Ethanol adaptation in a thermotolerant yeast strain Kluyveromyces marxianus IMB3

The maximum ethanol concentration produced from glucose in defined media at 45°C by the thermotolerant yeast Kluyveromyces marxianus IMB3 was 44 g L⁻¹. Acclimatisation of the strain through continuous culture at ethanol concentrations up to 80 g L⁻¹, shifted the maximum ethanol concentration at which growth was observed from 40 g L⁻¹ to 70 g L⁻¹. Four isolates were selected from the continuous culture, only one of which produced a significant increase in final ethanol concentration (50 ± 0.4 g L⁻¹), however in subsequent fermentations, following storage on nutrient agar plates, the maximum ethanol concentration was comparable with the original isolate. The maximum specific ethanol production rates (approximately 1.5 g (gh)⁻¹) were also comparable with the original strain except for one isolate (0.7 g (gh)⁻¹). The specific ethanol productivity decreased with ethanol concentration; this decrease correlated linearly (rval 0.92) with cell viability. Due to the transience of induced ethanol tolerance in the strain it was concluded that this was not a valid method for improving final ethanol concentrations or production rates. Received 18 July 1997/ Accepted in revised form 19 February 1998     

Production of xylitol by Candida peltata

Candida peltata NRRL Y-6888 to ferment xylose to xylitol was evaluated under different fermentation conditions such as pH, temperature, aeration, substrate concentration and in the presence of glucose, arabinose, ethanol, methanol and organic acids. Maximum xylitol yield of 0.56 g g⁻¹ xylose was obtained when the yeast was cultivated at pH 6.0, 28°C and 200 rpm on 50 g L⁻¹ xylose. The yeast produced ethanol (0.41 g g⁻¹ in 40 h) from glucose (50 g L⁻¹) and arabitol (0.55 g g⁻¹ in 87 h) from arabinose (50 g L⁻¹). It preferentially utilized glucose > xylose > arabinose from mixed substrates. Glucose (10 g L⁻¹), ethanol (7.5 g L⁻¹) and acetate (5 g L⁻¹) inhibited xylitol production by 61, 84 and 68%, respectively. Arabinose (10 g L⁻¹) had no inhibitory effect on xylitol production. Received 24 December 1998/ Accepted in revised form 18 March 1999     

Analysis and optimization of a two-substrate fermentation for xylitol production using Candida tropicalis

Xylitol, a functional sweetener, was produced from xylose using Candida tropicalisATCC 13803. A two-substrate fermentation was designed in order to increase xylitol yield and volumetric productivity. Glucose was used initially for cell growth followed by conversion of xylose to xylitol without cell growth and by-product formation after complete depletion of glucose. High glucose concentrations increased volumetric productivity by reducing conversion time due to high cell mass, but also led to production of ethanol, which, in turn, inhibited cell growth and xylitol production. Computer simulation was undertaken to optimize an initial glucose concentration using kinetic equations describing rates of cell growth and xylose bioconversion as a function of ethanol concentration. Kinetic constants involved in the equations were estimated from the experimental results. Glucose at 32 g L⁻¹ was estimated to be an optimum initial glucose concentration with a final xylose concentration of 86 g L⁻¹ and a volumetric productivity of 5.15 g-xylitol L⁻¹ h⁻¹. The two-substrate fermentation was performed under optimum conditions to verify the computer simulation results. The experimental results were in good agreement with the predicted values of simulation with a xylitol yield of 0.81 g-xylitol g-xylose⁻¹ and a volumetric productivity of 5.06 g-xylitol L⁻¹ h⁻¹. Received 16 June 1998/ Accepted in revised form 28 February 1999     

Characterisation of a novel thermotolerant yeast, Kluyveromyces marxianus var marxianus: development of an ethanol fermentation process

The fermentation characteristics of the novel, thermotolerant, isolate Kluyveromyces marxianus var marxianus were determined to evaluate its aptitude for use in an ethanol production process. Sustainable growth was not observed under anaerobic conditions, even in the presence of unsaturated fatty acid and sterol. A maximum ethanol concentration of 40 g L⁻¹ was produced at 45°C, with an initial specific ethanol production rate of 1.7 g g⁻¹ h⁻¹. This was observed at ethanol concentrations below 8 g L⁻¹ and under oxygen-limited conditions. The low ethanol tolerance and low growth under oxygen-limited conditions required for ethanol production implied that a simple continuous process was not feasible with this yeast strain. Improved productivity was achieved through recycling biomass into the fermenter, indicating that utilising an effective cell retention method such as cell recycle or immobilisation, could lead to the development of a viable industrial process using this novel yeast strain. Received 14 February 1998/ Accepted in revised form 19 May 1998     

