Appl Microbiol Biotechnol (1990) 33:148-153

/lppli Microbiology Biotechnology © Springer-Verlag 1990

The use of free and immobilised Arthrobacter simplex in organic solvent/aqueous two-liquid-phase reactors M . D. H o c k n u l l and M . D . Lilly .

SERC Centre for Biochemical Engineering, Department of Chemical and Biochemical Engineering, UniversityCollege London, Torrington Place, London, WC1E 7JE, United Kingdom Received 7 September 1989/Accepted 8 December 1989 Summary. The use of free and immobilised Arthrobacter

simplex (NCIB 8929) for steroid Aa-dehydrogenation in two-liquid-phase, stirred-tank reactors has been compared. Product formation is related to the logarithm of the water-octanol partition coefficient (log P) of the organic solvent employed, but the relationship is different for the two forms of the biocatalyst. No reaction was seen with either biocatalyst in media containing solvents of log P 95% purity) steroid product was demonstrated.

Introduction

The Al-dehydrogenation of hydrocortisone by Arthrobacter simplex is being used as a model system to study the use of metabolically active microorganisms in organic solvent/aqueous two-liquid phase systems. At its simplest, the Al-dehydrogenation reaction may be considered as a two-component reaction. The steroid substrate is oxidised by a single enzyme, generating reduced flavin adenine dinucleotide (FADH). For continued reaction, the oxidised form of the cofactor (FAD +) must be regenerated. This regeneration proceeds via the electron transport chain of the organism (Medentsev et al. 1985). Although cofactor regeneration can be achieved by the use of artificial electron acceptors such as phenazine methosulphate (PMS) or menadione (Silbiger and Freeman 1988), or by the use of coupled enzyme reactions (Wong 1987), these are often expensive

Offprint requests to: M. D. Lilly

or technically difficult to achieve, thus adding greatly to process costs. The use of metabolically active microorganisms as biocatalysts enables the exploitation of the organism's ability to regenerate necessary cofactors. The use of water-immiscible organic solvents offers several advantages in biocatalysis (Lilly 1982) - for example the possibility of operating with high concentrations of poorly-water-soluble reactants and products in the reactor whilst avoiding inhibition of the reaction by either reactant or product. There are however a number of disadvantages, including the need for adequate mass transfer between the two-liquid phases and biocatalyst inactivation by the organic phase. In a recent report (Hocknull and Lilly 1988) the causes of Al-dehydrogenation activity loss in A. simplex exposed to essentially water-immiscible organic solvents were investigated. In particular, the roles of the organic solvent dissolved in the aqueous phase and contact with the liquid-liquid interface were distinguished. For hydrocortisone oxidation by A. simplex in twoliquid-phase stirred tanks, two broad classes of organic solvent were distinguished. Firstly, there are those organic solvents that are not toxic to the Gram-positive bacteria when dissolved in an aqueous buffer, but which cause a partial loss of Al-dehydrogenation activity when direct contact between the liquid-liquid interface and the biocatalyst is permitted. The second class of organic solvents also affect the stability of the A 1dehydrogenation system of A. simplex when dissolved in the aqueous buffer, the rate of loss correlating with the logarithm of the water/octanol partition coefficient (log P) of the organic solvent. For A. simplex, solvents in the first group include all those organic solvents with a log P > 3.6 with the exception of the alkanols which, together with solvents of log P < 3.0, constitute the second group of organic solvents. Contact with the liquid-liquid interface reduces the measured activity half-life of free A. simplex in twoliquid-phase stirred tanks (Hocknull and Lilly 1988). Prevention of such contact may lead to enhanced sta-

149 b i l i t y a n d h e n c e r e a c t o r p e r f o r m a n c e . This p a p e r c o m p a r e s t h e u s e o f free a n d i m m o b i l i s e d m e t a b o l i c a l l y active A. simplex in t w o - l i q u i d - p h a s e reactors.

M a t e r i a l s and m e t h o d s

Chemicals. Hydrocortisone was kindly donated by Diosynth BV, Oss, The Netherlands. Prednisolone and phenazine methosulphate (PMS) were obtained from Sigma Chemical Company, London, UK. Organic solvents were supplied by BDH, Poole, Dorset, UK, and were of analytical grade unless otherwise stated.

