1 Introduction

Activities within the biorefinery area have meant a revitalization of the interest for immobilized cells and enzymes [1, 2]. Such preparations offer possibilities for reuse of the biocatalyst, improved volumetric productivity and stable operational conditions [3]. A key issue with immobilized preparations has been diffusion limitations due to the presence of the biocatalyst throughout the solid phase [4, 5]. Alternatively, preparations with the catalyst adsorbed on the surface of the matrix have been used. Leading themes in development of immobilized preparations have been high catalytic density, low diffusion limitation, environmentally friendly and low cost processing [6, 7].

All published immobilization techniques have their advantages and limitations [8, 9]. High catalyst density for immobilized enzymes has been obtained by cross-linking enzyme crystals or by stabilizing enzyme aggregates (CLEA) by cross-linking [10, 11]. With regard to cells, preparations containing cells entrapped in three-dimensional polymer networks is the most commonly used, besides the adsorbed cells. An alternative that is presented here involves to build three-dimensional macroporous networks, often in monolith shape, built from cross-linked cells [12, 13].

One way of creating such macroporous monolith materials is to carry out the gel forming process at subzero temperature. The simple combination of two steps: freezing–thawing of a reaction mixture resulted in monolith preparation with desired size of interconnected pores [14]. These materials, often called “cryogels”, have been proven to be useful in biotechnology as they provide exceptional properties of highly porous structure and good operational stability [15]. A set of studies have been performed in developing immobilized biocatalysts based on cryostructuration technique. For example, the freezing of monomers mixed with the catalyst system for polymerization and with proper biological molecules/structures yields polyacrylamide cryogels with entrapped enzymes or cells in the walls of cryogel structure. These preparations provide large pores that help to overcome problems of mass transfer [16]. However, the thin polymer walls of the cryogel have a high content of polymer which contributes to a reduced diffusion within the gel which negatively influences the performance of the entrapped enzyme. The situation will be similar, but less pronounced when cells are entrapped since the cell membranes constitute a diffusion hindrance per se.

In an effort to produce a preparation of immobilized cells with a minimum of diffusion restrictions combined with good mass transfer a new type of cryogels was designed. By using a minimum of cross-linking reagents it was possible to form a cryogel consisting of cross-linked microbial cell and a few per cents of crosslinker. The thin polymer filaments separating the pores make mass transfer efficient and these thin walls do not constitute any major diffusion restriction for exchange of substrate and products to/from the cells. A natural first choice of cross-linker was glutaraldehyde (GTA) which has been used extensively both for immobilizing CLEA [17, 18]. A very serious drawback with GTA turned out to be the effect on cell viability upon exposure of the cells for the cross-linker [19]. In the present application the GTA-treatment also inactivated the enzyme studied. Therefore, efforts were spent to design and use new macromolecular crosslinkers that would form the desired structures while maintaining cell viability.

As model organism studied here was chosen Escherichia coli with a cloned thermostable and his-tagged β-glucosidase.

2 Materials and Methods

2.1 Materials

Dextran, technical grade T500 was supplied from Pharmacia (Uppsala, Sweden). Polyethyleneimine (PEI, MW 1800) was obtained from Polysciences Inc. (Warrington, PA, USA). GTA solution (50 % v/v), potassium periodate (KIO4), sodium ampicillin, poly (vinyl alcohol) (PVA) (MW 89,000–98,000), isopropyl- β D-thiogalactopyranoside (IPTG) and 4-nitrophenyl β-D-glucopyranoside (pNPG) were purchased from Sigma-Aldrich (Steinheim, Germany). SpectroPor Dialysis membranes (MWCO 1; 3, 5 and 12 kDa) were obtained from SpectrumLab, Inc. (Rancho Dominguez, USA). LB broths (Bacto TM) and glucose were purchased from Difco Laboratories (Detroit, MI, USA). All other chemicals were of analytical grade and obtained from commercial sources. A recombinant strain of E. coli containing (His)6-β-glucosidase originally from thermophilic bacterium Thermotoga neopolitana was provided by Department Biotechnology, Lund University, Sweden.

2.2 Synthesis of Polymeric Crosslinkers

Oxidized dextran (OxDex) was performed by reacting dextran with KIO4 according to the standard procedure described previously [20]. Oxidation of dextran with excess of KIO4 leads to formation of 1,5-dialdehyde moieties from glucose, while the polymeric structure is maintained [21]. The method was slightly modified in order to enhance aldehyde content of the final polymer. Dextran T500 (2 g) was suspended in 20 ml of water and a first portion of KIO4 (0.75 g) was added. The reaction mixture was stirred until KIO4 was completely dissolved. A second portion of KIO4 (0.75 g) and a third portion (0.5 g) were dissolved in the reaction mixture. After 2 h of mixing at dark conditions, the reaction mixture was dialysed (dialysis membrane with pore size 3.5 kDa) against water (5 l) during 3 days with daily change of water.

