Reversible bacterial immobilization based on the salt-dependent adhesion of the bacterionanofiber protein AtaA

Background Immobilization of microbial cells is an important strategy for the efficient use of whole-cell catalysts because it simplifies product separation, enables the cell concentration to be increased, stabilizes enzymatic activity, and permits repeated or continuous biocatalyst use. However, conventional immobilization methods have practical limitations, such as limited mass transfer in the inner part of a gel, gel fragility, cell leakage from the support matrix, and adverse effects on cell viability and catalytic activity. We previously showed a new method for bacterial cell immobilization using AtaA, a member of the trimeric autotransporter adhesin family found in Acinetobacter sp. Tol 5. This approach is expected to solve the drawbacks of conventional immobilization methods. However, similar to all other immobilization methods, the use of support materials increases the cost of bioprocesses and subsequent waste materials. Results We found that the stickiness of the AtaA molecule isolated from Tol 5 cells is drastically diminished at ionic strengths lower than 10 mM and that it cannot adhere in deionized water, which also inhibits cell adhesion mediated by AtaA. Cells immobilized on well plates and polyurethane foam in a salt solution were detached in deionized water by rinsing and shaking, respectively. The detached cells regained their adhesiveness in a salt solution and could rapidly be re-immobilized. The cells expressing the ataA gene maintained their adhesiveness throughout four repeated immobilization and detachment cycles and could be repeatedly immobilized to polyurethane foam by a 10-min shake in a flask. We also demonstrated that both bacterial cells and a support used in a reaction could be reused for a different type of reaction after detachment of the initially immobilized cells from the support and a subsequent immobilization step. Conclusions We invented a unique reversible immobilization method based on the salt-dependent adhesion of the AtaA molecule that allows us to reuse bacterial cells and supports by a simple manipulation involving a deionized water wash. This mitigates problems caused by the use of support materials and greatly helps to enhance the efficiency and productivity of microbial production processes. Electronic supplementary material The online version of this article (doi:10.1186/s12934-017-0740-7) contains supplementary material, which is available to authorized users.

downstream process including product separation [6]. Recently, a systems metabolic engineering approach targeting an upstream process received considerable attention by aiming to develop a novel biosynthetic pathway producing high-value products and/or improve their productivity in microbial cells [6][7][8][9][10][11]. As for the downstream process, cell immobilization is important because it simplifies product separation, enables the cell concentration to be increased, stabilizes the enzymatic activity, and permits repetitive or continuous use of precious and expensive biocatalysts [12][13][14][15]. Conventional methods for cell immobilization are gel entrapment, covalent bonding to solid surfaces, cross-linkage, and physical adsorption [16,17]. These methods, however, have practical limitations, such as limited mass transfer in the inner part of a gel [18,19], gel fragility, cell leakage from the support matrix, and adverse effects on cell viability and catalytic activity [12].
We previously invented a method for bacterial cell immobilization using the adhesive protein AtaA found in Acinetobacter sp. Tol 5 [20][21][22], which belongs to the trimeric autotransporter adhesin (TAA) family [23]. Although AtaA shares a fibrous architecture consisting of an N-terminus-passenger domain (PSD) containing head and stalk domains-transmembrane anchor (TM)-C-terminus with TAA family members [24], which usually bind to target biotic surfaces, AtaA uniquely confers nonspecific high adhesiveness to both abiotic and biotic surfaces on bacterial cells transformed with its gene. Large amounts of growing, resting, even lyophilized transformant cells can be quickly and firmly immobilized onto any material surfaces selected according to the application [25]. Cells immobilized directly on surfaces through AtaA are not embedded in extracellular polymeric substances with mass transfer limitations, show enhanced tolerance [22], increase chemical reaction rates, and can be repeatedly used in reactions without inactivation [25]. However, similar to all other immobilization methods, the use of support materials increases the cost of bioprocesses and subsequent waste materials. These might be inevitable problems as long as support materials are used in the immobilization process. A way to minimize these drawbacks should be developed so as to, for example, reduce the amount of support materials, use inexpensive materials or waste materials, and reuse support materials.
AtaA is a homotrimer of polypeptides comprising 3630 amino acids. In a previous study, we developed a method to isolate its PSD, which is secreted to the bacterial cell surface through the TM and is responsible for biological functions, by genetically introducing a recognition site for human rhinovirus 3C (HRV 3C) protease [26]. Specific cleavage by the protease reaps AtaA PSD nanofibers 225 nm in length from the cell surface. This enables biochemical and biophysical analyses of the purified huge AtaA PSD in the native molecular state. Here, we demonstrate a new phenomenon: AtaA PSD cannot adhere to surfaces in deionized water (dH 2 O). Based on this molecular property of AtaA, we developed a unique method for the reversible immobilization of bacterial cells, which can solve the problems caused by the use of support materials.

