SEARCH

SEARCH BY CITATION

Keywords:

  • bacterial cellulose;
  • Gluconacetobacter xylinus;
  • Hestrin–Schramm medium;
  • NMR spectroscopy;
  • scanning electron microscopy

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Aims:  To determine the effect of carbon sources on cellulose produced by Gluconacetobacter xylinus strain ATCC 53524, and to characterize the purity and structural features of the cellulose produced.

Methods and Results:  Modified Hestrin Schramm medium containing the carbon sources mannitol, glucose, glycerol, fructose, sucrose or galactose were inoculated with Ga. xylinus strain ATCC 53524. Plate counts indicated that all carbon sources supported growth of the strain. Sucrose and glycerol gave the highest cellulose yields of 3·83 and 3·75 g l−1 respectively after 96 h fermentation, primarily due to a surge in cellulose production in the last 12 h. Mannitol, fructose or glucose resulted in consistent rates of cellulose production and yields of >2·5 g l−1. Solid state 13C CP/MAS NMR revealed that irrespective of the carbon source, the cellulose produced by ATCC 53524 was pure and highly crystalline. Scanning electron micrographs illustrated the densely packed network of cellulose fibres within the pellicles and that the different carbon sources did not markedly alter the micro-architecture of the resulting cellulose pellicles.

Conclusions:  The production rate of bacterial cellulose by Ga. xylinus (ATCC 53524) was influenced by different carbon sources, but the product formed was indistinguishable in molecular and microscopic features.

Significance and Impact of the Study:  Our studies for the first time examined the influence of different carbon sources on the rate of cellulose production by Ga. xylinus ATCC 53524, and the molecular and microscopic features of the cellulose produced.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Gluconacetobacter xylinus (previously named Acetobacter xylinum and Acetobacter xylinus) is a Gram-negative, obligately aerobic rod-shaped bacterium belonging to the family Acetobacteraceae (Yamada et al. 1997, 1998; Kersters et al. 2006). A notable feature of this micro-organism is its ability to produce cellulose, with a high degree of crystallinity, thus distinguishing it from plant cellulose (Cannon and Anderson 1991; Ross et al. 1991). As such, extensive research has been conducted using Ga. xylinus as a model organism for basic and applied studies on the biochemistry and genetics of cellulose formation, with significant contributions to the elucidation of the mechanisms of biogenesis of cellulose in plants (for reviews, see Cannon and Anderson 1991; Ross et al. 1991; Brown and Saxena 2000; Römling 2002; Brown 2004). More recently, Ga. xylinus has been used in a novel in vitro construction approach to understand the molecular assembly, architecture and mechanical behaviour of the primary cell wall of higher plants. Specifically, Ga. xylinus strain ATCC 53524 was used to create composites of cellulose/xyloglucan (Whitney et al. 1995, 1999, 2000, 2006; Chanliaud et al. 2004), cellulose/pectin (Chanliaud and Gidley 1999; Chanliaud et al. 2002) and cellulose/mannan-based polysaccharides (Whitney et al. 1998). These composite networks were simple two component systems, with defined molecular features and without the complexity and heterogeneity of plant tissues. Given the success of these studies, the potential for this bacterial model to be used in future investigations of plant cell wall architecture is promising. As such, it would be advantageous to have greater control over cellulose production by Ga. xylinus (ATCC 53524). This would require an understanding of this strain’s ability to metabolize different carbon sources and the impact this has on cellulose production. In this study, we report the effect of six different carbon sources on the production of cellulose by Ga. xylinus (ATCC 53524). The purity, molecular organization and micro-architectural properties of the cellulose produced were characterized.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Bacterial strain and culture conditions

Gluconacetobacter xylinus (formerly Acetobacter xylinus) strain ATCC 53524 from the American Type Culture Collection (Manassas, VA, USA) was used in this study. The bacterial strain was cultivated in Hestrin and Schramm (HS) medium (Hestrin and Schramm 1954) or in the modified media by simply replacing glucose with glycerol, mannitol, fructose, sucrose or galactose. In addition to the 20 g l−1 carbon source, the medium contained 5 g l−1 peptone, 5 g l−1 yeast extract, 2·7 g l 1 Na2HPO4 and 1·15 g l−1 citric acid. The starting pH of all media used in this study was 5·0.