PEI was modified by introducing reactive aldehyde groups mainly via modification of the primary and partially via secondary amino groups of the polymer. PEI (0.1 g) was dissolved in 10 ml of water and then mixed with 1.88 ml of GTAsolution (50 % v/v) under stirring. The mixture was incubated for 1 h during which the solution changed its colour from transparent to orange, indicating the formation of Schiff’s bases in the structure. The solution was then transferred into a dialysis bag (with pore size 1 kDa) and dialysed against water (5 l) during 3 days with replacement of the water twice per day.

Modification of PVA was done similarly to the modification of PEI. Reactive aldehyde groups were introduced by reaction with GTA followed by a dialysis step. PVA (0.1 g) was dissolved in water (5 ml), with subsequent addition of GTA solution (0.3 ml). The mixture was stirred at room temperature. After 1 h of incubation, the polymer solution was treated to remove unreacted GTA. The solution was dialysed (with pore size 12 kDa) against water (5 l) at +4 °C during 3 days with daily change of water.

2.3 Preparation of Cryogels Based on Crosslinked Cells

E.coli Bgl A cells were cultivated in LB medium with 0.1 mg/ml of ampicillin (at +37 °C, 150 rpm) until stationary phase was reached. Further, cells were harvested by centrifugation (10 min, at 10,000×g) followed by washing with ice-cold sodium chloride solution (0.9 %). The supernatant was discarded and the cell pellet (150 mg) was mixed with ice-cold aqueous solution of crosslinkers (0.5 ml). In order to avoid bubble formation, the re-suspension of cells in crosslinker solution was carried out by gentle mixing. The reaction mixture was then quickly transferred into glass tubes (7 mm in diameter) and frozen at −12 °C. After 1-3 days of reaction, samples were thawed at room temperature. The formed cryogels, based on crosslinked cells, were washed with sodium chloride solution (0.9 %).

2.4 Characterization of Cryogels

The water permeability measurements were done in order to estimate the flow-through of prepared cryogels. Velocity of water passing through the cryogel column (10 mm long, 7 mm in diameter) was determined. Triplicate measurements of samples for each type of cryogel were carried out.

Mechanical stability of produced cryogels was evaluated through performing the compression test using TA-XT2 instrument (Stable Micro System, Godalming, Surrey, UK) Samples (10 mm long, 7 mm in diameter) were compressed 50 % of their initial height at a speed of 1 mm/s. The elastic modulus was calculated in the linear region at 30 % of deformation using the following equation:

$$\varvec{E} = \frac{{(\varvec{F}/\varvec{A})}}{{({\varvec{\Delta}}\varvec{l}/\varvec{l})}}$$

where E is elasticity modulus in Pa, F is applied force defined in N, A (m 2) is area of the cryogel sample, Δ l (m) is the change in length and l (m) is a length of sample. Triplicate measurements were done.

Pore size distribution and morphology of cryogels made from crosslinked cells were analysed using scanning electron microscopy (SEM). Sample discs (1 mm thick) from different cryogels were fixed in GTA solution (2.5 % in 0.1 M sodium phosphate buffer, pH 7.4) overnight at +4 °C, with gentle mixing. The cryogel discs were rinsed with water to remove GTA traces and were further dehydrated gradually in ethanol solution (20, 50, 75, 95 and 99.5 % respectively). The samples were then dried at critical-point and sputter-coated with gold/palladium (40:60). Images of crosslinked cells were taken using Hitachi SU3500 scanning electron microscope.

2.5 Viability Test

To evaluate viability of the crosslinked cells, cryogels prepared in presence of different crosslinkers were placed in LB medium together with glucose and ampicillin (1 and 0.1 mg/ml, respectively). In order to prevent contamination, cryogels were washed with sterile sodium chloride solution (0.9 %) three times under sterile conditions before incubation. During incubation in a shaker (+37 °C, at 70 rpm), the glucose consumption and cell growth was checked with time. The measurement of glucose in LB medium was done by ACCU-CHEK Aviva device (USA) whereas the cell growth was observed by monitoring the optical density of the medium at 600 nm. All experiments were performed in triplicates. Experiments with sliced cryogel and with free cells (150 mg) as a control were carried out by the same procedure as described above.