Cultivation of bacteria. A. simplex NCIB 8929 was grown and stored as a frozen cell paste as described previously (Hocknull and Lilly 1988). It is known that thawed and fresh A. simplex behave in a similar manner in two-liquid-phase systems (Hocknull and Lilly 1988; Hocknull 1989).

Immobilisation of A. simplex in calcium alginate. A thawed bacterial suspension in 50 mM TRIS-HC1 buffer (pH 7.8) was mixed at room temperature with three volume parts of 4% (w/v) sodium alginate. After mixing, the bacteria-alginate mixture was dispersed in a 0.2 M CaC12 solution in 50 mM TRIS-HC1 buffer (pH 7.8) by pumping the suspension through a hypodermic needle. The beads were allowed to harden for 2 h at room temperature and then used immediately. Beads of 3 mm mean diameter were obtained.

Steroid Al-dehydrogenation by free and immobilised bacteria. Reactions were done in 70-ml working volume, stirred-tank reactors. Hydrocortisone (35 mg) dissolved in 35 ml organic solvent was combined with 20 ml TRIS-HC1 buffer solution (pH 7.8) and the two liquids left to equilibrate under operational conditions (agitation 750 rpm, 30°C) for 10 min. After this time, either free bacteria (70 mg wet wt) or 15 g calcium alginate beads containing 1.6 g dry weight bacteria/kg were added to the reactor. The aqueous (including immobilised cells)/organic volumetric phase ratio was 1. Samples (0.2 ml) were removed periodically for steroid analysis. Prednisolone concentrations are expressed per litre of reaction mixture. For aqueous reactions in the absence of organic solvent, the solvent was replaced by an equal volume of buffer solution.

presence of PMS was used to assay steroid A~-dehydrogenation activity. Analyses of hydrocortisone and prednisolone concentrations in samples were done by HPLC (Hocknull and Lilly 1988).

Results

Steroid A 1_dehydrogenation by free and immobilised bacteria Figure 1 shows the At-dehydrogenation of hydrocortis o n e b y free a n d i m m o b i l i s e d A. simplex in t h e a b s e n c e o f o r g a n i c solvent. T h e s p e c i f i c activity o f t h e free b a c t e r i a a n d t h e e x p r e s s e d activity o f t h e i m m o b i l i s e d b a c t e r i a w e r e i n i t i a l l y 10.0 a n d 6.7 m g / g d r y wt p e r m i n u t e , r e s p e c t i v e l y . T h e effectiveness f a c t o r o f t h e i m m o b i l ised b i o c a t a l y s t (the r a t i o o f i m m o b i l i s e d to free b a c t e ria activity) w a s t h e r e f o r e 0.67. W h e n t h e b a c t e r i a w e r e r e l e a s e d b y s o l u b i l i s i n g t h e a l g i n a t e gel, t h e i r a c t i v i t y was t h e s a m e as b e f o r e i m m o b i l i s a t i o n . T h e A l - d e h y d r o g e n a t i o n b y free a n d i m m o b i l i s e d b a c t e r i a in t h r e e o r g a n i c s o l v e n t / a q u e o u s t w o - l i q u i d p h a s e systems is s h o w n in Fig. 2. It h a s b e e n p r o p o s e d t h a t " a c t i v i t y r e t e n t i o n " in t w o - l i q u i d - p h a s e r e a c t o r s is r e l a t e d to t h e l o g a r i t h m o f t h e w a t e r - o c t a n o l p a r t i t i o n c o e f f i c i e n t (log P) o f t h e o r g a n i c s o l v e n t e m p l o y e d ( L a a n e et al. 1985). "[he r e l a t i o n s h i p b e t w e e n p r o d u c t f o r m e d in 2 h b y free a n d i m m o b i l i s e d b i o c a t a l y s t in two-liquid-phase reactions and the log P of the organic s o l v e n t u s e d f o r a r a n g e o f s o l v e n t s is s h o w n in Fig. 3. T h e results are e x p r e s s e d as p e r c e n t a g e s o f t h e v a l u e s o b t a i n e d with w h o l l y a q u e o u s r e a c t i o n m e d i a . T h e relat i o n s h i p is m a r k e d l y d i f f e r e n t f o r the t w o f o r m s o f t h e b i o c a t a l y s t . W i t h b o t h free a n d i m m o b i l i s e d b a c t e r i a , n o r e a c t i o n was o b s e r v e d in t h e p r e s e n c e o f s o l v e n t s o f

