- Email: [email protected]
Production of L-tryptophan from D,L-5-indolylmethylhydantoin by resting cells of a mutant of Arthrobacter species (DSM 3747)
Journal of Biotechnology, 14 (1990) 363-376
363
Elsevier BIOTEC 00523
Production of L-tryptophan from D,L-5-indolylmethylhydantoin by resting cells of a mutant of Arthrobacter species (DSM 3747) C h r i s t i a n e G r o s s , C h r i s t o p h Syldatk, V e r a M a c k o w i a k a n d F r i t z W a g n e r Institute of Biochemistry and BiotechnologyA, Technical University of Braunschweigo Braunschweig, F.R.G.
(Received 14 April 1989; accepted 7 January 1990)
Summary The reaction parameters and the stereospecificity of the enzymatic cleavage of D,L-5-indolylmethylhydantoin in producing L-tryptophan with resting cells of Arthrobacter sp. D S M 3747 were studied. When intact cells were tested, the optimal p H was between 8.5 and 9.0 and the optimal temperature was 5 0 ° C . Both, L-N-carbamoylase and hydantoinase could be stabilized over 24 h at 30 and 40 ° C by the addition of D,L-5-indolylmethylhydantoin. Furthermore, the hydantoinase was stable over 24 h at 50°C by the addition of 0.5 m M Mn 2+ ions. The treatment with sodium desoxycholate turned out to be successful in overcoming the poor availability of D,L-5-indolylmethylhydantoin for the cells. The optimal temperature with permeabilized cells decreased to 3 0 ° C and therefore ensured a good.enzyme stability. While the L-N-carbamoylase proved to be absolutely L-specific, the hydantoinase led to a mixture of enantiomers of N-carbamoyltryptophan. The produced D-N-carbamoyl-tryptophan caused an inhibition of the L-N-carbamoylase. The transformation yield from D,L-5-indolylmethylhydantoin always reached 100%. Arthrobacter sp.; L-Tryptophan; Transformation; Stereospecificity
Introduction Aromatic just as rare L-amino acids are gaining increasing importance for the supplementation of feedstuff and for pharmaceutical applications. One approach to
Correspondence to: Dr. C. Gross, Institute of Biochemistryand BiotechnologyA, Technical University of
Braunschweig, Konstantin-Uhde-Str. 5, D-3300, Braunschweig, F.R.G. 0168-1656/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
364
producing L-amino acids is the use of hydantoin-hydrolyzing enzymes. While the ability to produce D-amino acids from racemic 5-monosubstituted hydantoins is widespread among microorganisms (Syldatk et al., 1986), only a few species seem to be able to produce L-amino acids along this route. According to the research reported to date, productivities of the hydantoin-hydrolyzing enzymes from different bacterial sources vary considerably. The enzymes however seem to possess several common features. L-N-Carbamoylase represents an enzyme with a strict stereospecificity for aromatic L-N-carbamoylamino acids (Syldatk et al., 1987; Yokozeki et al., 1987a; Nishida et al., 1987), while stereospecificity of the hydantoinase depends on the C-5-substituent of the hydantoinring. For example, the L-enantiomer of N-carbamoyltryptophan is predominantly formed from the hydantoin substituted with the bulky indolylmethyl rest. Thus, for the production of L-amino acids a process consisting of one enzymatic (hydantoinase) and one chemical step (N-carbamoylamino acid hydrolysis) could not be adopted as it had been for the production of optically pure D-amino acids (Dinelli et al., 1978). Although in most cases it was possible to quantitatively convert the employed racemic 5-monosubstituted hydantoins, most procedures lead to an unsatisfactory yield of L-amino acids (Guivarch et al., 1980; Yokozeki et al., 1987b). We previously reported the optimization of growth conditions and enzyme formation by a mutant of Arthrobacter sp. DSM 3747 employing D,L-5-indolylmethylhydantoin as a model substrate (Syldatk et al., 1990). This paper describes the development of a microbial process for the L-tryptophan production from D,L-5-indolylmethylhydantoin, using resting cells of a mutant of Arthrobacter sp. DSM 3747. Aiming at optimization, the influence of the configuration of the products formed was taken into particular consideration.
Materials and Methods
Chemicals 5-Indolylmethylhydantoin (5-IMH) and N-carbamoyltryptophan (N-Carbtrp) were supplied by Professor Krohn, Institute of Organic Chemistry, T U Braunschweig. Yeast extract was obtained from Ohly G m b H (Hamburg, F.R.G.). All other chemicals were purchased from Sigma (Miinchen, F.R.G.), Merck (Darmstadt, F.R.G.) and Serva (Heidelberg, F.R.G.). Microorganism All experiments were performed with a mutant of Arthrobacter sp. DSM 3747 which was selected as described before (Gross et al., 1987a). Culture conditions Cultivation was carried out in 500-ml Erlenmeyer flasks with 200 ml medium on a rotation shaker at 100 rpm and 27°C for 24 h or in a 20-1 bioreactor (BraunMelsungen, Melsungen, F.R.G.). A complex medium was used for these cultivations (Cotoras and Wagner, 1984). It was further optimized by increasing the concentra-
365 tion of the inducer to 2 g 1-1 D,L-5-IMH. The biomass grown in the bioreactor was harvested after 10 h of growth by centrifugation at 16000 x g and 20°C for 30 min. At this time, maximal specific enzyme activity had been reached with 0.45 m M N-Carbtrp + L-Trp g-1 (cell dry mass = C D M ) h -1 for hydantoinase and 0.37 m M L-Trp g-1 h -~ for L-N-carbamoylase at 27°C.
Reaction conditions with resting cells Wet biomass obtained after centrifugation of the culture was applied directly to the biotransformation. The reaction was done either in 100-ml Erlenmeyer flasks with 10 ml buffer or in a 5-1 bioreactor of the type Biostat V (Braun-Melsungen, Melsungen, F.R.G.). p H adjustment in the 5-1 bioreactor was carried out with 10% N a O H . A concentration of 4.37 m M D , L - 5 - I M H could be solubilized in the reaction mixture, at higher concentrations it was added as a solid component. The data even of higher concentrations of D , L - 5 - I M H are given in m M to simplify the comparison of the product and substrate concentrations. All experiments were performed in a N 2 atmosphere to prevent the degradation of L-Trp via a pathway initiated by the oxygen-dependent tryptophan oxygenase (Gross et al., 1987b). The conditions of the experiments are given in the text. It could be shown that the growing cells were able to partly store the inducer D,L-5-IMH. Under reaction conditions the stored D , L - 5 - I M H was liberated by the cells as L-tryptophan (L-Trp), so in some cases the reaction yield exceeded 100%. Protein concentration In the case of detergent addition, the protein concentration in the supernatant of the reaction mixture was measured by the method of Lowry et al. (1951) before starting the reaction by adding the substrate. Bovine serum albumin was used as calibration standard. Quantitatioe determination of the reaction products The concentrations of the reaction products were measured in the supernatant of the reaction mixture after centrifugation at 3000 x g and 20°C for 15 min. This quantitative determination was done by H P L C as described before (Gross et al., 1987b). The optical purities were analyzed by H P L C (Spectra Physics, Hamburg, F.R.G.) using different chiral columns (Serva, Heidelberg, F.R.G.), and conditions for the separation of the D- and L-enantiomers. A Chiral Val Cu-column was used for the separation of D,L-5-IMH; eluent 10 m M CUSO4/5% acetonitril, flow rate 0.7 ml min-1, detection at 260 nm. D- and L-N-Carbtrp could be determined on a Chiral Pro Cu-column; eluent 1 m M CuSO 4, flow rate 1.3 ml rain-1, detection at 230 nm. Before measuring the enantiomer ratio of N-Carbtrp, it was usually necessary to remove L-Trp from the samples by ion-exchange chromatography on Amberlite IR-120 (H + form) because of interference with determination when it was present at high concentrations. The analysis of D- and L-Trp was performed on a Chiral Val Cu-column; eluent 1 m M CuSO4, flow rate 1 ml rain -1, detection at 280 nm.
