Introduction

Cellulose is the most abundant biopolymer found in nature and is produced by a wide variety of organisms, ranging from plant to algae and prokaryotic organisms (Delmer and Amor 1995; Klemm et al. 2005). However, bacterial cellulose (BC) has been more interesting and useful due to its higher purity and highly crystalline nanostructure. BC structures are formed by extracellular-excreted nanofibers produced by various species of bacteria such as Acetobacter, Aerobacter, Azotobacter, Agrobacterium, Achromobacter, Gluconacetobacter, Rhizobium, Sarcina and Salmonella (Brown 2004; Jahn et al. 2011; Morgan et al. 2013). Among the mentioned genera, the most effective cellulose producing specie is Gluconacetobacter xylinus (formerly Acetobacter xylinum) which is often used as model organisms in bacterial cellulose production studies (Keshk et al. 2006; Nguyen et al. 2008). At present, G. xylinus is reclassified as Komagataeibacter xylinus according to its 16S rRNA gene sequences (Yamada et al. 2012). K. xylinus is an aerobic gram negative bacterium, has ability in producing cellulose extracellularly at temperatures between 25 and 30 °C and pH from 3 to 7, using glucose, fructose, sucrose, mannitol, and other carbohydrates, as carbon sources (Castro et al. 2011).

BC has distinctive characteristics including high mechanical strength, high purity, and extremely fine cellulose fiber network structure of around 100 nm and 100 µm in diameter and length, respectively. This ribbon-like network structure is consisted of bundles of microfibrils of 2-4 nm in diameter (Gindl and Keckes 2004; Hirai et al. 2002; Iguchi et al. 2000; Nakagaito et al. 2005). In its native state, BC is a water-swollen network of cellulose nanofibers. BC has coherent 3D network structure with porosity, high water holding capability, and high biocompatibility. Due to these special characteristics, BC has become a very useful biomaterial in many applications such as paper making (Yamanaka et al. 1989), membrane (Iguchi et al. 2000), food industry (Budhiono et al. 1999), and biomedical materials (Chang and Zhang 2011).

The BC structure and production yield can be affected by several factors including types of BC producing bacterial strains, fermentative media, carbon sources, and growth conditions which can be altered to obtain BC with desirable properties. The BC production yield is mostly influenced by the relation of types of bacterial strain and carbon sources used. For example, Gluconacetobacter hansenii NCIM 2529 produced the highest BC yield from a sucrose containing medium (Mohite et al. 2013), whereas Gluconacetobacter xylinus ATCC 53524 and PTCC 1734 had the highest ability in BC production when mannitol was used as a carbon source (Mohammadkazemi et al. 2015; Ruka et al. 2012). The determination of an optimal medium and an appropriate set of growth conditions for high levels of BC production would be beneficial for implementing this technology to an industrial scale.

Interestingly, the BC morphology and characteristic properties are greatly affected by different cultivation methods, static and agitated. In the static condition, a BC pellicle in a gelatinous form is produced at the surface of the culture medium. In contrast, various sizes (10 µm–1 mm) and shapes (spherical, ellipsoidal, stellate or fibrous) of well-dispersed BC in the culture medium are synthesized in the agitated condition, depending on types of BC producing bacterial strains (Dudman 1960; Hestrin and Schramm 1954; Yoshinaga et al. 1997). The BC produced through the agitated cultivation method displayed some microstructural changes which were the decrease in degree of polymerization, crystallinity index, cellulose Iα content, and crystallite size, but the increase in water holding capacity as compared to static cultivation (Czaja et al. 2004; Krystynowicz et al. 2002; Watanabe et al. 1998). Consequently, the BC with different characteristics suitable for different applications can be produced by varying these variable factors.

