Peptide antibiotic production through immobilized biocatalyst technology

o

Peptide antibiotic production through immobilized biocatalyst technology ERICK J. VANDAMME Laboratory o f General and Industrial Mierobiok~gy, University o f Ghent, Coupure Links, 653, B-9000 Ghent, Belgium

Summary. The use o f immobilized biocatalysts for producing known or new antibiotics is presented. A n evaluation o f the applicability o f this concept in the lascinating field o f peptide antibiotic hioconversions and fermentations is also given. The use o f immobilized enzymes, organelles and cells to synthesize antibiotics as an alternative method to conventional fermentation is discussed. In vitro total enzymatic antibiotic synthesis is illustrated with the 'multienzyme thiotemplate m e c h a n i s m ' o f Bacillus brevis, the producer o f gramicidin S. Total synthesis o l peptide antibiotics, based on immobilized living cells, has recently been demonstrated with penicillin, bacitracin, nisin and a few other antibiotics. A s an industrial example o f the use o f enz.vmes or cells to convert peptide antibiotics into therapeutically useful derivatives, f)'ee and immobilized penicillin ae.vlases, producing the penicillin nucleus 6-aminopenicillanic acid {6-APA), are reviewed as well as their potential to synthesize semisynthetic ~3-1actams (penicillins, cephalosporins). A cylases, aeetylesterases and ~-amino acid ester hydrolases acting on cephalosporin-compounds and yielding valuable intermediar.v or end products have also gained wide interest. Stereospecific enzymic side-chain preparations for semisynthetic penicillin and cephalosporin production have recently reached the industrial stage. Bioconversion possibilities with the novel J3-lactam compounds are suggested. These examples o f simple single-step, as well as complex multi-step, enzyme reactions point to the vast potential o f immobilized biocatal.vst technolog.v in fermentation science, in organic synthesis and in biotechnological processes in general. Keywords: Antibiotics; immobilized biocatalyst; bioconversion; fermentation; multi-step enzymic synthesis; ~-lactams

Introduction Antibiotics are very important and chemically very complex compounds, made exclusively by microbiological synthesis. Since the discovery of the first antibiotics and penicillin, more than 6000 natural microbial compounds have been described, all of which display antibiotic activity. Only about 150 are produced on a large scale and find use in medicine and agriculture, mainly as antibacterial, antifungal or antiviral agents. A minority find use as anti-

tumour agents, immune modulators, as antihelminthic, herbicidal, antiprotozoal, piscicidal or anticoccidial agents. feed additives, food preservatives, plant disease controllers or abcission agents) -3 At present, all contmercial antibiotics except two (chloramphenicol and cellocidin) are produced by microorganisms in conventional fermentation processes. In a few cases, the natural microbial product can be chemically or enzymatically converted into a so-called semisynthctic antibiotic with superior therapeutic properties. 4-6 In the last decade, several attempts have been made to apply the imnlobilized-enzyme or immobilized-cell principle to antibiotic fermentation and bioconversion processes. 7'8 At present, several antibiotic bioconversions involving single-step reactions are already being performed with immobilized enzymes or cells as biocatalysts on an industrial scale. The replacement of conventional fermentation (which usually involves a complex multi-step reaction sequence) with immobilized enzyme or immobilized living-cell technology also seems to offer great potential, but still needs further study at a fundamental level. This review evaluates the potential or real value of such multi-step enzymic total synthesis of antibiotics, of antibiotic fermentations using immobilized living cells and of immobilized biocatalyst antibiotic bioconversions.

Multi-step total enzymic synthesis of antibiotics A challenging field for the application of immobilized enzyme technology is tile total in vitro enzymic synthesis of antibiotics involving multi-step reactions and cofactor supplies. As already mentioned, few ccmamercial antibiotics are made by chemical synthesis, despite the fact that chemical routes are known for many important antibiotics. The economics of multi-reaction chemical synthesis of such complicated biological molecules are simply too unfavourable. It is significant that the development of the new semisynthetic penicillins did not start when the chemical synthesis route to the penicillin nucleus, 6-aminopenicillanic acid (6-APA). was discovered, but only later when it was found that under certain conditions the fungus Penieillium chrysogenum could excrete 6-APA, which could then bc converted into new pcnicillins. As organic synthesis is virtually impossible, antibiotic synthesis by traditional fermentation also has its drawbacks. Apart from the use of sophisticated production facilities, the impressive increase in antibiotic productivity in industrial fermenta-