Primary inocula were prepared by transferring a single colony from the HS medium working culture plate into 10 ml of each of the six modified HS media. Incubations were performed at 30°C for 48 h under static conditions. After incubation, broths were shaken vigorously to (partially) release attached cells from the cellulose pellicles. The resulting cell suspensions were used as inocula in subsequent experiments.

Experimental approach

Cellulose yield from Ga. xylinus (ATCC 53524) cultivated in the traditional HS medium was compared with modified HS medium containing each of the five other carbon sources. Cellulose pellicles were harvested every 12 h over a 96-h experimental period. At each time point, the pH of the medium was monitored. Total viable counts were recorded at the start and end of the experimental period. Cellulose pellicles harvested at 48 h from the six different media were analysed by nuclear magnetic resonance (NMR) spectroscopy and scanning electron microscopy to examine the purity and structural (molecular and micro-architectural) properties of the cellulose produced.

pH and total viable counts

The pH of the medium was recorded every 12 h over the 96 h experimental period using a TPS basic benchtop pH meter (TPS Pty Ltd, Springwood, Australia). The number of viable bacteria in the inoculum at the start and at the end of fermentation was determined by the spread plate technique. Cell suspensions used for plating were obtained by vigorously shaking the broths to (partially) dislodge the cells embedded in the cellulose pellicles. The diluent used was 0·1% peptone (pH 5·0) and the plating medium was HS agar containing the same carbon source as the sample. Colonies were counted after 4 days of incubation at 30°C using an Omron H7EC digital colony counter (Applethorn, Australia).

Consumption of carbon sources

Consumption of glucose, glycerol and galactose carbon sources was measured at 48 and 96 h, using the commercially available d-Glucose HK, Lactose/Galactose and Glycerol assay kits from Megazyme (Bray, Ireland), following the manufacturer’s instructions.

Bacterial cellulose yield

Cellulose pellicles were harvested from the broth media, rinsed with distilled water to remove excess media, and then immediately boiled (at 90°C) in 0·1 mol l−1 NaOH solution for 30 min to inactivate attached bacterial cells. After boiling, the pellicles were purified by extensive washing in distilled water at room temperature until the pH of the water became neutral. The purified cellulose was freeze dried overnight (10°C, 0·006 mbar) and dry weight was recorded for each pellicle at room temperature.

NMR spectroscopy

Cross-polarization/magic angle spinning (CP/MAS) 13C-NMR spectroscopy can be used to provide information on the ratio of crystalline and noncrystalline material in cellulose, and the nature of the crystalline material. Naturally occurring crystalline cellulose exists in two forms, cellulose Iα and Iβ which are distinguished by 13C-NMR (Earl and VanderHart 1981). Solid state 13C CP/MAS NMR experiments were performed at a 13C frequency of 75·46 MHz on a Bruker MSL-300 spectrometer. Approximately 50 mg of cellulose was packed in a 4-mm diameter, cylindrical, PSZ (partially stabilized zirconium oxide) rotor with a KelF end cap. The sample was placed in the centre of the rotor and the remaining space filled with KBr. The rotor was spun at 5–6 kHz at the magic angle (54·7°). The 90° pulse width was 5 μs and a contact time of 1 ms was used for all samples with a recycle delay of 3 s. The spectral width was 38 kHz, acquisition time 50 ms, time domain points 2k, transform size 4k with no line broadening. At least 2400 scans were accumulated for each spectrum. The spectra were referenced to external adamantane. In the case of the small amounts of bacterial cellulose produced from the galactose medium and the sucrose medium after 48 h, 20 000 scans were accumulated but the signal to noise ratio was still low.

Scanning electron microscopy

Cellulose samples (from 48 h fermentations) were freeze-substituted according to the method of Wharton (1991). Approximately 1 cm2 sample pieces were frozen in liquid nitrogen for approximately 10 s and immediately transferred to a solution of 3% glutaraldehyde in methanol at −20°C for 24 h. Cellulose samples were then transferred to methanol (100%, without the glutaraldehyde) at −20°C for a further 24 h, removed from the freezer, allowed to warm to room temperature, and dried using a Balzers critical point drier. Thereafter, the freeze-substituted samples were coated with approximately 10 nm of Pt using an Eiko IB-5 sputter coater and examined using a field emission Scanning Electron Microscope (JEOL JSM 6300F) at 6 kV and 3–5 mm working distance.