366 Results and Discussion
Effect of pH The effect of pH is shown in Fig. 1. The highest yield of L-Trp from D,L-5-IMH was obtained between pH 8.5 and 9.0. The influences of the various buffers were
2,0
A
1.5-E
v
Ii .J ÷
1.0--
)
D.. f_
u i
Z
0.5
/ •
I
6
I
I
8
1.5 --
,
,
L
10
12
pH B
E1.o! 0.5~
I 6
l
I
8
I 10
12
pH Fig. 1. Effect of p H on product formation; conditions: 1.75 m M D , L - 5 - I M H , 4.2 g 1-~ C D M , 27 ° C, 3 h.
(A) Influence on the hydantoinase; (B) influence on the L-N-carbamoylase. e, Tris-maleate; ~, glycine-NaOH; A, Tris-HCl; m, diethanolamine-HC1.
367
2.0
_g
1.0
20
z,O
/0
Tempereture ('C)
Fig. 2. Effect of temperature on product formation. Conditions: 1.75 mM D,L-5-IMH, 0.1 M glycineNaOH, pH 8.5, 4.2 g 1-1 CDM, 3 h. ,, L-Trp; O, N-Carbtrp; II, L-Trp+ N-Carbtrp. quite different, especially Tris-HC1 and diethanolamine-HC1 lead to an unfavourable ratio of L-Trp to N-Carbtrp. Among several reaction solutions tested (Gross et al., 1987b), 0.1 M potassium phosphate buffer or glycine-NaOH gave the best results regarding a good conversion yield in favour of the production of L-Trp.
Effect of temperature After 3 h of incubation, the optimal temperature for the L-Trp formation was 50°C, as given in Fig. 2.
Effect of products The influence of NH~- and H C O f ions, L-N-Carbtrp and L-Trp on the biotransformation was investigated in order to prevent a possible inhibition. NH4CI, N a H C O 3 and NaC1, to detect possible effects of the counterions and the ionic strength, were added in concentrations of 1, 10 and 50 m M to the reaction mixtures. N o inhibition could be detected, but in all cases an increase in conversion of maximal 38% was observed, which led in most cases to a deterioration of the ratio of L - T r p / N - C a r b t r p . Adding either L-N-Carbtrp or L-Trp up to a concentration of 20 m M to the reaction mixture resulted in an increase in transformation after 24 h of incubation. F o r the 20 m M concentration of L-N-Carbtrp the yield was double that of the control containing no L-N-Carbtrp. Hydantoinase was activated especially beyond a L-N-Carbtrp concentration of 10 mM, while the increase in the L-Trp concentration could be caused either by activation or a shift in reaction equilibrium produced by the high substrate concentration for L-N-carbamoylase. The strong effect of L-NCarbtrp concentrations greater than 10 m M may be due to improved diffusion through the cell membrane which results in a sufficient intracellular concentration. For the complete range of L-Trp concentrations the yield was 30% higher than that
368 TABLE 1 Influence of temperature on the stability of enzymatic activities in the presence and absence of various additives Addition
Temp.
Fresh cells *
mM L-Trp
mM L-Trp + N-Carbtrp
1.76
1.95
None 0.5 mM MnSO4 0.5 mM ZnSO4 87.4 raM, D,L-5-IMH
30 o C
0.54 0.93 0.54 2.30
0.86 1.54 1.31 2.91
None 0.5 mM MnSO4 0.5 mM ZnSO4 87.4 mM D,L-5-IMH
40 o C
0.59 0.29 0.54 1.62
1.00 0.94 1.59 3.08
None 0.5 mM MnSO4 0.5 mM ZnSO4 87.4 mM D,L-5-IMH
50 o C
0.04 0.05 0.02 0.25
0.07 1.83 0.04 0.50
The remaining activity of resting cells, which were preincubated for 24 h at 30, 40 and 50 ° C, was compared to the activity of cells measured immediately after cultivation (*). Conditions: 4.37 mM D,L-5-IMH, 0.1 M glycine-NaOH, pH 8.5, 4.2 g 1-1 CDM, 27°C, 3 h.
of the control to which n o L-Trp had b e e n added. L-Trp stimulated the entire t r a n s f o r m a t i o n from D , L - 5 - I M H to L-Trp. The increased yield was correlated with a higher ratio of L - T r p / N - C a r b t r p .
Stability of hydantoinase and L-N-carbamoylase Cells were p r e i n c u b a t e d at different p H values at 2 7 ° C for 24 h. T h e best stability of e n z y m a t i c activities was observed at p H 8 - 1 0 with a residual activity of 50-60% a n d 4 0 - 5 0 % for h y d a n t o i n a s e a n d L-N-carbamoylase, respectively. T h e results of i n c u b a t i n g the cells at various temperatures for 24 h before m e a s u r i n g the r e m a i n i n g activity c a n be seen from T a b l e 1. W i t h o u t additions, b o t h enzymes t u r n e d out to be very labile even at 30 a n d 4 0 ° C with a drop in activity of 50% a n d 70% for h y d a n t o i n a s e a n d L-N-carbamoylase, respectively. A t 5 0 ° C b o t h enzymatic activities were negligibly low. By a d d i n g D , L - 5 - I M H at a c o n c e n t r a t i o n that could n o t be converted completely even at 5 0 ° C d u r i n g the p r e i n c u b a t i o n period, b o t h e n z y m a t i c activities could be restored at 30 a n d 40°C. T h e activation of h y d a n t o i n a s e at 30 a n d 4 0 ° C a n d of L - N - c a r b a m o y l a s e at 30°C m a y be explained by the p e r m e a b i l i z a t i o n of the cells d u r i n g the i n c u b a t i o n period, which yielded higher conversion rates i n the following enzyme test. I n contrast to L - N - c a r b a m o y lase h y d a n t o i n a s e could b e stabilized b y the a d d i t i o n of M n 2÷ ions up to 50°C a n d b y Z n 2÷ ions up to 40°C.
Permeabilization of the cells Since L - N - c a r b a m o y l a s e in particular could n o t be sufficiently stabilized for a longer period of time at higher temperatures, it was necessary to increase the
369 TABLE 2 Influence of several detergents on cell permeabilization. Conditions: 43.7 mM D,L-5-1MH, 0.1 mM glycine-NaOH, pH 8.5, 4.2 g 1-1 CDM, 27 o C, 24 h. Addition
g 1-1
mM L-Trp
mM L-Carbtrp
mM L-Trp + N-Carbtrp
None Triton X-100 Triton X-405 Tween 20 Tween 40 Tween 60 Span 80 NaDOC CTAB * 20 o C
10 10 10 10 10 10 10 10
8.28 6.66 7.20 2.50 3.53 4.70 4.46 17.05 8.18 10.99
0.65 4.41 0.53 0.20 0.28 0.61 0.65 6.20 9.59 1.17
8.93 11.07 7.73 2.70 3.81 5.31 5.11 23.25 17.74 12.16
-
* Cetyltrimethyiammonium bromide. reaction rate in order to shorten the duration of reaction. Thus a prolonged use of t h e cells w o u l d b e m a d e p o s s i b l e . A t f i r s t t h e e n z y m a t i c a c t i v i t i e s o f cells g r o w n i n t h e p r e s e n c e o f D , L - 5 - N - 3 m e t h y l h y d a n t o i n ( S y l d a t k e t al., 1990) w e r e c o m p a r e d t o t h o s e i n u s u a l l y g r o w n
TABLE 3 Dependence of cell permeabilization and transformation yield on the length of incubation with NaDOC Time (h)
Protein CDM -1 (%) (supernatant)
Product concentration (mM) pellet a
supernatant a
b
0
0.4
L-Trp N-Carbtrp L-Trp + N-Carbtrp
15.44 1.86 17.30
b 7.11
0.15 0.04 0.19
0.17
1
12.0
L-Trp N-Carbtrp L-Trp + N-Carbtrp
45.82 5.55 51.37
40.87
0.98 1.22 2.20
0.98
2
17.1
L-Trp N-Carbtrp L-Trp + N-Carbtrp
45.09 5.06 50.15
42.83
2.35 1.22 3.57
2.25
3
22.2
L-Trp N-Carbtrp L-Trp + N-Carbtrp
43.12 4.74 47.86
40.96
4.17 0.73 4.90
4.02
4
25.5
L-Trp N-Carbtrp L-Trp + N-Carbtrp
37.78 3.48 41.26
39.49
5.10 0.39 5.49
4.85
Intact cells were preincubated with NaDOC up to 4 h before applying them to the reaction mixture. Conditions: 0.1 M phosphate buffer, pH 8.5, 4.2 g 1-1 CDM, 27°C, 24 h. a: addition of 43.7 mM D,L-5-IMH; b: addition of 40.5 mM L-N-Carbtrp.