Repeated batch transformations of high soluble concentrations of hydrocortisone. Dodecanol (55 ml) containing 1.67 g/1 dissolved hydrocortisone was combined with 15 g calcium aiginate beads containing 1.6 g dry weight bacteria/kg in a 70-ml stirred-tank reactor. The reaction was allowed to proceed for 24 h under operating conditions of 350 rpm, 30 ° C. Samples (0.2 ml) were withdrawn periodically for steroid analysis. After 24 h, the dodecanol was drawn off and replaced with 55 ml fresh dodecanol containing hydrocortisone.

Stability ofimmobilised bacteria. Exposure of immobilised bacteria to organic solvents was done in a 1-1 stirred-tank reactor (LH Fermentation, Stoke Poges, UK). TRIS-HC1 buffer, pH 7.8, (425 ml) was added to 175 ml organic solvent in the reactor and left to equilibrate under operational conditions (30 ° C, 500 rpm) for 0.5 h. Calcium alginate beads (100 g) containing 4.5 g dry weight bacteria/kg were then added to the reactor. Periodically, 5 g beads were removed from the reactor and dissolved in 30 ml of 50 mM potassium phosphate buffer, containing 50 mM TRIS-sodium citrate, for 15 min. The resulting cell suspension was used to assay the specific activity of the bacteria as described by Hocknull and Lilly (1988).

Specific activity assay and steroid analys&. The assay of steroid A~-dehydrogenation activity was done using a method described previously (Hocknull and Lilly 1988). A similar assay but in the

4~1~

o

t-

I//o ~

6

!

I

1.

:~ :9 ~,~.me {l~)

f

1

!

4

~

Fig. 1. A ~-Dehydrogenation reaction by free ( A ) and immobilised

(D) Arthrobacter simplex in an aqueous reaction mixture

150 g.4

1~.3

.,.

~

.~ ,~ lal v

E

L

~

.~

~.2

N ~

a

.~

.~

• 1

11~

~

~.~

@

1

2 bi~e

3

4

5

(h)

Fig. 2. A1-Dehydrogenation by free (closedsymbols) and immobilised (open symbols) A. simplex in two-liquid-phase reactors containing as organic solvent; - diisopentyl ether, ,t, A ; ethyl decanoate, V, V ; dioctylphthalate, m, []

log P < 2 . 5 . With immobilised bacteria, reactor productivity rose rapidly with increasing organic solvent log P until full activity retention was seen with organic sol-

'

B.5

;

1~

'.

15

'

~6

ini~i~l hydrooo.~one (9/I) Fig. 4. The effect of initialhydrocoaisone concentration on the initial reaction rate in dodecanol-aqueous immobilised cell reactors

vents o f log P > 4 . However, with free bacteria as the biocatalyst, reactor productivity rose linearly with solvent log P over the range 2.5-9.8.

Influence of reactor hydrocortisone concentration on initial reaction rate

125

1~ o L

0 D o D U" 0 q0

5~ N ~.

-g "~l

[

2

~

~

|

4

8

8

i

1~1

lo~ P Fig. 3. Product formed in 2 h (expressed as a percentage of the aqueous control) by free 03) and immobilised (A) A. simplex in two-liquid-phase stirred tank reactors as a function of the logarithm of the water-octanol partition coefficient (10g P) value of the organic solvent used; ~,, immobilized cells with dodecanol saturated with steroid

The results in Fig. 3 were for experiments in which the initial a m o u n t of hydrocortisone was constant and, with the exception of dodecanol, the organic solvents were saturated with steroid and some solid steroid remained. In the presence of dodecanol, a lower a m o u n t of product formation was observed than would have been predicted f r o m the log P value of this solvent. This is due to the higher solubility of the steroid substrate and product in dodecanol c o m p a r e d to other organic solvents. The non-saturation of the organic phase has the effect o f reducing the concentration of steroid in the aqueous phase and hence in the alginate beads, resulting in a reduced reaction rate and therefore product formation over a given time. Figure 4 shows the rate of A~-dehydrogenation, measured over the first 10 min of reaction, by immobilised A. simplex in d o d e c a n o l / a q u e o u s two-liquid-phase reactors, with varying initial hydrocortisone concentrations. The higher activity obtained at and above 1 g/1 hydrocortisone is plotted in Fig. 3 (filled symbol) and now fails on the general curve.