Effect of ionic strength on the adhesive property of the AtaA molecule
To investigate the adhesive property of the AtaA molecule, AtaA PSD was isolated by the enzymatic reaping method from a Tol 5 derivative strain, 4140, transformed with p3CAtaA [26,27]. KCl solutions (50 µL) of various concentrations containing 5 μg/mL of the purified AtaA PSD were incubated in 96-well polystyrene (PS) and glass plates at 28 °C for 2 h. The protein solution was removed from each well using a micropipette and the well was rinsed three times with 200 μL of phosphate-buffered saline containing 0.05% Tween-20 (PBS-T). The AtaA PSD adsorbed to the well plates was assessed by ELISA. Interestingly, we found that the amounts of AtaA PSD molecules adsorbed onto surfaces of hydrophobic PS and hydrophilic glass dropped sharply at ionic strengths lower than 10 mM, with the molecules hardly adhering in dH 2 O, despite their high adhesiveness at higher ionic strengths (Fig. 1a). AtaA PSD cannot be considered to be denatured in dH 2 O because AtaA PSD has high structural stability [26,28]. Indeed, AtaA PSD molecules isolated from cells by the enzymatic reaping method mentioned above were dissolved in dH 2 O, and subsequently KCl solution was added to the adherence assay to attain the final ionic strengths intended. Furthermore, by using a quartz crystal microbalance (QCM), which enables the quantification of molecules adhered to its quartz crystal sensor chip as a frequency shift, we confirmed that the adhesiveness of AtaA PSD can be recovered in a salt solution by adding KCl salt to fresh water. AtaA PSD molecules did not adhere to a gold-coated sensor chip of QCM in dH 2 O, but started adhering to the chip immediately after KCl solution was added (Fig. 1b). Evidently, AtaA PSD is not denatured in dH 2 O and recovers its adhesiveness in a salt solution.

Effect of the ionic strength on bacterial cell adhesion mediated by AtaA
The identification of the ionic strength-dependent stickiness of the AtaA PSD prompted us to examine whether or not bacterial cell adhesion mediated by AtaA also depends on ionic strength. Tol 5 cells were grown, harvested, washed with dH 2 O, suspended in dH 2 O and KCl solutions of various concentrations at an OD 660 of 0.5, and placed into 96-well plates. After a 2-h incubation at 28 °C without shaking, the cell suspension was removed from each well by a micropipette and the well was rinsed three times with 200 μL of dH 2 O or KCl solution of each concentration using a micropipette. The cells immobilized onto the well surfaces were quantified by crystal violet staining. At ionic strengths higher than 20 mM, a large amount of Tol 5 cells was immobilized onto PS and glass surfaces, although the amount gradually increased as the ionic strength increased (Fig. 2a). However, Tol 5 cell adhesion dropped at ionic strengths lower than 5 mM for both PS and glass surfaces, and Tol 5 cells were unable to adhere to either surface in dH 2 O. To confirm that such ionic strength-dependent adhesion of bacterial cells can be decisively attributed to the adhesive properties of the AtaA molecule, the cell adhesion of Acinetobacter baylyi ADP1 and its transformant with ataA, ADP1 (pAtaA), to PS and glass surfaces was examined at various ionic strengths by the same procedure used for Tol 5 cells. ADP1 cells expressing AtaA showed high adhesiveness to both PS and glass surfaces at ionic strengths higher than 20 mM with a gradual increase in adhesion with ionic strength, whereas wild-type ADP1 was hardly immobilized at any ionic strength (Fig. 2b). However, at an ionic strength lower than 5 mM, even ADP1 cells expressing AtaA showed the same diminished adhesion as Tol 5 cells, adhering to neither PS surface nor glass surface in dH 2 O. Therefore, the adhesion profiles of Tol 5 cells and ADP1 (pAtaA) cells at various ionic strengths directly reflect the properties of the AtaA molecule.