2.6 Expression of β-Glucosidase

In order to produce β-glucosidase in E. coli Bgl A cells, the inducer Isopropyl- β-D-IPTG was used. When the optical density at 600 nm of the cell suspension in the cultivation medium with antibiotics reached to 0.7, IPTG was added (to a final concentration of 1 mM) and further cultivated overnight. Then, induced cells were used for preparation of cryogel. The expression of β-glucosidase in crosslinked cells was induced by incubating cryogel (consisting of 150 mg of cells) in 150 ml of LB medium with addition of IPTG to 1 mM of final concentration. After incubation at +37 °C overnight, samples were thoroughly washed with PBS buffer and activity assay was conducted.

2.7 Enzymatic Assay

β-Glucosidase activity was measured in cells by using pNPG as a substrate. Cryogels, consisting of 50 mg of crosslinked cells, were placed in 3 ml of pre-warmed solution (0.1 M sodium phosphate buffer, pH 6.7) containing substrate (at a concentration of 5 mM). Reaction was run at +37 °C under stirring. The formation of para-nitrophenol was measured by spectrophotometer at 405 nm and concentration of formed product was calculated from a calibration curve for para-nitrophenol. The same procedure was applied to determine activity for free cells. Measurement of enzymatic activity for each type of cryogel was performed in triplicate samples. Cryogels prepared from synthetic crosslinkers were evaluated for enzyme stability. The samples were stored at +4 °C in phosphate buffered saline (PBS), pH 7.4 during 30 days. The enzyme activities of samples were checked at different stages followed by washing after assay and continued storage in PBS buffer solution prior to next measurement of activity.

3 Results and Discussion

3.1 Cryogels Built from Crosslinked Whole Cells

This study deals with macroporous monolith structures composed from metabolically active crosslinked cells that were produced through cryostructuration technique. By using the freezing-thawing procedure where cells and crosslinker are treated together, a biocatalyst with high catalyst density was produced. Cryogels from crosslinked cells were prepared using macromolecular crosslinkers: OxDex, PEI + GTA and PVA + GTA and also the low molecular weight compound GTA. The reactive macro polymers were synthesized to be used for crosslinking the cells. The strategy chosen involved the exploitation of aminogroups on the cell surface and aldehyde functions on the polymers. The detailed procedure is demonstrated in Fig. 1. The option to prepare macroporous structures through crosslinking cells by 0.5 % of GTA in the freezing-thawing procedure was reported earlier [13]. However, earlier studies have shown that low molecular weight crosslinkers such as GTA penetrate into the cells and kill them. Therefore, a strategy to go for macromolecular reagents was chosen. The polymeric crosslinkers are big enough not to penetrate through the cell wall and are thus capable of reacting only with amino groups on the cell surface thereby producing metabolically active crosslinked cells.

Fig. 1
figure 1

Scheme of cryogel preparation from whole cells through freezing-thawing procedure using polymeric crosslinkers. The reaction between amino groups presented on cell surface with macromolecular polymers is taking place in the unfrozen phase during reaction stage at −12 °C

These polymers were tested in preparation of cryogels from whole cells. It is desirable to keep polymeric fraction as low as possible in the biocatalysts in order to increase the density of cells and thereby the volumetric activity. It was observed that formation of cryo preparations strongly depends on the concentration of crosslinker in the sample. The critically lowest concentration of each crosslinker was determined that was capable of creating stable macroporous cryogel structures. It was found that required amounts of polymeric crosslinkers to form stable structures were slightly higher (1 %) in comparison to that of GTA, data shown in the Table 1. The experiments showed that use of different crosslinking structures resulted in variation of the properties of the final preparations. By carrying out water permeability test it was demonstrated that crosslinkers like OxDex and PVA + GTA form a structure with relatively slow flow through. In case of cryo(OxDex), speed of the water passing through the column (10 mm long and 7 mm in diameter) was 0.04 ml/min while no flow-through was observed for the cryo(PVA + GTA) sample. Data are displayed in Tables 1 and 2. Samples of cryo(GTA) and cryo(PEI + GTA) showed opposite picture and had high flow through of water. Interestingly, the evaluation of mechanical properties in cryogels produced using macromolecular crosslinkers revealed an almost opposite picture. Compression tests were carried out in order to evaluate rigidity and elasticity of formed cryogel structures. Results demonstrated that samples cryo(OxDex) and cryo(PVA + GTA) formed more stable cryogels with more elastic structures than that of cryo(PEI + GTA). Even the calculated elastic modulus was higher in cryo(PEI + GTA) but this sample was fragile and was destroyed at compression to 50 % of the initial length whereas other samples were able to restore their original shape after compression. SEM images demonstrate macroporous structure in all samples, a feature important for efficient mass transfer during application (Fig. 2). Visible observations demonstrated no differences in the structure of the prepared cryogels. Images showed cells tightly attached to each other and densely packed into wall structure. The use of one of the polymeric crosslinkers, formed the cryogel structure of desired stability and porosity in one preparation having comparable properties with those of cryo(GTA). Improvement of properties in cryogel was achieved via crosslinking of cells with a combination of the two synthetic polymers: PEI + GTA and PVA + GTA (P/P). The elastic and stable structure with high flow through in sample cryo(P/P) was formed as a result of interaction of two synthetic polyaldehydes. Measured flow through for cryo(P/P) was comparable to that of cryo(GTA) flow-through. (Table 1, sample 5) Also, a higher rigidity was obtained in the sample; even if calculated elasticity modulus was lower than in samples prepared from GTA, still the cryo(P/P) showed stability properties judged to be sufficient for the applications.