In this study, the aim was to evaluate the effect of various BC producing bacterial strains, carbon sources, and fermentation conditions on the BC production. The synthesis and structural characteristics of BC produced with six different carbon sources in static and agitated cultures by five different strains of K. xylinus, which are commercially available in Thailand, were investigated. The BC production yield was measured as the dry weight of cellulose within the volume of medium in liter (g/L). The morphology of resulting BCs was examined by the field emission scanning electron microscopy. To characterize the effects of variable factors on the crystalline arrangement of glucan chains within microfibrils and their crystallite size, the X-ray diffraction was used. Furthermore, the estimation of Iα and Iβ cellulose fractions in BC samples from different culture conditions was carried out using the FT-IR spectroscopy. The obtaining results would be useful database for other researchers to select optimal BC production conditions for their desired BC applications.

Materials and methods

Bacterial strains

Four strains of K. xylinus; K. xylinus TISTR 086, 428, 975, and 1011 (K086, K428, K975, and K1011, respectively) were obtained from the Thailand Institute of Scientific and Technological Research, and one K. xylinus (KX) was obtained from the Institute of Food Research and Product Development, Kasetsart University, Thailand. Each of K. xylinus strains was cultured on glucose yeast extract (GYE) agar containing 100 g d-glucose, 10 g yeast extract, 5 g peptone, 20 g CaCO3, 25 g agar per liter at 30 °C for 3 days. Working cultures were routinely prepared on GYE and stored at 4 °C until use. Chemicals for microbiology and other chemicals used were purchased from Sigma-Aldrich.

Media

The glucose yeast extract broth (GYB) was selected from the literature and modified for the present study. The GYB consisted of 50 g glucose and 5 g yeast extract in one-liter solution. Glucose (GC), fructose (FT), lactose (LT), maltitol (MT), sucralose (SC), and xylitol (XL) were used as different carbon sources in the modified GYB media. Before use, all the media were autoclaved at 121 °C for 15 min. The pH was adjusted to 5.0 with HCl or NaOH.

Growth conditions

For preparation of each K. xylinus strain seed culture, a single colony from a working plate of GYE agar was selected and inoculated in 10 mL of each of six modified GYB media. These seed cultures were incubated for 7 days at 30 °C under static condition. Following growth, bacterial cells were separated from cellulose pellicles in the seed cultures by vigorously shaking; as a result, the cell suspension for inoculation was obtained. Cultures were grown in 250-mL erlenmeyer flasks containing 100 mL of media and inoculated with 5% (v/v) of the cell suspension. The BC production was studied in two conditions: under static condition and agitated condition in which cultures were agitated in a shaking incubator at speed of 150 rpm. The incubation period was 7 days at 30 °C for both static and agitated conditions.

BC purification

After incubation, the BC pellicles and particles were harvested from the cultures and rinsed with distilled water to remove any residual media. The BC products were washed with 2% w/v NaOH at 80 °C for 1 h, and then washed repeatedly with distilled water until a neutral pH was obtained.

Dry weight and yield of BC

The BC production was investigated as the dry weight of cellulose within the volume of medium in liter (g/L). The dry weight of BC was determined by weighing the dried BC pellicles and pellets which were air dried in a desiccator at room temperature for 3 days until reaching constant weight.

Fourier transform infrared spectroscopy (FT-IR)

Each BC sample was air-dried on a glass slide in the form of a thin film. FT-IR spectra were obtained using an ATR Nicolet iD7 FT-IR spectrometer. All spectra were scanned between 4000 and 400 cm−1 with 128 convolutions at a resolution of 4 cm−1. Baselines for each sample spectrum were normalized using the Spectrum software. The fα fraction of the samples was calculated by the following equation (Yamamoto et al. 1996):

$$ f_{a} = 2.55f_{ \propto }^{IR} - 0.32 $$

where \( f_{ \propto }^{IR} \) of cellulose can be calculated as Aα/(Aα + Aβ) and Aα and Aβ are absorbencies at 750 and 710 cm−1, respectively.