The stability of immobilised bacteria in two-liquid phase stirred tanks In the absence of organic solvent and steroid substrate, immobilised A. simplex lost half its steroid A ~-dehydro-

151 ~

10

~., •~

E

!:

5

~0

,

~

i-

e

~-~ ~ ~

"

~

~

~

N

N .~

~

"~ •

~

A

~ ~

~

•~ •~

4

4~'~

~

a



~

4

N

~ ~ N 2

~ ~

O

~

O

• t~

~

~

~

~ .~

o

2

1

-~ •

~

~

~

I

I

I

I

I

I

~

5

iI~

15

~

25

e

~ime (h) Fig. 5. Maintenance stability of immobilised bacteria in two-

liquid-phase stirred tanks. Residual Al-dehydrogenation stability in dodecane (A) and decanol (A), ALdehydrogenase activity in dodecane ( v ) and decanol (V)

genation activity in 24h. During the experiment R(PMS), the ratio of activity in the presence and absence of PMS, remained constant. Figure 5 shows the maintenance stability (i.e. the stability in the absence of steroid substrate) in twoliquid-phase stirred tanks. In the presence of dodecane, which is known to be non-toxic when present in dissolved form only (Hocknull and Lilly 1988), the activity half-life (time after which half the activity had been lost) of the bacteria was 24 h. The R(PMS), however, did not remain constant. There was an initial rise in R(PMS), denoting a partial loss of dehydrogenation activity relative to the remaining dehydrogenase activity. For free-cell, two-liquid-phase systems such a phenomenon is due to direct contact between the biocatalyst and the liquid-liquid interface. It is likely that some limited contact between the bacteria within the immobilisation matrix and the liquid-liquid interface still occurs at the surface of the beads, accounting for the partial loss of dehydrogenation activity seen in the dodecane/immobilised-cell, stirred-tank reactors. The stability of the immobilised bacteria in the presence of an organic solvent, decanol, known to be toxic to the bacteria in dissolved form is also shown in Fig. 5. Steroid Al-dehydrogenation activity half-life in the presence of a decanol second liquid phase was 1.3 h while that of the ALdehydrogenase enzyme itself was 5.0 h. These values were lower than the values obtained previously for free bacteria (Hocknull and Lilly 1988) in decanol-saturated buffers and it is likely that some contact between the liquid-liquid interface and the bacteria in the beads was taking place.

5

10

15

~

25

~

85

~

~5

58

~i~e0~)

~l

Fig. 6. ALDehydrogenatiton of hydrocortisong by immobilised bacteria in a dodecanol/buffer two-liquid-phase reactor at an initial concentration of 5 g / l ; hydrocortisone, [] ; prednisolone, A

Two-liquid-phase batch reactions In addition to the reactor productivity, the initial steroid concentration in the reactor also affected the percentage of substrate initially present in the reactor converted to product. At low concentrations (0.1-2g/1 reactor liquid) conversions of greater, than 95% were possible. At higher concentrations the solubility in the organic solvent was exceeded and solid steroid was present in the reactor. This resulted in a lower percentage conversion at the end of the reaction. Figure 6 shows the conversion by immobilised bacteria of 5 g/1 hydrocortisone in a dodecanol/aqueous two-liquid-phase reactor. The bacteria recovered at the end of the experiment were still active, their specific activity being 6 mg/g dry wt per minute. Thus if the presence of solid steroid can be avoided in a two-liquidphase reactor, high conversions by metabolically active microorganisms are potentially possible. To examine this, four sequential batch transformations each lasting 24h were carried out. Hydrocortisone (1.31 g/1 total reaction volume) was added at the start of each reaction and a conversion of 95% was achieved in the first batch and over 90% in the three subsequent batches, in which the reaction rates were very similar. At the end of the experiment the specific activity of bacteria recovered from the reactor was 5 mg/g dry wt per minute (i.e. 50% of the initial activity still remained).