Table 1 Elasticity modulus (kPa) calculated at 30 % of strain and water permeability for cryogels prepared using different crosslinking reagents after 3 days in reaction at −12 °C
Table 2 Calculated elastic modulus (kPa) of cryogels applied for compression test after different days in reaction at −12 °C. The stress was calculated at 30 % of strain
Fig. 2
figure 2

SEM images of cryogels. Pictures of cells that were crosslinked with a, b 0.5 % of GTA, c, d 1 % of OxDex, e, f 1 % of PVA + GTA, g, h 0.55/0.35 % of P/P. Pictures demonstrate each sample with lower and higher magnifications

It was furthermore noticed that the reactivities of the synthesized macromolecular crosslinkers differ from that of GTA and longer time is needed for the formation of stable cryogel structures. Samples based on use of polymeric crosslinkers were kept in frozen state for 2–3 days as compared to samples containing GTA where 1 day was enough to form cryogels. In case of activated polymers, a cryogel structure was formed after 24 h, however, these cryogels did not recover their original shape after mechanical compression, Table 2. Samples prepared from synthetic polymers P/P were able to restore their original forms after compression to 50 % of their original length only after they had been under cryo-reaction for 3 days. Samples produced according to this method were used in further experiments in order to investigate metabolic activity of cryogels prepared from cells crosslinked by synthetic polymers.

3.2 Viability of Crosslinked Cells in Cryogels

Cryogels prepared from crosslinked cells were evaluated for metabolic activity present in cells after treatment with different types of polymers. Cell growth during incubation of cryogels in LB medium was observed thereby indicating that use of any of the macromolecular crosslinkers preserved cell viability during the cryostructuration procedure. As demonstrated on the Fig. 4, cells from sample cryo(OxDex) consumed the total amount of glucose in LB medium faster than did the cells from the sample of cryo(P/P). The sample from OxDex was not stable and started to dissolve from the beginning of incubation in LB medium due to that the natural carbohydrate structure of dextran was utilized by the cells as a carbon source. The metabolic activity in cryogel structure prepared using P/P was somewhat lower. There might be some diffusion limitation due to the more rigid structure. In order to prove this hypothesis, the sample cryo(P/P) was sliced into small pieces and incubated in LB medium at the same conditions as described above. As shown in Fig. 3, the curve displaying values of total glucose consumption by cells from sliced sample cryo(P/P) almost resembles the data of glucose consumption for free cells, indicating the high content of viable cells present in cryogel. No glucose consumption or cell growth were observed during incubation of cryo(GTA) for 24 h indicating efficient inactivation of metabolic activity in cells after GTA treatment. These results demonstrated that macromolecular structures preserve metabolic activity in crosslinked cells better than the low molecular weight does. Cryogels consisting of viable microorganisms can serve as inoculum in fermentation processes [22].

Fig. 3
figure 3

Plots of glucose consumption (a) and cell growth (b) during incubation of cryogels in LB medium at +37 °C with time. Concentration of glucose (g/l) was checked using ACCU-CHEK Aviva device, cell growth was monitored by measuring of the optical density at 600 nm. Each sample was run in triplicate

3.3 Enzymatic Activity of Crosslinked Cells

The β-glucosidase activity present in E. coli cells used for cryogel preparation was investigated (Fig. 4). About 90 % of β-glucosidase activity from free cells was retained in cells crosslinked with the mixed synthetic polymers cryo(P/P). As a control, enzymatic activity in sample cryo(P/P) prepared from non-induced cells was determined. The low enzymatic activity observed in noninduced cells is ascribed to the natural level of enzyme produced in the cells. The enzymatic activity of cryo(GTA) was dramatically diminished. Only 15 % of β-glucosidase activity in induced cells was left whereas non-induced cells showed some negligible activity during assay.