Increased astaxanthin production by a Phaffia rhodozyma mutant grown on date juice from Yucca fillifera

The wild strain and the astaxanthin-overproducing mutant strain 25–2 of Phaffia rhodozyma were analyzed in order to assess their ability to grow and synthesize astaxanthin in a minimal medium composed of g L⁻¹: KH₂PO₄ 2.0; MgSO₄ 0.5; CaCl₂ 0.1; urea 1.0 and supplemented with date juice of Yucca fillifera as a carbon source (yuca medium). The highest astaxanthin production (6170 μg L⁻¹) was obtained at 22.5 g L⁻¹ of reducing sugars. The addition of yeast extract to the yuca medium at concentrations of 0.5–3.0 g L⁻¹ inhibited astaxanthin synthesis. The yuca medium supported a higher production of astaxanthin, 2.5-fold more than that observed in the YM medium. Journal of Industrial Microbiology & Biotechnology (2000) 24, 187–190. Received 14 July 1999/ Accepted in revised form 02 December 1999     

Production of Cold-Adapted Amylase by Marine Bacterium <Emphasis Type="Italic">Wangia</Emphasis> sp. C52: Optimization,Modeling, and Partial Characterization

The aim of this work was to optimize the fermentation parameters in the shake-flask culture of marine bacterium Wangia sp. C52 to increase cold-adapted amylase production using two statistical experimental methods including Plackett–Burman design, which was applied to find the key ingredients for the best medium composition, and response surface methodology, which was used to determine the optimal concentrations of these components. The results showed starch, tryptone, and initial pH had significant effects on the cold-adapted amylase production. A central composite design was then employed to further optimize these three factors. The experimental results indicated that the optimized composition of medium was 6.38 g L⁻¹ starch, 33.84 g L⁻¹ tryptone, 3.00 g L⁻¹ yeast extract, 30 g L⁻¹ NaCl, 0.60 g L⁻¹ MgSO₄ and 0.56 g L⁻¹ CaCl₂. The optimized cultivation conditions for amylase production were pH 7.18, a temperature of 20°C, and a shaking speed of 180 rpm. Under the proposed optimized conditions, the amylase experimental yield (676.63 U mL⁻¹) closely matched the yield (685.60 U mL⁻¹) predicted by the statistical model. The optimization of the medium contributed to tenfold higher amylase production than that of the control in shake-flask experiments.     

Bioreduction of Cu(II) by Cell-Free Copper Reductase from a Copper Resistant <Emphasis Type="Italic">Pseudomonas</Emphasis> sp. NA

Environmental copper contamination is a serious human health problem. Copper reductase is produced by microorganisms to facilitate copper uptake by ATPases into the cells increasing copper biosorption. This study assessed the reduction of Cu(II) by cell-free extracts of a highly copper-resistant bacterium, Pseudomonas sp. strain NA, isolated from vineyard soil contaminated with copper. Both intact cells and cell-free extract of Pseudomonas sp. strain NA displayed substantial reduction of Cu(II). Intact cells reduced more then 80 mg L⁻¹ of Cu(II) from medium amended with 200 mg L⁻¹ of copper after 24 h of incubation. Cell-free extract of the isolate reduced more than 65% of the Cu(II) at initial copper concentration of 200 mg L⁻¹ after 24 h. Soluble protein production was high at 72 h of incubation at 100 mg L⁻¹ of copper, with more then 60 μg L⁻¹ of total soluble protein in cell-free extract recorded. Cu(II) reduction by isolate NA was increased when copper concentration increased for both intact cells and cell-free extract. Results indicate that Pseudomonas sp. strain NA produces copper reductase enzyme as the key mechanism of copper biotransformation.     