Cell detachment and reversible immobilization using AtaA
Our finding that the cell adhesion mediated by AtaA is inhibited by dH 2 O prompted us to examine the ability of a simple dH 2 O wash to detach bacterial cells already immobilized on material surfaces. After immobilization of ADP1 (pAtaA) cells onto the well surfaces in 100 mM KCl solution as described above, the wells were rinsed with 200 μL of dH 2 O or 100 mM KCl solution using a micropipette. This washing step was repeated three times. Thereafter, the cells still immobilized on the well surfaces were quantified by crystal violet staining. Most of the cells washed with dH 2 O were detached from both PS and glass surfaces, whereas the cells washed with 100 mM KCl solution were retained on the surfaces (Fig. 3).
We also confirmed the ability of dH 2 O to detach ADP1 (pAtaA) cells previously immobilized on a polyurethane foam support, which is often used in bioprocesses as a support. The cells were immobilized onto 1-cm 3 pieces of polyurethane foam support. A piece of the support with the immobilized cells was transferred into fresh 100 mM KCl solution, gently rinsed, picked up with tweezers, and shaken in dH 2 O or 100 mM KCl solution for video recording. When shaken in dH 2 O, the immobilized cells immediately began to detach and an increase in the turbidity of the surrounding H 2 O solution was observed. An additional movie file shows this in more detail (see Additional file 1). In contrast, the immobilized cells were not detached at all by being shaken in 100 mM KCl solution (see Additional file 2).
Next, we tried to repeat the immobilization and detachment of bacterial cells expressing AtaA. ADP1 (pAtaA) cells were repeatedly subjected to the immobilization/detachment process; cells suspended in 100 mM KCl solution were immobilized onto the PS well surface by a 2-h incubation (immobilization process) and subsequently detached by the dH 2 O wash using the same procedure described above (detachment process). The detached cells were collected by centrifugation, resuspended in 100 mM KCl solution at an OD 660 of 0.5, and placed into the new well for the next immobilization cycle. This immobilization/detachment process was repeated four times. As shown in Fig. 4a, the detached cells showed the same adhesion ability as the fresh cells used for the first adherence assay, and the cell adhesiveness did not decrease throughout four immobilization/ detachment cycles. The cells finally detached were subjected to flow cytometry to quantify the amount of AtaA displayed on the ADP1 (pAtaA) cell surface. This revealed that the amount of AtaA molecules on the cell surface did not decrease even after the fourth detachment compared with that on fresh cells before the first immobilization ( Fig. 4b), suggesting that AtaA was not impaired throughout the repeated detachment manipulations by washing with dH 2 O. The ADP1 (pAtaA) cells immobilized in each immobilization/detachment cycle were subjected to an esterase activity assay involving the addition of a reaction buffer containing a substrate directly to the well. The cellbound esterase activity of the immobilized cells on the PS surface was maintained at the same level throughout the four cycles (Fig. 4c), suggesting that the repeated immobilization/detachment cycle also did not deteriorate the integrity of the surface of ADP1 (pAtaA) cells.
We also examined whether or not ADP1 (pAtaA) cells can be reversibly immobilized onto polyurethane foam support. Six pieces of the polyurethane foam (a 1 cm cube) were placed into a 30-mL cell suspension of ADP1 (pAtaA) at an OD 660 of 2.0 in a 100-mL Erlenmeyer flask and shaken at 115 rpm for immobilization of the bacterial cells. The percentage of immobilized cells over time is shown in Fig. 4d ("1st"). More than 90% of the cells were immobilized within 10 min (immobilization process). Subsequently, the supernatant was discarded by decantation and 30 mL of dH 2 O was poured into the flask, which was then shaken at 115 rpm for 5 min. This washing step was repeated three times (detachment process). The detached cells from each washing step were collected by centrifugation and resuspended in 100 mM KCl solution at an OD 660 of 2.0. Six fresh pieces of the polyurethane foam were placed into the cell suspension for the next immobilization cycle. This immobilization/detachment cycle was also repeated four times. The time profiles of the immobilization were similar throughout the four cycles; more than 90% of ADP1 (pAtaA) cells detached from the polyurethane were re-immobilized onto the fresh polyurethane foam within 10 min (Fig. 4d).
Thus, we have succeeded in developing a novel method for the reversible immobilization of bacterial cells, which enables the reuse of cells without impairment, by means of AtaA expression and a simple manipulation involving a dH 2 O wash. Other conventional immobilization methods are unsuitable for the development of a reversible process without support destruction or cell inactivation.