0141-0229/831060403-14 $03.00

© 1983 Butterworth & Co. (Publishers) Ltd

Enzyme Microb. Technol., 1983, vol. 5, November

403

Review

tions is largely the result of forcing the nricroorganism to overproduce a useful metabolite by mutation or mutasynthesis, by directly influencing cellular metabolism and the cell's environment, (nutritional control, precursor addition), and by genetic engineering. 9 Most antibiotics are produced as secondary metabolites, i.e. their synthesis is delayed until the growth of the cells declines or stops. Consequently, such classical fermentation processes always have a 'non-productive' phase. Once all the necessary enzymes are formed in the cell, theoretically a 'linear' antibiotic production rate could be maintained over a long period if it were not for the fact that the involved enzymes are generally rapidly inactivated. Conversion of sugar substrates into antibiotics is rather inefficient because of growth, maintenance and the many side-reactions that occur in growing intact cells. Furthermore, strain degeneration, i.e. the selection of low productivity strains which grow faster, is a major problem in the fermentation industry. To solve these problems, attempts have been made to replace fermentation or cellular synthesis by acellular processes, i.e. total enzymic synthesis in vitro, xo, xl In such a process, it is the ultimate aim to use isolated, stabilized and innnobilized enzymes, which, in sequential reactions, perform the total synthesis of an antibiotic upon addition of its precursors, ATP and cofactors. This concept can be seen as an extension of the well-known and industrially applied simple enzymic bioconversions of antibiotics. This development, the acellular or cell-free total enzymic synthesis of antibiotics using immobilized enzymes, has yet to be exploited. In nrany such potential systems, extracts of cells have been shown to convert radioactive precursor(s) into an antibiotic, and partial enzymic synthesis of many antibiotic compounds is well documented. 2 Total enzymic synthesis h7 vitro has been studied particularly in the case of the oligopeptide antibiotics (gramicidin S, tyrocidines and bacitracin). :6 H5 0

CH2

vc.

%,. C¢-0

(CH3)~CH--? H

~H-CH2CH2CH 2 NH NH 2

O--~-~~ H

..I

I

. coo

NC CH(CH3)Z

NH Z CH2 CHz CH2 CH O//'C"N H ,,_H /C-.

H

C" H H/% _C O

,/¢,'a~ ~,c-._c_c-W, \

~c,"

o

H I ~ ",._1 CB H~

D- P h e ' ~

L-P~FO

L:" Leu

\

:l::

~L-Leu D- P h e ' /

Figure 1 Structure and

404

L-/

schematic representation

of gramicidinS

Enzyme Microb. Technol., 1983, vol. 5, November

The system that is best understood is the biosynthesis of gramicidin S (GS), a cyclic decapeptide antibiotic produced by certain Bacillus brevis strains (Figure 1). The number of enzymes in the biosynthetic pathway is small, i.e. GS synthetase 1 and 2. The enzymes, substrates and cofactors involved have been characterized and the biosynthetic mechanisnr has been resolved in detail. Both enzymes have been purified to homogeneity. Phosphopantotheine is the cofactor, which is tightly bound to GS synthetase 2. The constituent anaino acids, t,-Pro, t,-Val, L-Orn and L-Leu, are activated by enzyme 2 while activation and racemization of L-Phe to D-Phe is effected by enzyme 1. Enzyme 1 initiates the biosynthetic sequence while the peptide bonds (elongation and cyclization) are catalysed by enzyme 2. The amino acid sequence in gramicidin S is thus determined by the unique location of specific subunits on enzymes 1 and 2.13,14

The mechanism involved has been named 'multi-enzyme thiotemplate' biosynthesis, and it requires, in addition to the two enzymes, only the building-block amino acids or their analogues as precursors, ATP as an energy source, and Mg2+ ions (Figure 2). The overall biosynthesis scheme can be simplified as: 2 Phe 2 Pro GS synthetase 1 and 2 w, 2Val + I O A T P Mg2 + 20rn 2 Leu 1 G S + 10 AMP+ IOPP i