Kinetic models for phycocyanin production by high cell density mixotrophic culture of the microalga Spirulina platensis

Phycocyanin production by high cell density cultivation of Spirulina platensis in batch and fed-batch modes in 3.7-L bioreactors with a programmed stepwise increase in light intensity program was investigated. The results showed that the cell density in fed-batch culture (10.2 g L⁻¹) was 4.29-fold that in batch culture (2.38 g L⁻¹), and the total phycocyanin production in the fed-batch culture (0.795 g L⁻¹) was 3.05-fold that in the batch culture (0.261 g L⁻¹). An unstructured kinetic model to describe the microalga culture system including cell growth, phycocyanin formation, as well as glucose consumption was proposed. The data fitted the models well (r ² > 0.99). Furthermore, based on the kinetic models, the potential effects of light limitation and photoinhibition on cell growth and phycocyanin formation can be examined in depth. The models demonstrated that the optimal light intensity for mixotrophic growth of Spirulina platensis in batch or fed-batch cultures using a 3.7-L bioreactor was 80160 μE m⁻² s⁻¹, and the stepwise increase in light intensity can be replaced by a constant light intensity mode. Received 28 July 1998/ Accepted in revised form 8 October 1998     

Kinetic models for astaxanthin production by high cell density mixotrophic culture of the microalga Haematococcus pluvialis

High cell density cultivation of Haematococcus pluvialis for astaxanthin production was carried out in batch and fed-batch modes in 3.7-L bioreactors with stepwise increased light intensity control mode. A high cell density of 2.65 g L⁻¹ (batch culture) or 2.74 g L⁻¹ (fed-batch culture) was obtained, and total astaxanthin production in the fed-batch culture (64.36 mg L⁻¹) was about 20.5% higher than in the batch culture (53.43 mg L⁻¹). An unstructured kinetic model to describe the microalga culture system including cell growth, astaxanthin formation, as well as sodium acetate consumption was proposed. Good agreement was found between the model predictions and experimental data. The models demonstrated that the optimal light intensity for mixotrophic growth of H. pluvialis in batch or fed-batch cultures in a 3.7-L bioreactor was 90–360 μmol m⁻² s⁻¹, and that the stepwise increased light intensity mode could be replaced by a constant light intensity mode. Received 24 December 1998/ Accepted in revised form 23 April 1999     

Fed-batch production of recombinant human calcitonin precursor fusion protein using Staphylococcus carnosus as an expression-secretion system

A pH-auxostatic fed-batch process was developed for the secretory production of a fusion protein consisting of the pro-part of Staphylococcus hyicus lipase and two synthetic human calcitonin (hCT) precursor repeats under the control of a xylose-inducible promotor from Staphylococcus xylosus. Using glycerol as the energy source and pH-controlled addition of yeast extract resulted in the production of 2000 mg l⁻¹ of the fusion protein (420 mg l⁻¹ of the recombinant hCT precursor) within 14 h, reaching 45 g l⁻¹ cell dry mass with Staphylococcus carnosus in a stirred-tank reactor. Product titer and space-time yield (30 mg calcitonin precursor l⁻¹ h⁻¹) were thus improved by a factor of 2, and 4.5, respectively, compared to Escherichia coli expression-secretion systems for the production of calcitonin precursors. Two hundred grams of the fusion protein was secreted by the recombinant S. carnosus on a 150-l scale (scale-up factor of 50) with a minimum use of technical-grade yeast extract (40 mg fusion protein g⁻¹ yeast extract). Received: 18 January 2000 / Received revision: 14 April 2000 / Accepted: 14 April 2000     

The effect of complex carbon and nitrogen, salt, Tween-80 and acetate on delta-endotoxin production by a Bacillus thuringiensis subsp kurstaki

Delta-endotoxin production by a strain of Bacillus thuringiensis subsp kurstakion complex media based on crude gruel and fish meal was investigated. High proteolytic activities were concomitantly produced with the bioinsecticide. In such complex media, the repressive regulation due to readily consumed carbon sources was partially overcome. In order to improve substrate assimilation, 0.5 g L⁻¹ sodium chloride and 0.1% Tween-80 were supplemented to the production medium, increasing delta-endotoxin yields when using gruel concentrations below 59 g L⁻¹. At and beyond 75 g L⁻¹ gruel, delta-endotoxin yields were not affected in the presence of 0.5 g L⁻¹ NaCl and 0.1% Tween-80, but proteolytic activity yields were remarkably reduced. Thus, the use of sodium chloride and Tween-80 allowed reduction of the initial gruel concentration to 42 g L⁻¹ for the production of 3350 mg L⁻¹ delta-endotoxin, while it was only 3800 mg L⁻¹ with 92 g L⁻¹ gruel. Moreover, similar to 0.5 g L⁻¹ NaCl and 0.1% Tween-80, the use of 10 g L⁻¹ sodium acetate significantly improved delta-endotoxin production and also reduced the proteolytic activity to 250 U ml⁻¹. Received 05 November 1998/ Accepted in revised form 19 August 1999     