Reuse of bacterial cells in a different type of reaction
To show the merit of our reversible immobilization method, we attempted to demonstrate the reusability of bacterial cells for different types of reactions, an ester hydrolysis and toluene degradation, using the schema shown in Fig. 5a. At first, Tol 5 cells were immobilized onto three pieces of the polyurethane foam support in a 100-mL Erlenmeyer flask containing the cells suspended in 30 mL of 100 mM KCl solution (OD 660 = 1.0) by shaking at 115 rpm at 28 °C for 30 min. The cells loosely attaching to the support were removed by dipping them into 100 mM KCl solution and gentle squeezing. A piece of the support with the immobilized Tol 5 cells was placed into esterase reaction buffer in a test tube. After a 10-min incubation of the test tube at 28 °C, the reaction buffer turned from colorless to yellow due to the 4-nitrophenol produced by esterase on the immobilized Tol 5 cells (Fig. 5b). Next, three pieces of the support used in the esterase reaction were collected and washed in 100 mL of dH 2 O in a 500-mL Erlenmeyer flask by shaking at 115 rpm for 5 min. This washing step was repeated three times. The detached cells from each washing step were collected by centrifugation, resuspended in 30 mL For the induction of toluene-degrading gene expression, the steel wool support with the re-immobilized Tol 5 cells was picked up, touched with paper towel to remove extra water, transferred into a 25-mL vial, and incubated at 28 °C for 1 day under a toluene atmosphere. After this induction step, the immobilized cells on the steel wool support were subjected to a toluene-degradation reaction in a gas phase (Fig. 5c). The reaction was started by injecting 1 μL of toluene into the vial, and thereafter the toluene concertation of the gas in the vial was quantified by gas chromatography-mass spectrometry (GC/MS) and its time-dependent decrease was plotted (Fig. 5d). The cells immobilized on the steel wool support linearly degraded toluene for 5 h and thereafter the degradation rate lowered following the first-order reaction kinetics that depends on the toluene concentration. The slight decrease in the toluene concentration in the control vial without the bacterial cells (blank) suggests adsorption of toluene onto a butyl rubber septum on the cap or solubilization of toluene in a small amount of water from the wetting support. Thus, our reversible immobilization method uniquely allows us to reuse bacterial cells for different types of chemical reactions after detachment from a support, re-immobilization, and an appropriate induction or reactivation for second reaction.

Reuse of supports for a different type of reaction
To show the further merit of our reversible immobilization method, we attempted to demonstrate the reusability of the support used for the cell immobilization using the schema shown in Fig. 6a. Three pieces of the polyurethane foam support were placed into 30 mL of the suspension of resting ADP1 (pAtaA) cells in 100 mM KCl solution (OD 660 = 1.0) and shaken in a 100-mL Erlenmeyer flask at 115 rpm at 30 °C for 30 min. The cells loosely attaching to the support were removed by dipping them into 100 mM KCl solution and gentle squeezing. A piece of the support with the immobilized ADP1 (pAtaA) cells was placed into an esterase reaction buffer in a test tube. After a 10-min incubation of the test tube at 28 °C, the reaction buffer turned from colorless to yellow due to the 4-nitrophenol produced by esterase on the immobilized ADP1 (pAtaA) cells, as with the Tol 5 cells (Fig. 6b).
Three pieces of the support used in the reaction were collected and washed in 100 mL of dH 2 O in a 500-mL Erlenmeyer flask by shaking at 115 rpm for 10 min. This washing step was repeated three times to thoroughly remove ADP1 (pAtaA) cells. Subsequently, three pieces of the used support were transferred into 30 mL of a  [22], which has the ability to produce indigo from indole using its phenol hydroxylase [29,30], in 100 mM KCl solution (OD 660 = 1.0) and shaken in a 100-mL Erlenmeyer flask at 115 rpm at 30 °C for 30 min for cell immobilization. Pristine pieces of the polyurethane foam support were also subjected to the cell immobilization for a control experiment. The cells loosely attaching to the support were removed by dipping them into 100 mM KCl solution and gentle squeezing. Each piece of the support with the immobilized ST-550 (pAtaA) cells was placed into an indigo reaction solution and incubated for 12 h. Indigo produced by the immobilized ST-550 (pAtaA) was extracted with N,N-dimethylformamide (DMF). Figure 6c shows the solution extracted from each support. The quantity of indigo produced from each support is shown in Fig. 6d. The productivity with the reused support was similar to that with the pristine support, implying that the reused support retained its capability for bacterial immobilization after the immobilization/ detachment process. Thus, our reversible immobilization method uniquely allows us to reuse supports, even for different chemical reactions, via the immobilization of different bacterial cells.