X-ray diffraction (XRD)

X-ray diffraction diagrams of dried BC samples were recorded using a Rigaku Model SmartLab 4800 diffractometer with the CuKα radiation wave length (λ = 1.54 Å), generated at a voltage of 40 kV and a filament emission of 30 mA. Samples were scanned from 5–40° 2θ-range at scan speed of 2°/min and scan step of 0.02°. The crystallinity index (CrI) and crystallite size (CrS) were calculated based on X-ray diffraction measurements. Crystallinity index was calculated from the following equation:

$$ {\text{CrI }}(\% ) = \frac{{\left( {I_{110} - I_{am} } \right)}}{{I_{110} }} \times 100 $$

where I110 is the overall intensity of the peak at 2θ about 22.7° and Iam is the intensity of the baseline at 2θ about 18° (Mihranyan et al. 2004). The CrS was determined using the Scherrer equation as following (Cheng et al. 2009):

$$ {\text{CrS}} = \frac{{{\text{K}}\lambda }}{\beta \, \cos \theta } $$

where K is the shape factor (0.9), λ is the X-ray wavelength (1.54 Å), β is the full width at half maximum height (FWHM), and θ is the Bragg’s angle.

Field emission scanning electron microscopy (FE-SEM)

The dried BC samples were sputter coated with platinum in preparation for FE-SEM imaging. The field-emission SEM Hitachi S-4800 model was used, operating at accelerated voltage of 5 kV and magnification of 20 k.

Results and discussion

Production and yield of BC

The culture medium used in this study was the GYB which was a simple medium containing only glucose (or other sugar) as a carbon source, and yeast extract as a protein and mineral source for bacterial growth and bacterial cellulose production. The amounts of BC production from various types of carbon sources by five strains of K. xylinus were measured after 7 days incubation. In the static condition, the bacterial strains KX (1.14–1.84 g/L) and K975 (1.11–1.55 g/L) were comparably able to produce BC pellicles in the highest amount in all carbon sources, then followed by the K1011 (0.57–1.46 g/L); while the K086 (0.14–0.39 g/L) and K428 (0.09–0.22 g/L) had the lowest abilities in BC production as shown in Fig. 1a. The KX strain is generally used as a culture feed in Nata de Coco dessert productions in Thailand. The result of high BC production yield could confirm the benefit of the KX use in the commercial level. Moreover, the KX might be substituted by the K975 with the same high productivity. The highest amounts of BC produced by KX and K975 were 1.84 and 1.55 g/L, respectively when fermenting with glucose.

Fig. 1
figure 1

Yield of BC with different carbon sources by five strains of K. xylinus a in static condition, b in agitated condition

Full size image

Bacterial cellulose is synthesized by microorganisms using a metabolic pool of hexose phosphate which is obtained directly by the phosphorylation of exogenous hexoses, and indirectly through the pentose cycle and gluconeogenic pathway (Colvin and Leppard 1977; Ross et al. 1991; Schramm et al. 1957). Six sugars were selected to study the capability of five strains of K. xylinus in the BC biosynthesis with different carbon sources. Glucose and fructose are hexose monosaccharides with the same chemical formula (C6H12O6) but different in their ring structures. Xylitol is a sugar alcohol and monosaccharide (C5H12O5) derived from xylose (pentose sugar). Lactose is a disaccharide consisting of galactose and glucose. Maltitol is a sugar alcohol and disaccharide (C12H24O11) obtained from the hydrogenation of maltose. Sucralose is a selective chlorinated sucrose, disaccharide (C12H19Cl3O8). All five strains were able to grow and produce BC in the GYB medium with these six sugars. Glucose and fructose gave high BC yields in all strains. Even though these two sugars are utilized by the same pathway for BC synthesis (Schramm et al. 1957), glucose gave the highest amount of produced BC when comparing in each bacterial strains. This phenomenon potentially resulted from the higher ability of glucose in promoting cell growth, consequently high BC yields can be obtained since BC production correlates to cell growth (Ross et al. 1991). Interestingly, maltitol was able to give significantly high BC amounts in the strains KX and K975. The possible reason was maltitol is composed of two molecules of glucose which might be consumed effectively by these two bacterial strains. In the xylitol medium, all strains produced low to moderate levels of BC, except the KX was able to synthesize BC in high amount. In contrast, the sucralose medium gave the lowest BC productions in all five bacterial strains due to its chlorinated modified structure which may not be suitable for bacterial growth (Omran et al. 2013), resulting in low BC production.