Development of a cost-effective glucose-corn steep medium for production of butanol by Clostridium beijerinckii

Corn steep water (CSW) medium (1.6% solids plus 6% glucose) was evaluated for growth and butanol production by Clostridium beijerinckii NCIMB 8052 wild-type and hyper-amylolytic, hyper-butanol-producing mutant strain BA101. CSW alone was not a suitable substrate, whereas addition of glucose supported growth and butanol production by both strains. In a batch-scale fermentation using an optimized 6% glucose-1.6% solids CSW medium, C. beijerinckii NCIMB 8052 and strain BA101 produced 10.7 g L⁻¹ and 14.5 g L⁻¹ of butanol, respectively. The total solvents (acetone, butanol, and ethanol) produced by C. beijerinckii NCIMB 8052 and strain BA101 were 14 g L⁻¹ and 20 g L⁻¹, respectively. Initial fermentation in small-scale flasks containing 6% maltodextrin-1.6% solids concentration CSW medium resulted in 6 g L⁻¹ and 12.6 g L⁻¹ of butanol production by C. beijerinckii NCIMB 8052 and strain BA101, respectively. CSW can serve as an economic source of nitrogen, vitamins, amino acids, minerals, and other nutrients. Thus, it is feasible to use 6% glucose-1.6% solids CSW medium in place of semi-defined P2 medium. Received 9 February 1998/ Accepted in revised form 1 September 1998     

Evaluation of sugar cane hemicellulose hydrolyzate for cultivation of yeasts and filamentous fungi

Sugar cane bagasse hemicellulosic fraction submitted to hydrolytic treatment with 100 mg of sulfuric acid per gram of dry mass, at 140°C for 20 min, was employed as a substrate for microbial protein production. Among the 22 species of microorganisms evaluated, Candida tropicalis IZ 1824 showed TRS consumption rate of 89.8%, net cell mass of 11.8 g L⁻¹ and yield coefficient (Y_x/s) of 0.50 g g⁻¹. The hydrolyzate supplemented with rice bran (20.0 g L⁻¹), P₂O₅ (2.0 g L⁻¹) and urea (2.0 g L⁻¹) provided a TRS consumption rate of 86.3% and a cell mass of 8.4 g L⁻¹. At pH 4.0 cellular metabolism was inhibited, whereas at pH 6.0 the highest yield was obtained. The presence of furfural (2.0 g L⁻¹) hydroxymethylfurfural (0.08 g L⁻¹) and acetic acid (3.7 g L⁻¹) in the hydrolyzate did not interfere with cultivation at pH 6.0. Received 25 October 1996/ Accepted in revised form 10 March 1997     

An optimization study of carotenoid production by Rhodotorula glutinis DBVPG 3853 from substrates containing concentrated rectified grape must as the sole carbohydrate source

A multivariate statistical approach was employed for the optimization of conditions for carotenoid production by Rhodotorula glutinis DBVPG 3853 from a substrate containing concentrated rectified grape must as the sole carbohydrate source. Several experimental parameters (carbohydrate, yeast autolysate and salt concentrations, and pH) were tested at two levels by following a fractional factorial design. Carotenogenesis was most sensitive to both initial pH and yeast autolysate concentration. A Central Composite Design experiment was then performed by obtaining both second-order polynomial models and isoresponse diagrams where initial pH and yeast autolysate concentration were considered as variables. In this way it was possible to determine the conditions (pH = 5.78, yeast autolysate = 4.67 g L⁻¹) which maximize both the concentration of total carotenoids and that of β-carotene (6.9 mg L⁻¹ and 1100 μg L⁻¹ of culture fluid, respectively, after 120 h of fermentation). Journal of Industrial Microbiology & Biotechnology (2000) 24, 41–45. Received 23 February 1999/ Accepted in revised form 14 September 1999     