Discussion
Immobilization of biocatalysts simplifies product separation, stabilizes biocatalysts, and enables the repeated or continuous use of biocatalysts, which are typically expensive to produce [12,13]. However, they are usually discarded, together with the supports, after a reaction. In protein immobilization, many techniques for reversible immobilization of enzymes (e.g. lipase, amyloglucosidase, glucoamylase, and aminoacylase) have been studied to enable the regeneration and reuse of support materials [31][32][33][34][35]. Additionally, with regard to bacterial cell immobilization, reversibility should be beneficial. However, the reuse of gel supports is impossible after their use in entrapment immobilization, which is most frequently employed for bacterial cells. Biofilm reactors are also used in the production of valuable compounds, such as alcohols and organic acids, not just in wastewater treatment [15,36,37]. These bioreactors use biofilms formed on support materials as immobilized microbial cells [15,38]. Once biofilms are formed, it is difficult to completely detach them from supports by simple treatments. Therefore, when catalytic activity decreases or a chemical reaction has to be switched for another one, the support with biofilms would be discarded and a new biofilm would be reconstructed on the fresh support. However, a long startup time is required to rebuild an active biofilm. In this study, we found that the stickiness of the AtaA molecule is drastically diminished at a lower ionic strength and is completely lost in dH 2 O (Fig. 1). Cell adhesion mediated by AtaA also depends on ionic strength in the same manner as the AtaA molecule, and even bacterial cells previously adhered to supports through AtaA can be detached in dH 2 O (Figs. 2, 3). Based on this adhesion property, we have established a reversible immobilization method for microbial cells (Fig. 4) and demonstrated the reuse of both cells and supports by means of this reversible immobilization (Figs. 5, 6).
Cells immobilized with AtaA can be detached from supports by a simple manipulation involving a dH 2 O wash and active cells can be quickly immobilized onto the same previously used support. Because this method is not based on the characteristics of support materials but the unique adhesion property of AtaA, various materials can be employed as reusable supports. For example, supports that have a structure with pores, fibers, or slits for a large surface area and are formed in a combined unit or integrated into a reactor vessel might be used.
In our new reversible immobilization method, both cells and supports can be reused for different types of chemical reactions. Three patterns can be considered about reused processes; (1) used cells are re-immobilized onto a new support, (2) fresh cells are immobilized onto a used support, and (3) used cells are re-immobilized onto a used support. In other words, one of or both of cells and a support are reused. It is expensive to grow bacterial cells on a medium containing many kinds of chemicals, such as nutrients, inducers, and antibiotics, using energy for sterilization, agitation, aeration, and temperature control.
In this study, we demonstrated that bacterial cells can be reused for a different type of chemical reaction after a simple induction or reactivation step. We can choose a different support material that is suitable for the subsequent reaction. For example, we used polyurethane foam for the first reaction of ester hydrolysis in a buffer and steel wool for the second reaction of toluene degradation in a gas phase. In addition, the polyurethane foam can also be reused for a different reaction, such as indigo production, after immobilization of a different bacterial strain. The reuse of support material mitigates problems caused by the use of support materials, such as the cost and waste of support materials. Reversible microbial cell immobilization would make bioreactors and bioprocesses simpler, more efficient, more cost-effective, and more convenient.