Besides the types of microorganism and carbon sources that could influence on the production of bacterial cellulose, the production methods had significant impacts as well. Two methods of BC production, static and agitated fermentation methods, were used in this research. As can be seen from Fig. 1b, the production of BC by three K. xylinus strains (KX, K1011, and K975) were highly influenced by the fermentation method. In the agitated condition, the BC productivity of K1011 and K975 increased significantly. It is generally known that the aerobic bacteria such as K. xylinus can proliferate better at the air/solution interface where oxygen is readily supplied, than in the culture nutrient-rich and oxygen-lean liquid media (Iguchi et al. 2000; Lee and Zhao 1999). In static fermentation, inside the BC pellicles may have concentration gradients of oxygen and nutrients (including carbon sources). This finding confirmed that the oxygen supply significantly influences on the cellulose production (Hestrin and Schramm 1954). The agitated condition could increase oxygen diffusion into the medium, resulting in the higher productivity of K975 and K1011 which the highest BC produced went up to 3.54 and 4.69 g/L, respectively. On the contrary, this action could induce the emergence of mutant cells which their ability in cellulose biosynthesis are depleted, causing a diminish in the production of cellulose. This consequence might explain the decrease in BC production yield of KX under agitated culture.

Morphology of BC

The BC production by different fermentation methods provided the BC with different morphology and properties. Under static condition, all K. xylinus strains formed pellicles at the surface of the culture medium. After 7 days incubation, KX, K1011, and K925 produced smooth and thick BC sheets; but K428 and K086 gave rough and thin BC pellicles which showed their low ability in BC biosynthesis. Nevertheless, the variation of carbon sources did not impact the morphology of BC produced by each strain. Examples of optical images of cellulose produced in the high BC-achieving carbon sources were presented in Fig. 2. The micro-architecture of cellulose pellicles was investigated by FE-SEM as shown in Fig. 3, the FE-SEM images of the obtained pellicles revealed no apparent difference in the morphological appearances and fibril diameters, as the cellulose produced under all conditions displayed the retention of its nanosized-interwoven structure.

Fig. 2
figure 2

Optical images of BC produced by five strains of K. xylinus a–e in static condition, f–j in agitated condition

Full size image
Fig. 3
figure 3

FE-SEM images of BC produced by five strains of K. xylinus a–e in static condition, f–j in agitated condition (all views with 20,000× magnification)

Full size image

In the agitated condition, the BC products were not in pellicle sheet forms, but appeared in forms of spherical shape and irregular granules that were well dispersed in culture media. The BC particles were accumulated in the agitated media with different sizes and shapes, including solid sphere-like BC (Fig. 2f, g, h, and j), and flocky asterisk-like BC (Fig. 2i). After purification with 2% NaOH solution at 80 °C, the resulting particles were white and semi-transparent appearances with diameters of 5–20 mm. The FE-SEM images as shown in Fig. 3f–j (Agitated-BC) illustrated that the BC microfibrils were in disorderly reticulated structures and consisted of 50–100 nm width ultrafine fibrils which were similar in size, not dependent of the different bacterial strains. The fibril arrangement of agitated BCs was looser and larger in pore size. There were some changes in the surface structures, the microfibrils were appeared in stretchable woven structures with the larger cracks webbing between cellulose microfibrils. The stretching cellulose microfibrils and larger holes between woven cellulose microfibrils may be a result of the disrupted microfibril formation by the agitated treatment during the fermentation process. According to their distinctive properties, the sphere-like BCs with the tenuous woven microfibrils have been found to be useful as adsorbents for heavy metal ions and sewage treatment (Zhu et al. 2011). In addition, according to their good biocompatibility, they might be useful in development of specific biomedical materials, such as bone type tissue scaffolds, bioseparation, immobilized reaction, and cell suspension culture (Hu and Catchmark 2011).