370
cells. A two-fold increase of the reaction rate could be reached by growing Arthrobacter sp. DSM 3747 in the presence of the new inducer. Increasing the substrate concentration was another approach to accelerate the reaction rate. Since increasing the D,L-5-IMH concentration by adding organic solvents to the reaction mixture could be achieved only to a certain extent (Gross et al., 1987b), permeabilization of the cells was investigated to improve the availability of the sparingly soluble substrate. In order to enable free exchange between intracellular enzymes and the surrounding medium several organic solvents (Gross et al., 1987b) and detergents were added to the reaction mixture to permeabilize the cells. The results are summarized in Table 2. Except Triton X-100 nonionic detergents were inhibitory. With cetyltrimethylammonium bromide hydantoinase activity was double that of the control, while no change in L-N-carbamoylase activity could be observed. Sodium desoxycholate (NaDOC) led to a 2.6-fold and 2-fold increase of hydantoinase and L-N-carbamoylase, respectively. Freezing and thawing of the biomass proved to be only slightly successful. Diffusion of the substrate D,L-5-IMH into the cells had obviously been a rate-limiting step of the reaction, while the subsequent loosening of D,L-5-IMH seemed to cause no further limitation. In the following experiments the conditions of permeabilization with N a D O C relating to the concentration, incubation time and temperature were optimized. As shown in Fig. 3 the optimal concentration was 0.75-1.0 g g - ] (NaDOC per CWM; w/w). This ratio was applicable even to the experiments with higher
30
/i/
E 213
Q.
g "6 1G
---....
// ~
lt-
t
I 0.5
"- - - ' - - ' - - - ~
I
I 1.0
g NoDoc per g
CWM
I
I 1.5
Fig. 3. Effect of various NaDOC-concentrations on product formation; conditions: 43.7 m M D,L-5-IMH, 0.1 M glycine-NaOH, p H 8.5, 4.2 g 1- 1 C D M , 27 o C, 24 h. A, L-Trp; O, N-Carbtrp; I , L-Trp + N-Carbtrp.
371 concentrations of biomass. In contrast to the reaction without N a D O C the optimal temperature decreased to 30°C. The increasing effect of detergents at higher temperatures which may lead to a rapid destruction of the cytoplasmic membrane could be responsible for this effect. Therefore the temperature optimum for intact cells at 50°C is a result of increasing reaction rate and membrane permeability. The permeabilization of biomass for the transformation of D,L-5-IMH came to completion after 1 h of incubation with N a D O C as shown in Table 3. At this point, 12% protein of the biomass employed could be monitored in the supernatant. A rather low constant amount of N-Carbtrp and a slowly increasing amount of L-Trp were detectable in the supernatant. This L-Trp was probably set free from the cells during the preincubation with N a D O C and therefore is derived from the intracellularly accumulated inducer. Neither hydantoinase nor L-N-carbamoylase activity decreased during an incubation period of 3 h with N a D O C and the enzymes therefore are not released from the ceils during this time. Whether the decrease in enzymatic activities following lengthy incubation was due to inactivation inside or outside the cell remains to be investigated.
Transformation with permeabilized cells Transformations with intact cells at 50°C and permeabilized cells at 30°C in a 5-1 bioreactor (43.7 mM D,L-5-IMH, 0.1 M phosphate buffer, pH 8.5, 12.5 g CDM) were compared. A decline of L-N-carbamoylase and hydantoinase activity after about 1.5 and 8.5 h, respectively, could be observed at 50°C, while no loss in activity occurred at 30°C. Nevertheless, the conversion of 87.4 mM D,L-5-IMH to L-Trp at 30°C took more than twice as long as the conversion of 43.7 mM D,L-5-IMH at the same temperature. While in both cases the cleavage of D,L-5-IMH was completed first in comparable times, that of the residual N-Carbtrp took longer in the reaction mixture containing 87.4 mM D,L-5-IMH. Product inhibition caused by L-Trp seemed unlikely because the reaction velocity remained nearly constant until the complete consumption of D,L-5-IMH despite the increasing concentration of L-Trp in the reaction mixture. This decrease in L-N-carbamoylase activity must depend on the difference in the remaining N-Carbtrp concentration, which is increased with increasing D,L-5-IMH concentrations. Simple substrate inhibition by L-N-Carbtrp could not be assumed because 20 mM N-Carbtrp, a concentration of L-N-Carbtrp unknown to be inhibitory, was not even reached at D,L-5-IMH concentrations of 43.7 m M or 87.4 raM. These results could be established after further investigations, which are described in the following passage.
Stereospecificity of conoersion The cleavage of D- and L-5-IMH was studied to determine the influence of the enantiomers on the reaction. The reaction courses of both enantiomers of 5-IMH leading to L-Trp showed only minor differences (Fig. 4). The conversion of 5-IMH was complete after 1 h, while another 3 h were necessary to complete the transformation of the residual N-Carbtrp to L-Trp. Thus the reaction catalyzed by L-Ncarbamoylase once again represented the rate-limiting step. Determination of the
372 601
60
A
-
~
B
/
g
/
g
~
2
4 Time (h)
20
6
2
4 Time(h)
6
Fig. 4. "rime course of product formation with intact cells in a 54 bioreactor; conditions: 43.7 mM D,L-5-IMH, 0.1 M phosphate buffer, pN 8.5, 12.5 g 1- ] CDM, 50°C. (A) Transformation of L-5-1MH; (B) transformation of D-5-IMH. A, L-Trp; @, N-Carhtrp; I , L-Trp + N-Carbtrp.
configuration of the N-Carbtrp formed revealed that both L-N-Carbtrp and D-NCarbtrp were produced from L-5-IMH as well as D-5-IMH. In both cases the ratio of L-N-Carbtrp to D-N-Carbtrp was rather constant at ca. 7 / 3 until the conversion
2O
E
10
'7
10
I
I
I
2O
Time (h)
Fig. 5. Dependence of L-N-carbamoylase activity on the configuration of the substrate; conditions: 40.5 mM (zx), D,L- (A), D- (A), L-N-Carbtrp, 0.1 M phosphate buffer, pH 8.5, 4.2 g 1-1 CDM, 27 ° C.