Conclusions
In summary, we found that the stickiness of isolated AtaA PSD and cell adhesion mediated by AtaA are drastically diminished in deionized water and that deionized water even detaches bacterial cells previously adhered to support in a salt solution. Using this phenomenon, we invented a unique reversible immobilization method that allows us to reuse bacterial cells and supports for different chemical reactions by a simple manipulation involving a dH 2 O wash. This method for the immobilization of bacterial cells using AtaA would make bioprocesses more cost-effective and enhance their commercial use for environmentally friendly chemical productions.

Bacterial strains and culture conditions
The bacterial strains used in this study are detailed in Table 1. These bacterial strains were grown as described previously [20].

Quartz crystal microbalance
The adhesiveness of AtaA PSD was measured using a QCM system (AFFINIX Q8; ULVAC, Kanagawa, Japan) as described previously [26]  For a cell detachment assay, each well was rinsed three times with 200 µL of 100 mM KCl solution or dH 2 O using a micropipette. The remaining cells were quantified by crystal violet staining as described above. For reattachment, the cells detached by rinsing each well with dH 2 O were collected by centrifugation, resuspended in 100 mM KCl solution at an OD 660 of 0.5, and added to a new well for the next immobilization cycle.

Immobilization of bacterial cells onto support materials
Polyurethane foam with a specific surface area of 37.5 cm 2 /cm 3 (CFH-30; Inoac Corporation, Nagoya, Japan) in the shape of a cube (1 cm 3 ) was used as a sponge support. The steel wool used in this study was the same as that previously used [25], which was purchased from Handy Crown (Tokyo, Japan).
To immobilize bacterial cells onto the polyurethane foam support, cells were suspended in 100 mM KCl solution in a 100-mL Erlenmeyer flask. The value of the OD 660 was adjusted to 2.0 for the visualization of the cell detachment from the support and for the analysis of the time profile of cell immobilization or to 1.0 for use in chemical reactions. Pieces of the support were placed into the cell suspension and shaken at 115 rpm at 28 or 30 °C for 10-30 min. For analysis of the time profile of cell immobilization, the OD 660 of the cell suspension was measured periodically. The immobilization ratio of the cells was calculated from the following equation: To immobilize bacterial cells onto the steel wool support, the cells detached from the three pieces of the polyurethane foam support used for the esterase reaction were resuspended in 30 mL of BS medium [39] in a 100-mL Erlenmeyer flask. Into this cell suspension, 300 mg of the steel wool support was placed and shaken at 115 rpm at 28 °C for 1 h.

Flow cytometry
Bacterial cells before and after the cell immobilization/ detachment test were resuspended in PBS containing 4% paraformaldehyde and incubated at room temperature for 15 min. The samples were washed with PBS and treated with anti-AtaA 699-1014 antiserum diluted 1:10,000 in PBS. After a 1-h incubation at room temperature, the samples were washed twice with NET buffer (150 mM NaCl, 5 mM EDTA, 50 mM Tris-HCl, 0.05% Triton X-100, pH 7.6) and treated with Alexa Fluor 488-conjugated antirabbit antibody (Cell Signaling Technology, MA) diluted 1:500 in NET buffer for 30 min. Finally, the samples were resuspended in dH 2 O, and the fluorescence was measured by FACS Canto II (Becton, Dickinson and Company, NJ).

Chemical reactions by immobilized bacteria
For the measurement of cell-bound esterase activity, cells immobilized on plate wells were reacted with 1.9 mM 4-nitrophenyl butyrate (4-NPB) in 200 μL reaction buffer (1.1% Triton X-100, 50 mM 3,3-dimethylglutaric acid, 50 mM Tris, 50 mM 2-amino-2-methyl-1,3-propanediol) at 28 °C for 30 min. Triton X-100 was eliminated from the reaction buffer when the esterase activity of Tol 5 was measured. The A 405 of 4-nitrophenol produced by the reaction was measured by a microplate reader. To measure the esterase activity of cells immobilized on the polyurethane support, a piece of the support with the immobilized cells was placed into 3 mL of the reaction buffer in a test tube and incubated at 28 °C for 10 min.