Crystallinity and crystallite size

Effect of bacterial strains

The XRD and FT-IR techniques were performed to characterize the crystalline structure and the mass fractions of cellulose Iα and Iβ of the BC samples produced in the chosen media. Cellulose is a homogeneous polycrystalline macromolecular compound composing of crystalline (ordered) and amorphous (less ordered) regions (Bi et al. 2014). Cellulose I is the naturally mostly found form of cellulose which consists of parallel chains (Delmer and Amor 1995), and occurs in two distinct allomorphs, meta-stable state celluloses Iα and stable state celluloses Iβ (Atalla and Vanderhart 1984). The XRD analysis results of bacterial cellulose harvested after being cultivated for 7 days were shown Fig. 4. There were three main diffraction peaks at 2θ = 14.5°, 16.7° and 22.7° which can be corresponded to the (100), (010), and (110) planes of cellulose Iα, or the (\(1\bar{1}0\)), (110), and (200) planes of cellulose Iβ (French 2014). Due to the co-existing peak positions of two allomorphs, it is difficult to differentiate the two allomorphs by only determining the XRD peak positions. Nevertheless, the intensity of the 14.5° peak was higher than the 16.7° peak, indicating the unique properties of typical BC samples produced by K. xylinus which contain mostly cellulose Iα (Tokoh et al. 1998). For the shapes of cellulose crystallites, they are in square cross-sectional shape if the intensities of the (100) and (010) reflections are nearly equal; however, the slightly stronger (100) reflection in Fig. 4a and b suggested that the shape of cellulose crystallites were rectangular (Bi et al. 2014). The shape of the cellulose crystallites has been speculated that it is affected by the arrangement of the cellulose synthesis enzymes at the BC producing bacteria cell wall membrane (Tsekos 1999). Furthermore, the crystallite size and crystallinity index of produced BC samples were calculated from the resulting XRD peaks, and given in Table 1. In general, bacterial cellulose has small crystallite sizes and high crystallinity, in which in this study they were in range of 4.7–6.8 nm and 55–81%, respectively. Interestingly, the difference of bacterial strains profoundly influenced on these BC structural characteristics. Under static condition, the K011 and K975 produced the largest crystallites about 6.0–6.8 nm, followed by the KX (5.7–6.6 nm) and K428 (5.1–5.3 nm), and the K086 gave the smallest crystallites of 4.7–5.0 nm. The BC produced by K1011, K975, and KX showed higher crystallinities than BC produced by K428 and K086, with values of up to 80, 81, 81, 64, and 60%, respectively. These results suggested that the three strains of K. xylinus; KX, K1011, and K975 had better ability in BC production and crystallization, leading to the higher crystallinity of BC microfibrils than those from the other two strains; K428 and K086.

Fig. 4
figure 4

X-ray diffraction patterns with three peaks, obtained from BC samples synthesized in a static and b agitated culture conditions

Full size image
Table 1 Crystallite size and crystallinity index determined by XRD, and cellulose Iα content determined by FT-IR measurements, of BC produced from five bacterial strains with different carbon sources in static and agitated conditions
Full size table

In addition, the types of cellulose producing organisms have effects on the ratio of cellulose Iα and Iβ produced. The FT-IR spectroscopy was performed to determine the exact values of mass fractions of these two allomorphs. As shown in Fig. 5, the regions of the FT-IR spectra peaks at 750 and 710 cm-1 were assigned to Iα and Iβ fractions, respectively. The cellulose Iα contents (%) of all BC samples were shown in Table 1. The results demonstrated that the variation in cellulose Iα contents of the BC produced by five different bacterial strains in this study were small, with the values ranged from 73 to 82%. It confirmed that the BC is composed of high cellulose Iα contents, whereas plant cellulose is rich in cellulose Iβ (Atalla and Vanderhart 1984).

Fig. 5
figure 5

FT-IR spectra with Iα mass fraction determined from peaks at 750 and 710 cm−1, as indicated, of bacterial cellulose from a static and b agitated culture conditions

Full size image

Effect of carbon sources

For K. xylinus bacteria, the change in the carbon sources has also been shown to affect the characteristics of BC produced (Klemm et al. 2005). From the Table 1, the crystallite size and Iα content data did not differ greatly between carbon sources among each bacterial strain; oppositely, the crystallinity of the cellulose produced in the different carbon sources was apparently different. Especially in the case of the KX strain, the highest crystallinity BC (81%) was obtained from the BC production with the glucose-containing medium, while the other sugar media provided the crystallinity of cellulose about 60–69%. This could be explained by the KX can efficiently use glucose as a carbon source to produce and crystallize BC microfibril more perfectly, resulting in the higher crystallinity (Benziman et al. 1980).