373 of 5-IMH was complete. The hydantoinase reaction therefore proved not to be stereospecific. In the following phase of transformation L-Trp was produced predominantly from L-N-Carbtrp, so the optical purity of the D-enantiomer increased to 100% after a reaction time of 5 h. These results lead to the conclusion that the mutant of Arthrobacter sp. DSM 3747 was able to produce L-Trp from both enantiomers of 5-IMH and N-Carbtrp. To elucidate the relationship of the decrease in L-N-carbamoylase activity after complete conversion of 5-IMH to the increase in D-N-Carbtrp, the reaction courses of the conversion of D,L-, D- and L-N-Carbtrp were investigated separately (Fig. 5). The hydrolysis of D,L- and D-N-Carbtrp to L-Trp was clearly slower than that of L-N-Carbtrp to L-Trp. The increasing deceleration of the conversion of D,L-NCarbtrp compared with that of L-N-Carbtrp - before a yield of 50% was reached indicated an increasing inhibition of the L-N-carbamoylase activity caused by the D-enantiomer of N-Carbtrp. These results could be reproduced with permeabilized cells, so it was assured, that the permeabilization barriere had no influence on this effect. These results further suggested that the decrease in reaction velocity, especially at higher substrate concentrations of D,L-5-IMH, was due to a higher proportion of N-Carbtrp and therefore also D-N-Carbtrp formed. D-N-Carbtrp subsequently causes direct inhibition of L-N-carbamoylase activity on the one hand and indirect
100
80
6O "7" I t~ 4Q
20
2
6
I0
Time ( h )
Fig. 6. Comparison of the relative rates of chemical racemization and racemization during enzymatic transformation; conditions: o, chemical racemization - 4.37 mM D-5-IMH,0.1 M phosphate buffer, pH 8.5, 50 o C; e, enzymatic transformation - 43.7 mM D-5-IMH, 0.1 M phosphate buffer, pH 8.5, 12.5 g 1-1 CDM, 50°C.
374 inhibition by substrate limitation on the other hand. These results gave the strong evidence that at least one racemase was involved in this transformation, probably at the level of 5-IMH, because both D- and L-N-Carbtrp were produced in the same ratio from D- and L-5-IMH, respectively. This may be explained by a lower affinity of this racemase to the L-enantiomer of 5 - I M H in combination with a non-stereoselective hydantoinase a n d / o r with a D- and L-hydantoinase of different reaction velocities. The conversion of D-N-Carbtrp is more likely to occur after the backreaction to 5 - I M H than through an additional racemase, as it had been proposed earlier by Yokozeki et al. (1987a) and Guivarch et al. (1980). While chemical racemization of N-Carbtrp could be excluded (Dinelli et al., 1978), the ratio of chemical and enzymatic racemization of 5 - I M H had to be determined. Thus the chemical racemization rate of D - 5 - I M H at p H 8.5 in 0.1 M potassium phosphate buffer was monitored at 30 and 50°C. The racemization rate at 30°C was negligibly low compared with the enzymatic reaction rates at 30°C. Therefore the racemization at 30 ° C is enzymatic. Chemical racemization proceeded more rapidly at 50 ° C, but again in comparison with the enzymatic conversion rates of D,L-5-IMH at 50 ° C it could be assumed that the racemization is mainly enzymatic (Fig. 6). The properties of this racemase with regard to 5 - I M H are currently under investigation. The following reaction scheme for the conversion of D , L - 5 - I M H to L-Trp by a mutant of Arthrobacter sp. DSM 3747 can be advanced from the above results: D-5-IMH
back¢e----reaction
, D-N-Carbtrp
Race-II mizati°~ydantoinase L-5-IMH
~-~
~ L-N-Carbtrp
L-N-carbamoylase
, L-Trp
Operational stability of the reaction system In order to determine the operational stability of this process a semicontinuous procedure was studied under optimized conditions and relatively low substrate concentrations to prevent an accumulation of D-N-Carbtrp. The reaction course at 3 0 ° C using permeabilized ceils was followed with 43.7 m M D,L-5-IMH as substrate. When transformation was complete following 2.83 h, the cells were harvested by centrifugation and new buffer and substrate were added. This procedure was repeated four times. No decrease of the enzymatic activities could be observed during the investigated time period of 11.3 h. A reaction yield of 146 m M L-Trp was obtained after this period of incubation. The half-life of enzymatic activities at 30°C in permeabilized cells has to be ascertained. The use of immobilized cells in facilitating separation of the reaction medium in order to prevent the accumulation of D-N-Carbtrp, when adding D , L - 5 - I M H continuously, seems to be promising for the application of a commercial process. These aspects will be investigated.
375
Acknowledgement This work was supported by grants of the BMFT and Riitgerswerke AG, Castrop Rauxel (F.R.G.).
References Cotoras, W. and Wagner, F. (1984) Stereospecific hydrolysis of 5-monosubstituted hydantoins. In: Proceedings 3rd European Congress on Biotechnology, Vol. 1, Verlag Chemie, Weinheim, pp. 351-354. Dinelli, D., Marconi, W., Cecere, F., Galli, G. and Morisi, F. (1978) A new method for the production of optically active aminoacids. In: Pye, E.K. and Weethall, H.H. (Eds.), Enzyme Engineering, Vol. 5, Plenum Publishing Cooperation, New York, pp. 477-481. Guivarch, M., Gillonier, C. and Brunie, J.-C. (1980) Obtention d'amino acides optiquement actifs ~ l'aide d'hydantoinases. Bull. Soc. Chim. France 1-2, 91-95. Gross, C., Syldatk, C. and Wagner, F. (1987a) Screening method for microorganisms producing L-amino acids from D,L-5-monosubstituted hydantoins. Biotechnol. Techniques 1, 85-90. Gross, C., Syldatk, C. and Wagner, F. (1987b) Development of a process for the production of aromatic L-amino acids from D,L-5-monosubstitued hydantoins. In: Neijssel, O.M., Van der Meer, R.R., Luyben, K.Ch.A.M. (Eds.), Proceedings 4th European Congress on Biotechnology, Vol. 2, Elsevier, Amsterdam, pp. 249-252. Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) Protein measurement. J. Biol. Chem. 193, 265-275. Nishida, Y., Nakamichi, K., Nabe, K. and Tosa, T. (1987) Enzymatic production of L-tryptophan from D,L-5-indolylmethylhydantoin by Flavobacterium sp. Enzyme Microb. Technol. 9, 721-725. Syldatk, C., Cotoras, D., Mi)ller, A. and Wagner, F. (1986) Mikrobielle, enantioselektive Hydrolyse von D,L-5-monosubstituierten Hydantoinen zur Gewinnung von D- und L-Aminosiiuren. Biotech. Forum 3, 10-19. Syldatk, C., Mackowiak, V., Dombach, G., Gross, C., RShrmann, A. and Wagner, F. (1987) Microbial, enantioselective hydrolysis of D,L-N-carbamoylaminoacids to the corresponding L-amino acids by resting cells of Arthrobacter sp. DSM 3747. In: Neijssel, O.M., Van der Meer, R.R., Luyben, K.Ch.A.M. (Eds.), Proceedings 4th European Congress on Biotechnology, Vol. 2, Elsevier, Amsterdam, pp. 245-248. Syldatk, C., Mackowiak, V., H6ke, H., Gross, C., Dombach, G. and Wagner, F. (1990) Cell growth and enzyme synthesis of a mutant of Arthrobacter sp. (DSM 3747) used for the production of L-amino acids from D,L-5-monosubstituted hydantoins. Biotechnology, 14, 285-302. Yokozeki, K., Hirose, K. and Kubota, K. (1987a) Mechanism of asymmetric production of L-aromatic amino acids from the corresponding hydantoins by Flavobacterium sp. Agric. Biol. Chem. 51, 737-746. Yokozeki, K., Sano, K., Eguchi, C., Swagami, H. and Mitsugi, K. (1987b) Optimal Conditions for the enzymatic production of L-aromatic amino acids from the corresponding 5-substituted hydantoins. Agric. Biol. Chem. 51,729-736.