(1)

Using the two GS synthetases, isolated from high-yield gramicidin S-producing B. brevis ATCC 9999 cells, gram quantities of gramicidin S have already been produced in vitro by the total enzymic method. 11,12,14 The successful operation of continuous enzyme reactors fur such nrulti-step synthetic reactions depends largely upon the effective regeneration of ATP with cheap energy sources. ATP regeneration, in the context of total enzymic synthesis, has been investigated using immobilized enzymes and bacterial chromatophores, is' 16 yeast enzymes, 17 immobilized yeast cells 18 and polyphosphates. 19 The two GS synthetases could be coimmobilized to construct a bioreactor, coupled to an enzymic ATP-regeneration system, which may allow for a continuous process to be developed. Furthermore, instead of using expensive amino acids as substrates, cheap protein hydrolysates should be tested. A real problem is the transient formation of the GS synthetases during growth of B. brevis, which makes it difficult to harvest large quantities of these labile enzymes (Figure 3). Recent experiments have confirmed the presence of particulate GS synthetases, which might be easier to recover and to stabilize outside the cells. Use of these catalysts might be favourable to the soluble GS synthetases in view of" total enzymic synthesis of gramicidin S. ~4 It is also hoped to apply the knowledge gamed here to similar systems with high commercial value (tyrocidines, linear gramicidins and bacitracins). However, progress towards such a goal is slow, and the application of immobilized enzyme technology to the total enzymatic antibiotic synthesis concept has not yet been successful. Enzyme complexes localized on membranes or in cell organelles have indeed been found to be able to catalyse partial or even total antibiotic synthesis. The essential

Peptide antibiotic production." Erick J. Vandamme SH

Biosynthesiso f

PPI /

PPi ~

/

S

I . ATP F

L-PRO SH ~ ATD L-ORN L-LEU H S

Gnunicidin

VI

SH j

HS

(~SH

' Hl~/y~

% ~

SH /

o""

e , ~I S

~ ;,

,,v k

~,,

--I

~"

11

,_EU-S\ ~. ~.s%~~J

1

~

LEU.S

k

5-L --PHE n! --D --PHE ,v

H y " ~ ~ S _

D_pHE VIII

, . e - * " ~ " ~

~

VII

--] ""%'

SH

.,

7o. .rmo-,_

%

~ ~

-

"*%e" o-

-

,.. E

IX

Figure2 Model for the functioning of t h e m u l t i e n z y m e t h i o t e m p l a t e m e c h a n i s m i n g r a m i c i d i n S b i o s y n t h e s i s ( l = G S s y n t h e t a s e 1 , 2 = G s synthetase 2).~4 I, Schematic representation of GS synthetases 1 and 2. The zig-zag line represents 4'-phosphopantetheine. II, Activation of amino acids and binding on GS synthetases 1 and 2. III, Thioester binding of amino acids and conformational change of GS synthetase 2. IV, Racemization of L-Phe to its D-isomer. V, GS synthetases 1 and 2 complex, loaded with thioester-bound amino acids. Vl, Transfer of D-phenylalanine on GS synthetase I to a thiol site of GS synthetase 2. VII, Formation of the first intermediate dipeptide. VIII, Transfer of the dipeptide to the 4'-phosphopantetheine arm. IX, Two pentapeptides cyclize 'head to tail' in an intramolecular reaction on GS synthetase 2 [Reproduced from Vandamme, E. J. in Topics in Enzyme and Fermentation Biotechnology (Wiseman, A., ed.) Ellis Horwood, Chichester, 1981, vol. 5, p. 185, by permission of E t lis Horwood Limited"© ]