Fig. 4
figure 4

Specific activity of β-glucosidase measured in free cells and in crosslinked cells. The bars from left to right represent () free cells, () cells pre-induced prior to cryo-structurization using P/P, () cells pre-induced prior to cryo-structurization using GTA, () non-induced cells cryo-structured using P/P and () non-induced cells cryo-structured using GTA. Activity is defined as amount of para-nitrophenol (nmol) formed per minute for 1 mg of cells

High yields of retained enzymatic activity as demonstrated for the cells crosslinked using low amounts of macromolecular reagents indicate retained enzymatic activity in the cells. To evaluate if a large portion of the metabolic machinery was intact, induction of protein expression was performed. Preparation cryo(P/P) produced from non-induced cells was induced with IPTG in order to evaluate if over expression of β-glucosidase could take place in crosslinked cells. The subsequent assay for β-glucosidase activity revealed increased enzyme activity. As demonstrated in the Fig. 5, the induced β-glucosidase activity reached almost 50 % of that of crosslinked cells. A negative control constituting the sample from GTA is illustrated as well; no activity appeared in inactivated cells. These results provide strong evidence that macromolecular structures act mildly to cells thereby preserving their viability while GTA reagent completely inactivated the metabolic activity of the cells.

Fig. 5
figure 5

Specific activity of β-glucosidase induced in cryogels based on crosslinked cells. The bars from left to right represent () cells pre-induced prior to cryo-structurization with P/P; () cryo(P/P) produced from non-induced cells; () crosslinked cells induced post-immobilization; () pre-induced cells cryo-structured using GTA, () non-induced cells cryo-structured using GTA and () GTA-crosslinked cells induced post-immobilization. Activity is defined as amount of para-nitrophenol (nmol) formed per minute for 1 mg of cells

The storage stability of cryogel consisting of viable crosslinked cells was evaluated. It is important to see how enzyme activity drops with time due to the naturally occurring cell lysis in viable crosslinked cells. It was shown that activity of cryo(P/P) decreased slightly and after 30 days almost 80 % of initial enzymatic activity was retained (Fig. 6). The synthetic polymers help to stabilize cells and prevent their fast lysis which is important criteria for reusability of biocatalyst.

Fig. 6
figure 6

Residual activity of cryo(P/P) preparation. Samples were stored in PBS buffer solution (pH 7.4) at +4 °C during 30 days. Average of results from three samples is displayed

Finally, the best proportion of combinations of synthetic polymers was screened to evaluate the polymers influence on the stability, properties and enzyme activity. Almost equal enzyme activity was presented in cells crosslinked using a range from the lowest amount of polymers (0.45 %) to the highest amount (1.3 %) (Fig. 7). These results indicate that synthetic polymers gently act toward cells and amount of polymers did not affect on metabolic activity. The volume of polymeric fraction was essential only for stability of the cryogel system. The elasticity modulus was improved with rise of crosslinker amount used for cryogel preparation. Crosslinker amount of 0.9 % was enough to stabilize the cryogel structure. The measured elasticity modulus for this sample was 3.15 kPa. This is important since our focus was to obtain preparations with high catalyst content using as low polymer content as possible in the biocatalyst.

Fig. 7
figure 7

Specific activity and elasticity modulus for cryogels prepared from synthetic polymers P/P at different proportions. Activity is defined as amount of product produced per minute for 1 mg of cells. Elasticity modulus is calculated at 30 % of strain and given in kPa

4 Conclusion

Biocatalysts with high catalyst density and favorable mass transfer conditions were produced by crosslinking whole cells in presence of polymeric reagents in a freezing-thawing procedure. The operation with polyaldehyde crosslinkers like modified structures of PVA and PEI leads to a high degree of retained activity. Prepared cryogels consisting of viable cells represent advanced materials that might be considered as a multifunctional catalyst with the opportunity of inducing a desired protein expression. Monoliths, due to their stable interconnected porous structure showed no diffusion limitation in operation in a flow-through reactor for enzymatic conversion. Moreover, the production of such biocatalysts represent an eco-friendly approach since this material mainly consists of immobilized cells and low content of polymeric crosslinkers that could be easily degraded after utilization and therefore they are not hazardous to environment.