363
Elsevier BIOTEC 00523
Production of L-tryptophan from D,L-5-indolylmethylhydantoin by resting cells of a mutant of Arthrobacter species (DSM 3747) C h r i s t i a n e G r o s s , C h r i s t o p h Syldatk, V e r a M a c k o w i a k a n d F r i t z W a g n e r Institute of Biochemistry and BiotechnologyA, Technical University of Braunschweigo Braunschweig, F.R.G.
(Received 14 April 1989; accepted 7 January 1990)
Summary The reaction parameters and the stereospecificity of the enzymatic cleavage of D,L-5-indolylmethylhydantoin in producing L-tryptophan with resting cells of Arthrobacter sp. D S M 3747 were studied. When intact cells were tested, the optimal p H was between 8.5 and 9.0 and the optimal temperature was 5 0 ° C . Both, L-N-carbamoylase and hydantoinase could be stabilized over 24 h at 30 and 40 ° C by the addition of D,L-5-indolylmethylhydantoin. Furthermore, the hydantoinase was stable over 24 h at 50°C by the addition of 0.5 m M Mn 2+ ions. The treatment with sodium desoxycholate turned out to be successful in overcoming the poor availability of D,L-5-indolylmethylhydantoin for the cells. The optimal temperature with permeabilized cells decreased to 3 0 ° C and therefore ensured a good.enzyme stability. While the L-N-carbamoylase proved to be absolutely L-specific, the hydantoinase led to a mixture of enantiomers of N-carbamoyltryptophan. The produced D-N-carbamoyl-tryptophan caused an inhibition of the L-N-carbamoylase. The transformation yield from D,L-5-indolylmethylhydantoin always reached 100%. Arthrobacter sp.; L-Tryptophan; Transformation; Stereospecificity
Introduction Aromatic just as rare L-amino acids are gaining increasing importance for the supplementation of feedstuff and for pharmaceutical applications. One approach to
Correspondence to: Dr. C. Gross, Institute of Biochemistryand BiotechnologyA, Technical University of
Braunschweig, Konstantin-Uhde-Str. 5, D-3300, Braunschweig, F.R.G. 0168-1656/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
364
producing L-amino acids is the use of hydantoin-hydrolyzing enzymes. While the ability to produce D-amino acids from racemic 5-monosubstituted hydantoins is widespread among microorganisms (Syldatk et al., 1986), only a few species seem to be able to produce L-amino acids along this route. According to the research reported to date, productivities of the hydantoin-hydrolyzing enzymes from different bacterial sources vary considerably. The enzymes however seem to possess several common features. L-N-Carbamoylase represents an enzyme with a strict stereospecificity for aromatic L-N-carbamoylamino acids (Syldatk et al., 1987; Yokozeki et al., 1987a; Nishida et al., 1987), while stereospecificity of the hydantoinase depends on the C-5-substituent of the hydantoinring. For example, the L-enantiomer of N-carbamoyltryptophan is predominantly formed from the hydantoin substituted with the bulky indolylmethyl rest. Thus, for the production of L-amino acids a process consisting of one enzymatic (hydantoinase) and one chemical step (N-carbamoylamino acid hydrolysis) could not be adopted as it had been for the production of optically pure D-amino acids (Dinelli et al., 1978). Although in most cases it was possible to quantitatively convert the employed racemic 5-monosubstituted hydantoins, most procedures lead to an unsatisfactory yield of L-amino acids (Guivarch et al., 1980; Yokozeki et al., 1987b). We previously reported the optimization of growth conditions and enzyme formation by a mutant of Arthrobacter sp. DSM 3747 employing D,L-5-indolylmethylhydantoin as a model substrate (Syldatk et al., 1990). This paper describes the development of a microbial process for the L-tryptophan production from D,L-5-indolylmethylhydantoin, using resting cells of a mutant of Arthrobacter sp. DSM 3747. Aiming at optimization, the influence of the configuration of the products formed was taken into particular consideration.
Materials and Methods
Chemicals 5-Indolylmethylhydantoin (5-IMH) and N-carbamoyltryptophan (N-Carbtrp) were supplied by Professor Krohn, Institute of Organic Chemistry, T U Braunschweig. Yeast extract was obtained from Ohly G m b H (Hamburg, F.R.G.). All other chemicals were purchased from Sigma (Miinchen, F.R.G.), Merck (Darmstadt, F.R.G.) and Serva (Heidelberg, F.R.G.). Microorganism All experiments were performed with a mutant of Arthrobacter sp. DSM 3747 which was selected as described before (Gross et al., 1987a). Culture conditions Cultivation was carried out in 500-ml Erlenmeyer flasks with 200 ml medium on a rotation shaker at 100 rpm and 27°C for 24 h or in a 20-1 bioreactor (BraunMelsungen, Melsungen, F.R.G.). A complex medium was used for these cultivations (Cotoras and Wagner, 1984). It was further optimized by increasing the concentra-
365 tion of the inducer to 2 g 1-1 D,L-5-IMH. The biomass grown in the bioreactor was harvested after 10 h of growth by centrifugation at 16000 x g and 20°C for 30 min. At this time, maximal specific enzyme activity had been reached with 0.45 m M N-Carbtrp + L-Trp g-1 (cell dry mass = C D M ) h -1 for hydantoinase and 0.37 m M L-Trp g-1 h -~ for L-N-carbamoylase at 27°C.
Reaction conditions with resting cells Wet biomass obtained after centrifugation of the culture was applied directly to the biotransformation. The reaction was done either in 100-ml Erlenmeyer flasks with 10 ml buffer or in a 5-1 bioreactor of the type Biostat V (Braun-Melsungen, Melsungen, F.R.G.). p H adjustment in the 5-1 bioreactor was carried out with 10% N a O H . A concentration of 4.37 m M D , L - 5 - I M H could be solubilized in the reaction mixture, at higher concentrations it was added as a solid component. The data even of higher concentrations of D , L - 5 - I M H are given in m M to simplify the comparison of the product and substrate concentrations. All experiments were performed in a N 2 atmosphere to prevent the degradation of L-Trp via a pathway initiated by the oxygen-dependent tryptophan oxygenase (Gross et al., 1987b). The conditions of the experiments are given in the text. It could be shown that the growing cells were able to partly store the inducer D,L-5-IMH. Under reaction conditions the stored D , L - 5 - I M H was liberated by the cells as L-tryptophan (L-Trp), so in some cases the reaction yield exceeded 100%. Protein concentration In the case of detergent addition, the protein concentration in the supernatant of the reaction mixture was measured by the method of Lowry et al. (1951) before starting the reaction by adding the substrate. Bovine serum albumin was used as calibration standard. Quantitatioe determination of the reaction products The concentrations of the reaction products were measured in the supernatant of the reaction mixture after centrifugation at 3000 x g and 20°C for 15 min. This quantitative determination was done by H P L C as described before (Gross et al., 1987b). The optical purities were analyzed by H P L C (Spectra Physics, Hamburg, F.R.G.) using different chiral columns (Serva, Heidelberg, F.R.G.), and conditions for the separation of the D- and L-enantiomers. A Chiral Val Cu-column was used for the separation of D,L-5-IMH; eluent 10 m M CUSO4/5% acetonitril, flow rate 0.7 ml min-1, detection at 260 nm. D- and L-N-Carbtrp could be determined on a Chiral Pro Cu-column; eluent 1 m M CuSO 4, flow rate 1.3 ml rain-1, detection at 230 nm. Before measuring the enantiomer ratio of N-Carbtrp, it was usually necessary to remove L-Trp from the samples by ion-exchange chromatography on Amberlite IR-120 (H + form) because of interference with determination when it was present at high concentrations. The analysis of D- and L-Trp was performed on a Chiral Val Cu-column; eluent 1 m M CuSO4, flow rate 1 ml rain -1, detection at 280 nm.