Mineralization of pentachlorophenol in a contaminated soil by Pseudomonas sp UG30 cells encapsulated in κ-carrageenan

A fermentation process in Escherichia coli for production of supercoiled plasmid DNA for use as a DNA vaccine was developed using an automated feed-back control nutrient feeding strategy based on dissolved oxygen (DO) and pH. The process was further automated through a computer-aided data processing system to regulate the cell growth rate by controlling interactively both the nutrient feed rate and agitation speed based on DO. The process increased the total yield of the plasmid DNA by approximately 10-fold as compared to a manual fed-batch culture. The final cell yield from the automated process reached 60 g L⁻¹ of dry cell weight (OD₆₀₀ = 120) within 24 h. A plasmid DNA yield of 100 mg L⁻¹ (1.7 mg g⁻¹ cell weight) was achieved by using an alkaline cell lysis method. Plasmid yield was confirmed using High Performance Liquid Chromatography (HPLC) analysis. Because cells had been grown under carbon-limiting conditions in the automated process, acetic acid production was minimal (below 0.01 g L⁻¹) throughout the fed-batch stage. In contrast, in the manual process, an acid accumulation rate as high as 0.36 g L⁻¹ was observed, presumably due to the high nutrient feed rates used to maintain a maximum growth rate. The manual fed-batch process produced a low cell density averaging 10–12 g L⁻¹ (OD₆₀₀ = 25–30) and plasmid yields of 5–8 mg L⁻¹ (approximately 0.7 mg g⁻¹ cells). The improved plasmid DNA yields in the DO- and pH-based feed-back controlled process were assumed to be a result of a combination of increased cell density, reduced growth rate (μ) from 0.69 h⁻¹ to 0.13 h⁻¹ and the carbon/nitrogen limitation in the fed-batch stage. The DO- and pH-based feed-back control, fed-batch process has proven itself to be advantageous in regulating cell growth rate to achieve both high cell density and plasmid yield without having to use pure oxygen. The process was reproducible in triplicate fermentations at both 7-L and 80-L scales. Received 22 March 1996/ Accepted in revised form 20 September 1996     

Growth-associated production of poly(hydroxybutyric acid) by Azotobacter beijerinckii from organic nitrogen substrates

Poly(hydroxybutyric acid) (PHB) was produced by a selectant of Azotobacter beijerinckii in media containing only organic nitrogen sources such as N substrates. The chosen compounds were casein peptone, yeast extract, casamino acids and urea, each combined with carbon substrates glucose or sucrose. The PHB was synthesized under growth-associated conditions. The concentrations amounted to more than 50% of cell dry mass on casein peptone/glucose as well as urea/glucose medium within 45 h fermentation time. Corresponding to these yields, productivities of about 0.8 g PHB l⁻¹ h⁻¹ were discovered. The highest values increased to 1.06 g PHB l⁻¹ h⁻¹ on casein peptone/glucose medium and 1.1 g PHB l⁻¹ h⁻¹ on yeast extract/glucose medium after a period of 20 h. It was found that oxygen limitation was essential for successful product formation, as demonstrated earlier. These data from basic research may support further investigations into the use of technical proteins from renewable sources as substrates for PHB production by a strain of A. beijerinckii. Received: 3 June 1997 / Received revision: 29 August 1997 / Accepted: 15 September 1997     

Semi-defined Medium for in vitro Cultivation of the Fastidious Insect Pathogenic Fungus Cordyceps unilateralis

Cordyceps unilateralis is a fastidious fungal pathogen affecting ants. Up to now, only the complex and expensive Grace’s insect cell culture medium has been used for in vitro cultivation (as blastospores and mycelium) of this fungus. To obtain an inexpensive and less complicated medium, the effects of carbon and nitrogen sources, salt solution and carbon-to-nitrogen (C:N) ratio on the growth of this fungus were examined. Glucose was the most important factor for blastospore formation, and yeast extract could be used as a nitrogen source for blastospore formation and mycelial growth. A suitable C:N ratio (glucose: yeast extract) was 33.3:1. As a result, a new semi-defined medium was achieved, composed of 26.68 g L⁻¹ glucose, 3.3 g L⁻¹ yeast extract and salt solution. This medium supported blastospore formation and mycelial growth of all tested C. unilateralis isolates.     