Statistical analysis

Results are presented as means of triplicate measurements with error bars representing standard errors. The significance of the effect each carbon source had on bacterial cellulose production was analysed using two-sample t-test on Minitab 15 for Windows® with significance based on a level of 5% (P < 0·05).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Effect of carbon sources on cell growth

To ensure growth conditions were optimum for Ga. xylinus (ATCC 53524) in each of the six modified HS media, primary inocula were cultivated in the respective modified HS media. This approach was taken to acclimatize the strain to the various carbon sources in the modified HS media, and thus minimize the metabolic shock the micro-organisms could experience if suddenly exposed to a different carbon source (other than glucose in the HS medium) during the scale-up phase of the experiment. A consequence of this approach was that during scale-up, the modified HS media were inoculated with primary inocula varying in cell numbers since cell density in the respective modified HS media varied. For this reason, direct comparisons between cell numbers of Ga. xylinus (ATCC 53524) in the different media could not be made, although magnitude of differences in colony counts at the start and end of the experiment for each medium could be compared. After 96 h of fermentation, all carbon sources tested in this study appeared to support growth of the strain and the production of cellulose (Table 1). In addition, the final pH of all fermentation media differed from the initial pH of 5·0, with modified HS media containing glucose, fructose, sucrose or galactose as carbon sources having elevated final pH values (Table 1) and those containing mannitol or glycerol as carbon sources having slightly lower final pH values (Table 1).

Table 1.   Effect of various carbon sources on the growth and cellulose production by Gluconacetobacter xylinus strain ATCC 53524
Carbon SourceIncrease in cell growth (log10 CFU ml−1)aCellulose yield at 48 h (g l−1)Cellulose yield at 96 h (g l−1) cFinal pHb
  1. aValues are the difference between cell growth at t = 0 h and t = 96 h. All values are presented as mean colony counts.

  2. bInitial pH = 5·0. Results are presented as means of triplicates.

Glucose1·051·893·105·33
Mannitol0·552·043·374·91
Glycerol2·630·823·754·81
Fructose1·101·792·815·09
Sucrose1·530·343·835·23
Galactose0·570·100·095·52

Effect of carbon sources on cellulose yield

The amount of cellulose produced on different carbon sources was recorded at 12 h intervals over the 96 h experimental period (Fig. 1). After 12 h, immeasurably small amounts of cellulose were produced, with cellulose yields increasing to measurable levels from 24 h onwards (Fig. 1). Cellulose production was significantly (P < 0·05) stimulated by all the carbon sources (2% w/v) at the end of fermentation (96 h), with the exception of galactose (Fig. 1). After 48 h, mannitol gave the highest cellulose yield (2·04 g l−1) with glucose and fructose also producing relatively high levels of cellulose, of 1·89 and 1·79 g l−1 respectively (Table 1). The amounts of cellulose produced with the other carbon sources were relatively low (Table 1). After 96 h of fermentation, the highest level of cellulose production was obtained using sucrose (3·83 g l−1). Glycerol, mannitol, glucose and fructose all gave good cellulose yields of 3·75, 3·37, 3·10 and 2·81 g l−1 respectively. Galactose once again appeared to be the least suitable carbon source for cellulose production (Table 1).

image

Figure 1.  Bacterial cellulose production by Ga. xylinus strain ATCC 53524 during the course of fermentation using different carbon sources: glucose (◆), mannitol ( bsl00001 ), glycerol (bsl00066), fructose ( × ), Sucrose (•) and galactose (○).

Download figure to PowerPoint

The consumption of glucose, glycerol (high cellulose yields) and galactose (lowest cellulose yield) in culture media was measured at the half-way point (48 h) and at the end (96 h) of fermentation (Table 2). At 48 h, glucose and glycerol were consumed at a similar rate (13·5% and 12·6% carbon source consumed respectively), giving similar cellulose yields per gram of carbon source consumed by 48 h (Table 2). At this time point only 1% of galactose was consumed to produce 0·45 g cellulose per gram of carbon source consumed. After 96 h of fermentation, glucose and glycerol had final consumption percentages of 51·5% and 57% respectively, with similar cellulose production per gram of carbon source consumed (0·3 g). No change was observed for the residual galacotse concentration. Therefore, it was evident that Ga. xylinus strain ATCC 53524 had the ability to utilize a range of carbon sources to produce cellulose. The choice of the most productive carbon source for cellulose production is dependent on the time course of the experiment to be conducted (Fig. 1).