intermediate, 8-(oe-aminoadipyl)cysteinylvaline, in/3-1actam antibiotic biosynthesis is produced by a lysed inycelial Penicillium chrysogenum preparation in the presence of an energy-generating system.2° Kohsaka and Demain 21 demonstrated cephalosporin synthesis by protoplast lysates of Cephalosporium acremonium. Vesicles isolated from P. chrysogenum PQ-96 protoplasts displayed 8-(oe-aminoadipyl)cysteinylvaline synthetase activity,22 phenacylcoenzyme A ligase activity, and acyl exchange activity of transferases, and they converted 14C labelled aminoadipylcysteinylvaline into benzylpenicillin. Only vesicles, with a diameter of 40 to 200 rim, displayed penicillin G total-biosynthesis capacity.22 Ku,zatkowski et al. 2z immobilized these vesicles in a calcium alginate gel and 44% of the activity of native vesicles was retained. This activity remained stable after 240 h storage at 4°C, whereas the native vesicles had lost all biosynthetic activity within 60 h. The use of such stabilized superior organelle-biocatalysts in a continuous bioreactor might provide for improved penicillin production. There are four potential applications of such total enzymic synthesis.

(a) Rapid and efficient production of currently used and known antibiotics is possible. (b) In the production of biosynthetic intermediates as starting materials for chemical conversion or bioconversion into new antibiotics, certain enzymes can be omitted from such systems; new intermediates could be accumulated and made available to the chemist or microbiologist for further modification. (c) Total enzymic synthesis of novel antibiotics is feasible: specificities of antibiotic-forming enzymes are not very strict; however, by the use of an enzymic systein. enzyme specificity in vitro is even lower and substrates could be used which could otherwise not be incorporated by whole cells because of permeability constraints, toxicity, degradation or other problems. Experiments with cell-free systems have demonstrated that chemically related amino acids can substitute for the normal constituents of the antibiotics and, in this respect, the poorer specificity of cell-free enzyme systems can be used to produce novel antibiotics. New antibiotics are urgently needed for the effective control of Gram-negative infections, and antifungal, antiprotozoal, antiviral and antitumour agents

Enzyme Microb. Technol., 1983, vol. 5, November

405

Review 3000 k 2000~

h

f i OC.C

~ '

YP - medium

g t-

: ~k : : " :

2000

~

I ~18

i :: ? -

(:~

-11600 /

I

c ~

3.

I ~' q IO

c_.

O0

~1 : :

~/GS

2"

I:~ i

~c F

:

"

y"

I

~

~6

{

--

I! \i

(

.{./ 4

o

J ~oo

... l'i}

,o ~ . ~ - - ~ + _ . - < ~ L 2

t

~

~ ..

6

8 I0 12 22 24 Time (h) Figure3 Fermentation dynamics of gramicidinSandtransient formation of soluble GS synthetase 1 (=E~) and 2 (=E:) [Reproduced from Vandamme, E. J. in Topics in Enzymeand Fermentation Biotechnology (Wiseman, A., ed.) Ellis Horwood, Chichester, 1981, vol. 5, p. 185 by permission of Ellis Horwood Limited © ]

are urgently needed too. Furthermore, resistance to existing antibiotics is widespread among microorganisms and necessitates tile replacement of antibiotics now in use. (d) The concept can be applied to the synthesis of other complex molecules of medical, nutritional, agricultural or industrial importance. This concept is especially valuable in those cases where the desired product, such as gramicidin S, accumulates completely inside the microbial cells and cannot be liberated without killing or severely damaging the cells, a fact which has so far prevented tile use of inrmobilized whole living-cell technology to synthesize such products.

Antibiotic fermentations using immobilized living-cell technology The difficulties encountered with complex total enzymic synthesis processes using cell-free (inmrobilized) enzyme preparations might be circumvented by the use of innnobilized whole cells. Utilization of immobilized whole living cells for total synthesis of fermentation or organic products is only now emerging as a possibility. 23-26 This concept would be particularly valuable for those antibiotics which are, or can be, excreted into the fermentation medium. The production of penicillin G,2v bacitracin, 2~'29 cephamycin C,a° candicidin, al nisin, a2 and a few other antibiotic compounds, has been attempted with immobilized cells as catalysis. hnmobilized living-cell systems can offer several important advantages over conventional fermentation, such as the possibility of continuous operation (at high dilution rates will) no risk of washout) or plug-flow mode of action, column fermentation, reduction of non-productive growth phases; faster reaction rates become possible at increased cell density, higher yields can be achieved, together with easier rheok)gical control and control of slow cell reproduct ion .31 Morikawa et aL 27 were tire first to report the production of penicillin G from glucose with imnlobilized P. ch
406