In vitro cultivation of three <Emphasis Type="Italic">Kappaphycus alvarezii</Emphasis> (Rhodophyta,Areschougiaceae) variants (green,red and brown) exposed to a commercial extract of the brown alga <Emphasis Type="Italic">Ascophyllum nodosum</Emphasis> (Fucaceae,Ochrophyta)

Diseases such as “ice-ice” and “goose bumps” are responsible for loss of almost 70% of infected Kappaphycus production. To improve the growth of K. alvarezii and reduce the impact of diseases, the effect of the commercial Ascophyllum nodosum extract was tested. Five grams of apical branches of each K. alvarezii variant (n = 6) were bathed at different extract concentrations (e.g., 5, 10, 15, 20, 25, and 30 g L⁻¹) for 1 hour. To test the effect of the extract on epiphytes apical tips were not cleaned. After 2 weeks, the epiphytes such as Cladophora sp. and Ulva sp. disappeared. Polysiphonia subtilissima survived but its appearance was weak. The highest K. alvarezii growth rates were observed at extract concentrations of 15 and 20 g L⁻¹ (p < 0.001), almost double the expected for a 45-day production cycle used at Sepetiba Bay, Brazil. These results show that the extract was efficient at improving the growth of K. alvarezii and reducing the presence of some epiphytes.     

Use of Acadian marine plant extract powder from Ascophyllum nodosum in tissue culture of Kappaphycus varieties

Three varieties of Kappaphycus alvarezii (Kapilaran, KAP), Tambalang purple (PUR), Adik-adik (AA), and one variety of Kappaphycus striatum var. sacol (green sacol (GS) were used to determine the efficiency of Acadian marine plant extract powder (AMPEP) as a culture medium at different concentrations, for the regeneration of young plants of Kappaphycus varieties, using tissue culture techniques for the production of seed stock for nursery and outplanting purposes for the commercial cultivation of carrageenophytes. A shorter duration for shoot formation was observed when the explant was treated with AMPEP + Plant Growth Regulator (PGR = PAA + zeatin at 1 mg L⁻¹) compared to AMPEP when used singly. However, four explants responded differently to the number of days required for shoot formation. The KAP variety took 46 days to form shoots at 3–4 mg L⁻¹ AMPEP + PGR; while PUR required 21 days at 3–5 mg L⁻¹ AMPEP and 3–4 mg L⁻¹ AMPEP + PGR. AA required 17 days at 3–5 mg L⁻¹ AMPEP and AMPEP + PGR; and GS 25 days at 1 mg L⁻¹ AMPEP + PGR. It was observed that among the four explants used, PUR and AA initiated shoot formation with the use of AMPEP only at higher concentrations (3–5 mg L⁻¹) after a shorter period. Only PUR responded positively to ESS/2 for shoot initiation. The use of AMPEP alone and/or in combination with PGR as a culture medium in the propagation of microplantlets using tissue culture technique is highly encouraging.     

Simultaneous utilization of glucose and D-xylose by Candida shehatae in a chemostat

The kinetics of biomass formation, D-xylose utilization, and mixed substrate utilization were determined in a chemostat using the yeast Candida shehatae. The maximum growth rate of C. shehatae grown aerobically on D-xylose was 0.42 h⁻¹ and the Monod constant, K _s, was 0.06 g L⁻¹. The biomass yield, Y _{X/S}, ranged from 0.40 to 0.50 g g⁻¹ over a dilution rate range of 0.2–0.3 h⁻¹, when C. shehatae was grown on pure D-xylose. Mixtures of D-xylose and glucose (∼1 : 1) were simultaneously utilized over a dilution rate from 0.15 to 0.35 h⁻¹ at pH 3.5 and 4.5, but pH 3.5 reduced μ_max and reduced the dilution rate range over which D-xylose was utilized in the presence of glucose. At pH 4.5, μ_max was not reduced with the mixed sugar feed and the overall or lumped K _s value was not significantly increased (0.058 g L⁻¹ vs 0.06 g L⁻¹), when compared to a pure D-xylose feed. Kinetic data indicate that C. shehatae is an excellent candidate for chemostat production of value added products from renewable carbon sources, since simultaneous mixed substrate utilization was observed over a wide range of growth rates on a 1 : 1 mixture of glucose and D-xylose. Received 21 August 1997/ Accepted in revised form 28 May 1998