Discussion

For the steroid A 1-dehydrogenation of hydrocortisone by free A. simplex, which involves cofactor regenera-

152 tion, there is no real advantage in using a two-liquidphase reaction other than to ease product recovery. In the presence of all but solvents of very high log P (i.e. those of log P > 9) product formation is lower than that of a comparable aqueous reactor (Fig. 3). The best twoliquid-phase reactor, containing di-n-octylphthalate, was comparable to an aqueous reactor in terms of product formation rate. However this solvent has a number of undesirable properties including high viscosity and poor steroid capacity, making it unsuitable for use in a two-liquid-phase process. The low productivity in other two-liquid-phase reactors can be overcome by the use of artificial electron acceptors (Hocknull and Lilly 1987). The use of such compounds may be attractive even though they add to the cost of a potential process. Recently we have shown that the free cell biocatalyst rapidly loses a third of its initial activity in a stirred-tank reactor due to direct contact between the biocatalyst and the liquid-liquid interface (Hocknull and Lilly 1988) suggesting that, if such contact can be avoided, a significant improvement in biocatalyst stability and hence reactor productivity might be expected. This is most readily achieved by immobilisation of the biocatalyst by entrapment. Many reasons for immobilising biocatalysts in conventional (all aqueous) reactors have been cited including ease of reuse, enhanced product separation, a high biocatalyst concentration in the reactor and improved biocatalyst stability. In two-liquid-phase biocatalysis of steroids, however, product separation is readily achieved since the product partitions into the organic phase while the biocatalyst remains in the aqueous phase. However, there are still good reasons for immobilising the biocatalyst in two-liquid-phase reactors including enhanced biocatalyst stability (Omata et al. 1979, 1980) and the possibility of continuous operation. Enhanced stability will probably be the most beneficial effect of immobilisation of the biocatalyst in such reactors. Yamane et al. (1979), for example, noted a large increase in the stability of the steroid dehydrogenation activity of Nocardia rhodochrous on immobilisation. In most cases reported so far it is not usually clear whether this increase in stability results from the immobilisation procedure per se or from greater protection from the organic phase. The results of this study help to distinguish these two possibilities. The activity half-life of the immobilised bacteria in the absence of organic solvent was 24 h. Kloosterman and Lilly (1985) also reported an activity half-life of 24 h for immobilised .4. simplex in an aqueous reactor. This value however is higher than the value of 17 h obtained for free .4. simplex (Hocknull and Lilly 1988). As the activities of immobilised bacteria were measured on bacteria released by solubilisation of the matrix this difference cannot be a diffusional effect on immobilised activity and represents a real increase in stability. We have shown previously (Hocknull and Lilly 1988; Hocknull 1989) that water-immiscible organic solvents can be divided into two categories based upon their toxicity towards the Al-dehydrogenation system of non-immobilised A. simplex. Firstly there are those

solvents (e.g. alkanes) with log P > 3.6 but excluding the water-immiscible alkanols which are toxic only by virtue of direct contact between the liquid-liquid interface and the biocatalyst. Such contact results in a rapid, partial loss of activity depending on the degree of contact permitted. Secondly, there is a group of water-immiscible organic solvents with log P < 3.4 and the water-immiscible alkanols, up to dodecanol (log P = 5.1) that, in addition to being toxic via the liquid-liquid interface, are also toxic when present dissolved in the aqueous phase. This second toxic effect produces a slower (log P dependent) rate of activity loss. In this study the first of these groups of solvents is represented by dodecane, and the second by decanol. Figure 5 shows that the activity half-life of the immobilised bacteria in a dodecane/buffer, two-liquid-phase stirred tanks is equal to that of immobilised bacteria in an aqueous reactor. There is some loss of dehydrogenation activity relative to dehydrogenase activity. This is thought to result from some limited contact between bacteria near the surface of the calcium alginate bead and the liquid-liquid interface. In the presence of decanol, the activity half-life of the immobilised bacteria is 1.2 h. This value is similar to that obtained previously by us for free bacteria in decanol-saturated buffer. However, it is much larger than the value of 0.2 h obtained previously for free bacteria in an aqueous-decanol, two-liquid-phase stirred tank. Therefore, it is reasonable to conclude from these experiments that immobilisation by entrapment protects the biocatalyst by providing protection against contact with the liquidliquid interface. The results of the experiment with decanol suggest that no protection is provided against the organic solvent dissolved in the aqueous phase. This is perhaps to be expected since the immobilisation matrix employed, 4% calcium alginate, is hydrophilic so that the concentration of decanol in the beads would be similar to that in the liquid aqueous phase. The increase in stability on immobilisation from 0.2 h to 1.5 h represent a 7.5-fold increase. Previously we have shown (Kloosterman and Lilly 1985) that the stability of immobilised .4. simplex in aqueous reactions was greater when reactant was present and reaction taking place (operational stability) than in the absence of substrate (maintenance stability). A similar effect has now been observed in two-liquid-phase reactions. When repeated batches were done with the same immobilised bacteria about 50% of their activity remained after 96 h of use. During that time a total of 4.98 g/l hydrocortisone had been converted, equivalent to 14.5 g hydrocortisone converted/g dry weight of immobilised bacteria. As the second, third and fourth batches were essentially identical it seems likely that most .of the activity loss occurred during the first batch. These experiments demonstrate the feasibility of using immobilised A. simplex in a two-liquid-phase reaction to produce large quantities of prednisolone. The much greater stability observed during the batch reactions than when no steroid was present also highlights the importance of the metabolic state of microorganisms on their resistance to organic solvents.