366 Results and Discussion
Effect of pH The effect of pH is shown in Fig. 1. The highest yield of L-Trp from D,L-5-IMH was obtained between pH 8.5 and 9.0. The influences of the various buffers were
2,0
A
1.5-E
v
Ii .J ÷
1.0--
)
D.. f_
u i
Z
0.5
/ •
I
6
I
I
8
1.5 --
,
,
L
10
12
pH B
E1.o! 0.5~
I 6
l
I
8
I 10
12
pH Fig. 1. Effect of p H on product formation; conditions: 1.75 m M D , L - 5 - I M H , 4.2 g 1-~ C D M , 27 ° C, 3 h.
(A) Influence on the hydantoinase; (B) influence on the L-N-carbamoylase. e, Tris-maleate; ~, glycine-NaOH; A, Tris-HCl; m, diethanolamine-HC1.
367
2.0
_g
1.0
20
z,O
/0
Tempereture ('C)
Fig. 2. Effect of temperature on product formation. Conditions: 1.75 mM D,L-5-IMH, 0.1 M glycineNaOH, pH 8.5, 4.2 g 1-1 CDM, 3 h. ,, L-Trp; O, N-Carbtrp; II, L-Trp+ N-Carbtrp. quite different, especially Tris-HC1 and diethanolamine-HC1 lead to an unfavourable ratio of L-Trp to N-Carbtrp. Among several reaction solutions tested (Gross et al., 1987b), 0.1 M potassium phosphate buffer or glycine-NaOH gave the best results regarding a good conversion yield in favour of the production of L-Trp.
Effect of temperature After 3 h of incubation, the optimal temperature for the L-Trp formation was 50°C, as given in Fig. 2.
Effect of products The influence of NH~- and H C O f ions, L-N-Carbtrp and L-Trp on the biotransformation was investigated in order to prevent a possible inhibition. NH4CI, N a H C O 3 and NaC1, to detect possible effects of the counterions and the ionic strength, were added in concentrations of 1, 10 and 50 m M to the reaction mixtures. N o inhibition could be detected, but in all cases an increase in conversion of maximal 38% was observed, which led in most cases to a deterioration of the ratio of L - T r p / N - C a r b t r p . Adding either L-N-Carbtrp or L-Trp up to a concentration of 20 m M to the reaction mixture resulted in an increase in transformation after 24 h of incubation. F o r the 20 m M concentration of L-N-Carbtrp the yield was double that of the control containing no L-N-Carbtrp. Hydantoinase was activated especially beyond a L-N-Carbtrp concentration of 10 mM, while the increase in the L-Trp concentration could be caused either by activation or a shift in reaction equilibrium produced by the high substrate concentration for L-N-carbamoylase. The strong effect of L-NCarbtrp concentrations greater than 10 m M may be due to improved diffusion through the cell membrane which results in a sufficient intracellular concentration. For the complete range of L-Trp concentrations the yield was 30% higher than that
368 TABLE 1 Influence of temperature on the stability of enzymatic activities in the presence and absence of various additives Addition
Temp.
Fresh cells *
mM L-Trp
mM L-Trp + N-Carbtrp
1.76
1.95
None 0.5 mM MnSO4 0.5 mM ZnSO4 87.4 raM, D,L-5-IMH
30 o C
0.54 0.93 0.54 2.30
0.86 1.54 1.31 2.91
None 0.5 mM MnSO4 0.5 mM ZnSO4 87.4 mM D,L-5-IMH
40 o C
0.59 0.29 0.54 1.62
1.00 0.94 1.59 3.08
None 0.5 mM MnSO4 0.5 mM ZnSO4 87.4 mM D,L-5-IMH
50 o C
0.04 0.05 0.02 0.25
0.07 1.83 0.04 0.50
The remaining activity of resting cells, which were preincubated for 24 h at 30, 40 and 50 ° C, was compared to the activity of cells measured immediately after cultivation (*). Conditions: 4.37 mM D,L-5-IMH, 0.1 M glycine-NaOH, pH 8.5, 4.2 g 1-1 CDM, 27°C, 3 h.
of the control to which n o L-Trp had b e e n added. L-Trp stimulated the entire t r a n s f o r m a t i o n from D , L - 5 - I M H to L-Trp. The increased yield was correlated with a higher ratio of L - T r p / N - C a r b t r p .
Stability of hydantoinase and L-N-carbamoylase Cells were p r e i n c u b a t e d at different p H values at 2 7 ° C for 24 h. T h e best stability of e n z y m a t i c activities was observed at p H 8 - 1 0 with a residual activity of 50-60% a n d 4 0 - 5 0 % for h y d a n t o i n a s e a n d L-N-carbamoylase, respectively. T h e results of i n c u b a t i n g the cells at various temperatures for 24 h before m e a s u r i n g the r e m a i n i n g activity c a n be seen from T a b l e 1. W i t h o u t additions, b o t h enzymes t u r n e d out to be very labile even at 30 a n d 4 0 ° C with a drop in activity of 50% a n d 70% for h y d a n t o i n a s e a n d L-N-carbamoylase, respectively. A t 5 0 ° C b o t h enzymatic activities were negligibly low. By a d d i n g D , L - 5 - I M H at a c o n c e n t r a t i o n that could n o t be converted completely even at 5 0 ° C d u r i n g the p r e i n c u b a t i o n period, b o t h e n z y m a t i c activities could be restored at 30 a n d 40°C. T h e activation of h y d a n t o i n a s e at 30 a n d 4 0 ° C a n d of L - N - c a r b a m o y l a s e at 30°C m a y be explained by the p e r m e a b i l i z a t i o n of the cells d u r i n g the i n c u b a t i o n period, which yielded higher conversion rates i n the following enzyme test. I n contrast to L - N - c a r b a m o y lase h y d a n t o i n a s e could b e stabilized b y the a d d i t i o n of M n 2÷ ions up to 50°C a n d b y Z n 2÷ ions up to 40°C.
Permeabilization of the cells Since L - N - c a r b a m o y l a s e in particular could n o t be sufficiently stabilized for a longer period of time at higher temperatures, it was necessary to increase the
369 TABLE 2 Influence of several detergents on cell permeabilization. Conditions: 43.7 mM D,L-5-1MH, 0.1 mM glycine-NaOH, pH 8.5, 4.2 g 1-1 CDM, 27 o C, 24 h. Addition
g 1-1
mM L-Trp
mM L-Carbtrp
mM L-Trp + N-Carbtrp
None Triton X-100 Triton X-405 Tween 20 Tween 40 Tween 60 Span 80 NaDOC CTAB * 20 o C
10 10 10 10 10 10 10 10
8.28 6.66 7.20 2.50 3.53 4.70 4.46 17.05 8.18 10.99
0.65 4.41 0.53 0.20 0.28 0.61 0.65 6.20 9.59 1.17
8.93 11.07 7.73 2.70 3.81 5.31 5.11 23.25 17.74 12.16
-
* Cetyltrimethyiammonium bromide. reaction rate in order to shorten the duration of reaction. Thus a prolonged use of t h e cells w o u l d b e m a d e p o s s i b l e . A t f i r s t t h e e n z y m a t i c a c t i v i t i e s o f cells g r o w n i n t h e p r e s e n c e o f D , L - 5 - N - 3 m e t h y l h y d a n t o i n ( S y l d a t k e t al., 1990) w e r e c o m p a r e d t o t h o s e i n u s u a l l y g r o w n
TABLE 3 Dependence of cell permeabilization and transformation yield on the length of incubation with NaDOC Time (h)
Protein CDM -1 (%) (supernatant)
Product concentration (mM) pellet a
supernatant a
b
0
0.4
L-Trp N-Carbtrp L-Trp + N-Carbtrp
15.44 1.86 17.30
b 7.11
0.15 0.04 0.19
0.17
1
12.0
L-Trp N-Carbtrp L-Trp + N-Carbtrp
45.82 5.55 51.37
40.87
0.98 1.22 2.20
0.98
2
17.1
L-Trp N-Carbtrp L-Trp + N-Carbtrp
45.09 5.06 50.15
42.83
2.35 1.22 3.57
2.25
3
22.2
L-Trp N-Carbtrp L-Trp + N-Carbtrp
43.12 4.74 47.86
40.96
4.17 0.73 4.90
4.02
4
25.5
L-Trp N-Carbtrp L-Trp + N-Carbtrp
37.78 3.48 41.26
39.49
5.10 0.39 5.49
4.85
Intact cells were preincubated with NaDOC up to 4 h before applying them to the reaction mixture. Conditions: 0.1 M phosphate buffer, pH 8.5, 4.2 g 1-1 CDM, 27°C, 24 h. a: addition of 43.7 mM D,L-5-IMH; b: addition of 40.5 mM L-N-Carbtrp.