Table 2.   Consumption of select carbon sources (20 g l−1) and cellulose production efficiency by Gluconacetobacter xylinus strain ATCC 53524
Carbon source Consumption at 48 h (%)Cellulose produced per gram of carbon source consumed by 48 h (g) Consumption at 96 h (%) Cellulose produced per gram of carbon source consumed by 96 h (g)
Glucose13·50·6651·50·30
Glycerol12·60·5257·00·32
Galactose1·00·451·00·45

Cellulose purity and structure

NMR spectroscopy was used to determine the purity and molecular properties of the cellulose produced by Ga. xylinus ATCC 53524. All samples were pure as judged by the observation of only cellulose signals in the 13C-NMR spectrum (Fig. 2). By studying the relative intensities of the C4 signals at 84–86 ppm (noncrystalline) and 89–91 ppm (crystalline), it was possible to make a quantitative estimate of the crystallinity of the cellulose samples (VanderHart and Atalla 1984) (Fig. 2). By peak fitting the crystalline region of the C4 signals of the samples, it was possible to determine Ιαβ ratios (Yamamoto and Horii 1993) (Fig. 3). The Ια/Ιβ ratios were the same for all samples within experimental error (Table 3). All samples exhibited a similar level of crystallinity, although the error in these calculations is high owing to the low signal to noise ratio in the 84–86 ppm region (Fig. 2).

image

Figure 2. 13C CP/MAS spectrum of bacterial cellulose samples from different carbon sources and fermentation times.

Download figure to PowerPoint

image

Figure 3.  Peak fitting analysis of C4 region from the CP/MAS 13C-NMR spectrum of bacterial cellulose incubated in fructose HS medium for 48 h. 13C CP/MAS Spectrum (----) and Peak fitting (- - - -).

Download figure to PowerPoint

Table 3.   Percentage crystallinity and Ια/Ιβ ratios from 13C-NMR
Media and incubation time (pH 5·0)% Crystallinity (±5)%Ια (±2)Ιβ (±2)
Fructose 48 h856139
Glucose 48 h806238
Glucose 72 h806040
Glycerol 48 h806040
Glycerol 72 h806040
Mannitol 48 h 906535

The micro-architecture of cellulose pellicles was investigated by SEM. Although multiple SEM micrographs were taken for the pellicles from each modified HS medium, only one representative micrograph is presented here (Fig. 4). Micrographs obtained for these samples revealed a densely packed network of cellulose fibres with few subtle differences between each sample (Fig. 4). The cellulose fibres within pellicles appeared to be random in orientation at the micron length scale, although glycerol and fructose pellicles gave the impression of some local microfibril directionality (Fig. 4c,d). The majority of cellulose fibres appeared to be slightly broader in the glycerol sample (Fig. 4c) and thinner in the galactose sample (Fig. 4f) compared to the other samples. In the sucrose pellicle, cellulose fibres appeared to aggregate together, forming local web-like structures within the network (Fig. 4e). Though this morphology was consistently observed only in sucrose pellicle samples, it is possible that these structures could be artefacts from sample processing prior to SEM. Overall, it was evident from the micrographs that the carbon sources tested in this study did not markedly alter the major microfibril structures within the cellulose pellicles.

image

Figure 4.  Scanning electron micrographs illustrating the micro-architecture of cellulose micro-fibrils produced by Ga. xylinus strain ATCC 53524 when cultivated for 48 h on various carbon sources: glucose (a), mannitol (b), glycerol (c), fructose (d), sucrose (e) and galactose (f).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Effects of carbon sources on rate of cellulose production

In recent years, various carbon sources including monosaccharides, oligosaccharides, alcohols, sugar alcohols and organic acids, have been used to maximize bacterial cellulose production by various Ga. xylinus strains (Masaoka et al. 1993, 1996; Yang et al. 1998; Ishihara et al. 2002; Keshk and Sameshima 2005). In the present study, we focused on Ga. xylinus strain ATCC 53524, demonstrating its ability to metabolize a range of carbon sources, and highlighting the effect of these substrates on cellulose production.