153

Acknowledgements. The authors wish to thank the Science and Engineering Research Council for its support of this work and the award of a studentship to one of us (M.D.H.).

References Hocknull MD (1989) The influence of water immiscible organic solvents on bacterial steroid Al-dehydrogenation. PhD thesis, University of London Hocknull MD, Lilly MD (1987) The A~-dehydrogenation of hydrocortisone by Arthrobacter simplex in organic solvent/aqueous two-liquid phase environments. In: Laane C, Tramper J, Lilly MD (eds) Biocatalysis ir/organic media (Studies in Organic Chemistry 29). Elsevier, Amsterdam, pp 393-398 Hocknull MD, Lilly MD (1988) The stability of the A~-dehydrogenation system of Arthrobacter simplex in organic solvent/aqueous two-liquid phase environments. Enzyme Microb Technol 10:669-674 Kloosterman J, Lilly MD (1985) Maintenance and operational stability of immobilised Arthrobacter simplex for the Al-dehydrogenation of steroids. Enzyme Microb Technol 7:377-382 Laane C, Boeren S, Vos K (1985) On optimizing organic solvents in multi-liquid-phase biocatalysis. Trends Biotechnol 3:251-2

Lilly MD (1982) Two liquid phase biocatalytic reactors. J Chem Technol Biotechnol 32:162-169 Medentsev AG, Arinbasarova AY, Koscheyenko KA, Akimenko VK, Skryabin GK (1985) Regulation of 3 ketosteroid-l-en-dehydrogenase activity of Arthrobacter globiformis cells by a respiratory chain. J Steroid Biochem 23:365-368 Omata T, Lida T, Tanaka A, Fukui S (1979) Transformation of steroids by gel entrapped Nocardia rhodochrous cells in organic solvent. Eur J Appl Microbiol Biotechnol 8:143-155 Omata T, Tanaka A, Fukui S (1980) Bioconversions under hydrophobic conditions: effect of solvent polarity on steroid transformations by gel entrapped Nocardia rhodochrous cells. J Ferment Technol 58:339-343 Silbiger E, Freeman A (1988) Continuous non-aerated Al-dehydrogenation of hydrocortisone by PAAH-bead entrapped Arthrobacter simplex. Appl Microbiol Biotechnol 29:413-418 Wong C-H (1987) Nicotinamide cofactor-requiring enzymatic synthesis in organic solvent-water biphasic systems. In: Laane C, Tramper J, Lilly MD (eds) Biocatalysis in organic media (Studies in Organic Chemistry 29). Elsevier, Amsterdam, pp 197-208 Yamane T, Nakatani H, Sada E, Omata T, Tanaka A, Fukui S (1979) Steroid bioconversions in water insoluble organic solvents: Al-dehydrogenation by microbial cells and by cells entrapped in hydrophilic and lipophilic gels. Biotechnol Bioeng 21 : 1887-1903

aqueous two-liquid-phase reactors.

The use of free and immobilised Arthrobacter simplex (NCIB 8929) for steroid delta 1-dehydrogenation in two-liquid-phase, stirred-tank reactors has be...
566KB Sizes 0 Downloads 0 Views