370
cells. A two-fold increase of the reaction rate could be reached by growing Arthrobacter sp. DSM 3747 in the presence of the new inducer. Increasing the substrate concentration was another approach to accelerate the reaction rate. Since increasing the D,L-5-IMH concentration by adding organic solvents to the reaction mixture could be achieved only to a certain extent (Gross et al., 1987b), permeabilization of the cells was investigated to improve the availability of the sparingly soluble substrate. In order to enable free exchange between intracellular enzymes and the surrounding medium several organic solvents (Gross et al., 1987b) and detergents were added to the reaction mixture to permeabilize the cells. The results are summarized in Table 2. Except Triton X-100 nonionic detergents were inhibitory. With cetyltrimethylammonium bromide hydantoinase activity was double that of the control, while no change in L-N-carbamoylase activity could be observed. Sodium desoxycholate (NaDOC) led to a 2.6-fold and 2-fold increase of hydantoinase and L-N-carbamoylase, respectively. Freezing and thawing of the biomass proved to be only slightly successful. Diffusion of the substrate D,L-5-IMH into the cells had obviously been a rate-limiting step of the reaction, while the subsequent loosening of D,L-5-IMH seemed to cause no further limitation. In the following experiments the conditions of permeabilization with N a D O C relating to the concentration, incubation time and temperature were optimized. As shown in Fig. 3 the optimal concentration was 0.75-1.0 g g - ] (NaDOC per CWM; w/w). This ratio was applicable even to the experiments with higher
30
/i/
E 213
Q.
g "6 1G
---....
// ~
lt-
t
I 0.5
"- - - ' - - ' - - - ~
I
I 1.0
g NoDoc per g
CWM
I
I 1.5
Fig. 3. Effect of various NaDOC-concentrations on product formation; conditions: 43.7 m M D,L-5-IMH, 0.1 M glycine-NaOH, p H 8.5, 4.2 g 1- 1 C D M , 27 o C, 24 h. A, L-Trp; O, N-Carbtrp; I , L-Trp + N-Carbtrp.
371 concentrations of biomass. In contrast to the reaction without N a D O C the optimal temperature decreased to 30°C. The increasing effect of detergents at higher temperatures which may lead to a rapid destruction of the cytoplasmic membrane could be responsible for this effect. Therefore the temperature optimum for intact cells at 50°C is a result of increasing reaction rate and membrane permeability. The permeabilization of biomass for the transformation of D,L-5-IMH came to completion after 1 h of incubation with N a D O C as shown in Table 3. At this point, 12% protein of the biomass employed could be monitored in the supernatant. A rather low constant amount of N-Carbtrp and a slowly increasing amount of L-Trp were detectable in the supernatant. This L-Trp was probably set free from the cells during the preincubation with N a D O C and therefore is derived from the intracellularly accumulated inducer. Neither hydantoinase nor L-N-carbamoylase activity decreased during an incubation period of 3 h with N a D O C and the enzymes therefore are not released from the ceils during this time. Whether the decrease in enzymatic activities following lengthy incubation was due to inactivation inside or outside the cell remains to be investigated.
Transformation with permeabilized cells Transformations with intact cells at 50°C and permeabilized cells at 30°C in a 5-1 bioreactor (43.7 mM D,L-5-IMH, 0.1 M phosphate buffer, pH 8.5, 12.5 g CDM) were compared. A decline of L-N-carbamoylase and hydantoinase activity after about 1.5 and 8.5 h, respectively, could be observed at 50°C, while no loss in activity occurred at 30°C. Nevertheless, the conversion of 87.4 mM D,L-5-IMH to L-Trp at 30°C took more than twice as long as the conversion of 43.7 mM D,L-5-IMH at the same temperature. While in both cases the cleavage of D,L-5-IMH was completed first in comparable times, that of the residual N-Carbtrp took longer in the reaction mixture containing 87.4 mM D,L-5-IMH. Product inhibition caused by L-Trp seemed unlikely because the reaction velocity remained nearly constant until the complete consumption of D,L-5-IMH despite the increasing concentration of L-Trp in the reaction mixture. This decrease in L-N-carbamoylase activity must depend on the difference in the remaining N-Carbtrp concentration, which is increased with increasing D,L-5-IMH concentrations. Simple substrate inhibition by L-N-Carbtrp could not be assumed because 20 mM N-Carbtrp, a concentration of L-N-Carbtrp unknown to be inhibitory, was not even reached at D,L-5-IMH concentrations of 43.7 m M or 87.4 raM. These results could be established after further investigations, which are described in the following passage.
Stereospecificity of conoersion The cleavage of D- and L-5-IMH was studied to determine the influence of the enantiomers on the reaction. The reaction courses of both enantiomers of 5-IMH leading to L-Trp showed only minor differences (Fig. 4). The conversion of 5-IMH was complete after 1 h, while another 3 h were necessary to complete the transformation of the residual N-Carbtrp to L-Trp. Thus the reaction catalyzed by L-Ncarbamoylase once again represented the rate-limiting step. Determination of the
372 601
60
A
-
~
B
/
g
/
g
~
2
4 Time (h)
20
6
2
4 Time(h)
6
Fig. 4. "rime course of product formation with intact cells in a 54 bioreactor; conditions: 43.7 mM D,L-5-IMH, 0.1 M phosphate buffer, pN 8.5, 12.5 g 1- ] CDM, 50°C. (A) Transformation of L-5-1MH; (B) transformation of D-5-IMH. A, L-Trp; @, N-Carhtrp; I , L-Trp + N-Carbtrp.
configuration of the N-Carbtrp formed revealed that both L-N-Carbtrp and D-NCarbtrp were produced from L-5-IMH as well as D-5-IMH. In both cases the ratio of L-N-Carbtrp to D-N-Carbtrp was rather constant at ca. 7 / 3 until the conversion
2O
E
10
'7
10
I
I
I
2O
Time (h)
Fig. 5. Dependence of L-N-carbamoylase activity on the configuration of the substrate; conditions: 40.5 mM (zx), D,L- (A), D- (A), L-N-Carbtrp, 0.1 M phosphate buffer, pH 8.5, 4.2 g 1-1 CDM, 27 ° C.