Effect of growth conditions

For the purposes of comparison of the microstructural changes in cellulose samples from both different culture conditions and estimation of the influence of the shaking condition on disturbance in the crystallization process, the XRD technique was determined. Analysis of the reflections corresponding to all three peaks in those XRD profiles in Fig. 4b revealed that the peaks of the cellulose samples from agitated culture were shifted to wider angles. Additionally, when comparing of 2θ angle values, the (100) and (010) reflections of the agitated BC samples were positioned closer together than in the cellulose profile from the static BC. These d-spacing changes appeared to represent that the proportion of Iα and Iβ cellulose allomorphs was altered, as reported previously (Czaja et al. 2004; Watanabe et al. 1998). Likewise, from the calculation based on the FT-IR spectra, the cellulose Iα contents of agitated BC were lesser to the value range of 67 to 77%. Interestingly, the change in cellulose Iα contents of the K428 was very small, the values decreased from 73–80 to 70–76%. This result agreed with the previous study which reported that the cellulose Iα content in the flocky asterisk-like BC was higher than the one found in the solid sphere-like BC (Bi et al. 2014). This finding suggested that the crystallization of cellulose microfibril structures of produced BCs could be affected by their macrostructure morphology.

Moreover, the percent crystallinity of cellulose grown under agitated culture was lower than the cellulose synthesized under static condition, suggesting that the agitation could interfere the aggregation of BC microfibrils causing the lower crystallinity. The crystallinities of BC produced by the K975, K1011, K428, and K086 decreased to the range of 61–66%, 53–74%, 46–55%, and 48–53%, respectively. The considerable reduction of crystallinity was found in the BC produced by the KX, the values decreased from 60–81 to 41–52% which could confirm that the agitation caused the mutation of the KX strain resulting in the loss of BC production ability.

The results also demonstrated the existence of smaller crystallite sizes in cellulose from agitated culture. The stress occurred during agitation seemed to interrupt greatly in the process of nascent microfibrils crystallization; therefore, it was more favorable to form smaller size microfibrils and increase the cellulose Iβ contents (more stable allomorph) in the BC production under stressful conditions. This hypothesis reconciled with previous reports (Ruka et al. 2012; Yamamoto et al. 1996).

Conclusion

In this study, the structural characteristics of BC produced by five strains of K. xylinus available in Thailand with six categories of carbon sources (glucose, fructose, lactose, maltitol, sucralose, and xylitol) in static and agitated culture conditions were investigated. Some conclusions were as follows:

  • The yield of production and morphology of BC were mainly affected by types of bacterial producing strains and methods of fermentation. Whereas the difference of carbon sources had small effects on the BC yield depending on each K. xylinus strains.

  • Static fermentation method produced bacterial cellulose in the pellicle form, while agitated fermentation method generated fragmented-cellulose particles with predominantly spherical shape. Agitation caused woven cellulose microfibrils become looser and form larger porosity.

  • One of K. xylinus strains used was the KX provided by the Institute of Food Research and Product Development, Kasetsart University, Thailand which has been known to produce high yield of cellulose and popularly used in food industry in Thailand. This KX was able to highly biosynthesize BC from all carbon sources in the static condition but not the agitated condition.

  • The K975 had comparable ability of BC production as the KX, and was capable of producing BC in the agitated condition.

  • The K1011 was suitable for BC production in the agitated condition since its BC yield dramatically increased when incubated in the agitated culture comparing to the static one.

  • The K428 and K086 although had low BC yield in the static condition, their BC particles produced in the agitated condition were fascinating because of their small-size morphology which can provide large surface area. Especially, the asterisk-like BC particles made by the K428 would give extra-large surface area that is good for absorption ability.