins in E. coli W3110/pBZ001 was significantly lower than that in the wild-type strain. This phenomenon may be caused by the decreased synthetic capacity of the outer membrane proteins and the negative regulatory effect of cell lysis on the expression of these genes (Nitta et al. 2000), which, in turn, led to earlier cell lysis.

Numerous advances have been achieved in the engineering of σ transcriptional regulators, which have provided us with an effective means by which to tune the regulatory networks of bacteria, improve bacterial stress resistance, and increase the yields of desired products (Tripathi et al. 2014). Recently, Dragosits et al. (2012) constructed a novel, self-regulatory protein production system, which controls recombinant protein production through a stress-induced, negative feedback mechanism. To screen for molecular targets that increase cell resistance to stresses and to control cell lysis, further exploration of the mechanisms underlying the σ-dependent stress response and cell lysis is needed. Such knowledge could help to maximize the yields of heterologous proteins by balancing exogenous protein expression and cell growth.

Conclusion

The present study investigated how the induction of single-chain variable-domain antibody fragment (scFv) expression in Escherichia coli W3110/pBZ001 influenced the viability of host cells and their cell lysis pattern in a 5-L bioreactor. Furthermore, the transcriptional levels of genes related to the general stress response and cell lysis were compared between the W3110/pBZ001 and wild-type strains throughout identical fermentation processes. The scFv protein was rapidly synthesized from 10 to 13.5 h of induction, and most of the protein accumulated in the intracellular space until 24 h (Fig. 1). The viable but nonculturable (VBNC) cell number in W3110/pBZ001 began to increase rapidly 10 h after induction, and reached its maximum of approximately 7.34 × 1010 colony-forming units (CFU) mL−1 at 13.5 h after induction, while the rapid increase in the VBNC cell number in the wild-type strain was postponed for 6 h (Fig. 2). Because of scFv overexpression, multiple stress response pathways in E. coli W3110/pBZ001 were activated to accommodate the microenvironmental changes, as evidenced by the results shown in Figs. 3, 4, 5, and 6. VBNC cells lysed after 18 h of scFv induction in the W3110/pBZ001 strain, which was followed by a loss of cell viability, as evidenced by the substantial decrease in the expression of stress response genes (from 15 to 24 h of induction). σE participated in cell membrane repair, but no overexpression was detected during cell lysis (from 18 to 24 h of induction in the W3110/pBZ001 strain).

According to the findings of this study, we assume that recombinant E. coli cell lysis occurs mainly in VBNC cells. Comprehensive investigations of VBNC cell formation and changes in bioreactor conditions are essential for the industrial-scale production of scFv (Chou 2007). Further study is required in order to assess the effects of external stresses, such as process design parameters, on recombinant cells, as well as to understand the variations in the cellular profiles of VBNC cells (Lin 2000), which may contribute to the definition of the physiological states of the cells, i.e., survival or cell lysis.