373 of 5-IMH was complete. The hydantoinase reaction therefore proved not to be stereospecific. In the following phase of transformation L-Trp was produced predominantly from L-N-Carbtrp, so the optical purity of the D-enantiomer increased to 100% after a reaction time of 5 h. These results lead to the conclusion that the mutant of Arthrobacter sp. DSM 3747 was able to produce L-Trp from both enantiomers of 5-IMH and N-Carbtrp. To elucidate the relationship of the decrease in L-N-carbamoylase activity after complete conversion of 5-IMH to the increase in D-N-Carbtrp, the reaction courses of the conversion of D,L-, D- and L-N-Carbtrp were investigated separately (Fig. 5). The hydrolysis of D,L- and D-N-Carbtrp to L-Trp was clearly slower than that of L-N-Carbtrp to L-Trp. The increasing deceleration of the conversion of D,L-NCarbtrp compared with that of L-N-Carbtrp - before a yield of 50% was reached indicated an increasing inhibition of the L-N-carbamoylase activity caused by the D-enantiomer of N-Carbtrp. These results could be reproduced with permeabilized cells, so it was assured, that the permeabilization barriere had no influence on this effect. These results further suggested that the decrease in reaction velocity, especially at higher substrate concentrations of D,L-5-IMH, was due to a higher proportion of N-Carbtrp and therefore also D-N-Carbtrp formed. D-N-Carbtrp subsequently causes direct inhibition of L-N-carbamoylase activity on the one hand and indirect
100
80
6O "7" I t~ 4Q
20
2
6
I0
Time ( h )
Fig. 6. Comparison of the relative rates of chemical racemization and racemization during enzymatic transformation; conditions: o, chemical racemization - 4.37 mM D-5-IMH,0.1 M phosphate buffer, pH 8.5, 50 o C; e, enzymatic transformation - 43.7 mM D-5-IMH, 0.1 M phosphate buffer, pH 8.5, 12.5 g 1-1 CDM, 50°C.
374 inhibition by substrate limitation on the other hand. These results gave the strong evidence that at least one racemase was involved in this transformation, probably at the level of 5-IMH, because both D- and L-N-Carbtrp were produced in the same ratio from D- and L-5-IMH, respectively. This may be explained by a lower affinity of this racemase to the L-enantiomer of 5 - I M H in combination with a non-stereoselective hydantoinase a n d / o r with a D- and L-hydantoinase of different reaction velocities. The conversion of D-N-Carbtrp is more likely to occur after the backreaction to 5 - I M H than through an additional racemase, as it had been proposed earlier by Yokozeki et al. (1987a) and Guivarch et al. (1980). While chemical racemization of N-Carbtrp could be excluded (Dinelli et al., 1978), the ratio of chemical and enzymatic racemization of 5 - I M H had to be determined. Thus the chemical racemization rate of D - 5 - I M H at p H 8.5 in 0.1 M potassium phosphate buffer was monitored at 30 and 50°C. The racemization rate at 30°C was negligibly low compared with the enzymatic reaction rates at 30°C. Therefore the racemization at 30 ° C is enzymatic. Chemical racemization proceeded more rapidly at 50 ° C, but again in comparison with the enzymatic conversion rates of D,L-5-IMH at 50 ° C it could be assumed that the racemization is mainly enzymatic (Fig. 6). The properties of this racemase with regard to 5 - I M H are currently under investigation. The following reaction scheme for the conversion of D , L - 5 - I M H to L-Trp by a mutant of Arthrobacter sp. DSM 3747 can be advanced from the above results: D-5-IMH
back¢e----reaction
, D-N-Carbtrp
Race-II mizati°~ydantoinase L-5-IMH
~-~
~ L-N-Carbtrp
L-N-carbamoylase
, L-Trp
Operational stability of the reaction system In order to determine the operational stability of this process a semicontinuous procedure was studied under optimized conditions and relatively low substrate concentrations to prevent an accumulation of D-N-Carbtrp. The reaction course at 3 0 ° C using permeabilized ceils was followed with 43.7 m M D,L-5-IMH as substrate. When transformation was complete following 2.83 h, the cells were harvested by centrifugation and new buffer and substrate were added. This procedure was repeated four times. No decrease of the enzymatic activities could be observed during the investigated time period of 11.3 h. A reaction yield of 146 m M L-Trp was obtained after this period of incubation. The half-life of enzymatic activities at 30°C in permeabilized cells has to be ascertained. The use of immobilized cells in facilitating separation of the reaction medium in order to prevent the accumulation of D-N-Carbtrp, when adding D , L - 5 - I M H continuously, seems to be promising for the application of a commercial process. These aspects will be investigated.
375
Acknowledgement This work was supported by grants of the BMFT and Riitgerswerke AG, Castrop Rauxel (F.R.G.).
References Cotoras, W. and Wagner, F. (1984) Stereospecific hydrolysis of 5-monosubstituted hydantoins. In: Proceedings 3rd European Congress on Biotechnology, Vol. 1, Verlag Chemie, Weinheim, pp. 351-354. Dinelli, D., Marconi, W., Cecere, F., Galli, G. and Morisi, F. (1978) A new method for the production of optically active aminoacids. In: Pye, E.K. and Weethall, H.H. (Eds.), Enzyme Engineering, Vol. 5, Plenum Publishing Cooperation, New York, pp. 477-481. Guivarch, M., Gillonier, C. and Brunie, J.-C. (1980) Obtention d'amino acides optiquement actifs ~ l'aide d'hydantoinases. Bull. Soc. Chim. France 1-2, 91-95. Gross, C., Syldatk, C. and Wagner, F. (1987a) Screening method for microorganisms producing L-amino acids from D,L-5-monosubstituted hydantoins. Biotechnol. Techniques 1, 85-90. Gross, C., Syldatk, C. and Wagner, F. (1987b) Development of a process for the production of aromatic L-amino acids from D,L-5-monosubstitued hydantoins. In: Neijssel, O.M., Van der Meer, R.R., Luyben, K.Ch.A.M. (Eds.), Proceedings 4th European Congress on Biotechnology, Vol. 2, Elsevier, Amsterdam, pp. 249-252. Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) Protein measurement. J. Biol. Chem. 193, 265-275. Nishida, Y., Nakamichi, K., Nabe, K. and Tosa, T. (1987) Enzymatic production of L-tryptophan from D,L-5-indolylmethylhydantoin by Flavobacterium sp. Enzyme Microb. Technol. 9, 721-725. Syldatk, C., Cotoras, D., Mi)ller, A. and Wagner, F. (1986) Mikrobielle, enantioselektive Hydrolyse von D,L-5-monosubstituierten Hydantoinen zur Gewinnung von D- und L-Aminosiiuren. Biotech. Forum 3, 10-19. Syldatk, C., Mackowiak, V., Dombach, G., Gross, C., RShrmann, A. and Wagner, F. (1987) Microbial, enantioselective hydrolysis of D,L-N-carbamoylaminoacids to the corresponding L-amino acids by resting cells of Arthrobacter sp. DSM 3747. In: Neijssel, O.M., Van der Meer, R.R., Luyben, K.Ch.A.M. (Eds.), Proceedings 4th European Congress on Biotechnology, Vol. 2, Elsevier, Amsterdam, pp. 245-248. Syldatk, C., Mackowiak, V., H6ke, H., Gross, C., Dombach, G. and Wagner, F. (1990) Cell growth and enzyme synthesis of a mutant of Arthrobacter sp. (DSM 3747) used for the production of L-amino acids from D,L-5-monosubstituted hydantoins. Biotechnology, 14, 285-302. Yokozeki, K., Hirose, K. and Kubota, K. (1987a) Mechanism of asymmetric production of L-aromatic amino acids from the corresponding hydantoins by Flavobacterium sp. Agric. Biol. Chem. 51, 737-746. Yokozeki, K., Sano, K., Eguchi, C., Swagami, H. and Mitsugi, K. (1987b) Optimal Conditions for the enzymatic production of L-aromatic amino acids from the corresponding 5-substituted hydantoins. Agric. Biol. Chem. 51,729-736.