Efficient cellulose production by this bacterium lies in its ability to synthesize glucose from various carbon substrates, followed by glucose polymerization to cellulose. Ga. xylinus has two main operative amphibolic pathways: the pentose phosphate cycle for the oxidation of carbohydrates and the Krebs cycle for the oxidation of organic acids and related compounds (Ross et al. 1991). Over the 96 h fermentation period, the strain readily utilized mannitol, glucose and fructose for cellulose production (Fig. 1). Mannitol is known to be converted to fructose, and then metabolized by this organism to produce cellulose, while glucose and fructose are transported through the cell membrane and incorporated into the cellulose biosynthetic pathway (Ross et al. 1991; Oikawa et al. 1995, 1997). To date, only one other publication has investigated cellulose production by ATCC 53524 using glucose as carbon source (Ishihara et al. 2002). A cellulose yield of 2·61 g l−1 was reported (Ishihara et al. 2002), while our study observed a slightly higher yield of 3·10 g l−1. This difference could be due to the pH of the medium used in each study. Ishihara et al. (2002) used HS medium at a pH range of 5·7–6·5, while this study used HS medium of pH 5·0, the optimum pH recommended by ATCC for this strain.

For glycerol to be utilized by Ga. xylinus (ATCC 53524) for cellulose synthesis, the carbon of glycerol is introduced into the two metabolic cycles at the triose phosphate level (Weinhouse and Benziman 1976; Ross et al. 1991). The oxidation of triose phosphate is a primary reaction in this organism for the channelling of sugar carbon from the pentose cycle into the Krebs cycle (Weinhouse and Benziman 1976; Ross et al. 1991). This could explain the slight lag in the rate of cellulose production during the first 48 h of fermentation where glycerol was the sole carbon source (Fig. 1).

When sucrose was utilized as the carbon source, the second lowest yield for cellulose production was observed in the first 84 h of fermentation (Fig. 1). We suggest that this was due to the fact that sucrose could not be transported through the cell membrane and needed to be hydrolyzed in the periplasm to glucose and fructose (Velasco-Bedran and Lopez-Isunza 2007). Once this was achieved, only then could cellulose production commence, hence the late surge in cellulose production in the last 12 h of the 96 h fermentation experiment (Fig. 1).

Metabolism of galactose to cellulose did not occur very efficiently (Fig. 1) due to inefficient uptake from the medium (Table 2). It appears that Ga. xylinus (ATCC 53524) is unable to transport galactose efficiently across the cell membrane. The small amounts of galactose that were taken up were, however, apparently sufficient to maintain cell growth and essential cellular functions.

Preferred carbon source for Ga. xylinus ATCC 53524

It is evident from this study that depending on the time course of an experiment, Ga. xylinus ATCC 53524 could be grown on different carbon sources for cellulose production. We suggest the use of mannitol as the preferred carbon source for fermentation experiments to be conducted up to 72 h, although glucose and fructose produced nearly the same yield. If fermentation experiments are to be conducted over 96 h, the use of sucrose or glycerol for enhanced yields is suggested. Only limited studies of carbon sources other than glucose for Ga. xylinus strains have been reported. Until this investigation, only two studies have reported the ability of Ga. xylinus strains to metabolize mannitol. Oikawa et al. (1995) observed that the KU-1 strain, despite having a higher production rate of cellulose on d-arabitol (12·4 mg ml−1), could metabolize mannitol to give a bacterial cellulose yield of 4·6 mg ml−1, while Nguyen et al. (2008) found that Ga. xylinus strain K3 preferred mannitol as a carbon source.

Effects of carbon sources on cellulose microstructure

Although the rate and extent of production of cellulose was characteristic for the carbon source utilized, the microscopic and molecular organization of cellulose produced was highly conserved. This suggests that differences in productivity due to carbon source were due to substrate limitations, rather than any alteration in polymerization processes.

Gluconacetobacter xylinus ATCC 53524 is a preferred strain for the production of both pure cellulose and composites with plant polysaccharides such as xyloglucan (Whitney et al. 1995) and pectin (Chanliaud and Gidley 1999) that are characteristic of the primary cell wall of plants. This study has shown that different carbon sources lead to characteristic cellulose production profiles, but that there is no major difference in the nature of cellulose produced irrespective of the rate or extent of synthesis. This finding is useful for further development of models for the plant cell wall in two ways. Firstly, it shows how to design studies to separate effects of the rate and extent of cellulose production on composite formation and properties. Secondly it shows that isotopically labelled substrates based on, e.g. glycerol can be used efficiently to produce labelled cellulose and composites for detailed structural characterization.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The authors would like to thank Brigid McKenna for her technical assistance and Ismail Mohamed Al-Bulushi for his help with the statistical analysis. This research was supported by a Discovery Grant from the Australian Research Council (ARC) to M.J.G. and G.A.D.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References