Enzyme Microb. Technol., 1983, vol. 5, November

ATCC 12690 myceliunl as a catalyst. Mycelimn, harvested from a conventional batch fernlentation at its maximunl rate of penicillin production, was entrapped in polyacrylanfide (PAA) gels as tile preferred procedure. Collagen inmlobilization resulted ill low activity, while calcium alginate entrapment yielded tile highest activity but yielded a gel too fragile to be used under reactor conditions to allow repeated use. With 5% PAA gel-entrapped mycelium, penicillin was produced (0.77 U ml-l ) from glucose, (NH4)2SO4, phosphate buffer (pl-| 7.0) and phenyl acetate as precursor within 5 h, but it amounted to only 17'J of that produced by washed mycelium. However, tile half-life of the immobilized mycelium activity was 6 days. compared with 1 day for washed mycelium. The immobilized mycelium required oxygen for penicillin production. The total concentration and relative ratio of acrylamide and N, N'-methylenebisacrylamide (BIS) determine both the pore size of tile interstitial space within which whole cells are entrapped and the physical properties of the complex. Increasing the content of BIS to 16~. increased the synthetic activity of tile immobilized mycelium and provided excellent nrechanical rigidity. Morikawa et aL 27 obtained evidence that the PAA gel polymerization inactivates the mt, ltienzyme systems in tile viable mycelium, due mainly to toxicity of the monomers and heat evoh, tion during polymerization. Such negative effects have also been encountered by Freeman and Aharonowitz 3° when entrapping viable Streptom.rces clavuligerus cells for cephamycin C production (Figure 4a). In order to retain the advantages of PAA and to avoid cell and enzyme damage during the preparative steps, Freenmn and Aharonowitz 3° recently proposed the use of preformed linear, water-soluble PAA chains (MW 15-. 17 x 104), partially substituted with acylhydrazide groups. This prcpolymerized material was crosslinked in tile presence of viable cells by stoichonletric amounts of bi- or multifunctional aldehydes or ketoncs [glyoxal, glutardialdehyde, poly(vinyl alcohol)]. Use of crosslinking agents of various chain lengths allowed control of gel conrpactness. This crosslinking reaction carried out in cold, neutral, physiological conditions resulted in gel-entrapped cells with mechanical properties similar to those of conventional PAA gels. but no cell damage occurred. S. clartdigerus cells harvested from tile early growth phase (Figure 4h), entrapped by this prepolymerization PAA gel procedure, produced cephamycin C contmuot, sly for 96 h m yields comparable to those of free resting cells. Morikawa et al. 27 obtained only 15~ of the activity of washed Penicillium mycelium with tile conventional PAA entrapment procedure. The concept of using preformed linear, water-solt, ble PAA chains now deserves full attention and might improve the performance of entrapped living-cell systems considerably. However, other non-toxic matrices such as alginate, collagen, agar or carrageenan should be tested further. The economically important antibiotic bacitracin (Figure 5a) is also synthesized following thc multienzymc thiotemplate mechanisnl, which is involved in gramicidin S biosynthesis. 33'34 Cell-free synthesis of bacitracin has also been studied in detail. 34 While gramicidin S is retained ill the cells of its producer, bacitracin is excreted in the medium. Morikawa et aL 28,29 also studied bacitracin fermentation with immobilized cells. Bacillus KY 4515 cells were harvested from a batch fermentation when the rate of bacitracin production was maximal (40 U ml-~ h-J). After inmlobilization in 5'/, PAA gcl, the catalyst was cut into small blocks (8--27 mm 2) and was added to a complex (starch bouillon)

Peptide antibiotic production: Erick J. Vandarnme I

0CH 3 S HOOC -CH- (CHz)3 -CO-NH- ~ "~ HzN

a

L-lie !

0//'~N , ~ ~ CHz0CONH2

i

~LiCys 2

Thiazoline

-L.Leu

= D-GIu

"~ L-IIe---'7"~_L - L y s ~

ring

L..~n

coo.

"F

D-,~p

D-Orn

II L.Ile

il

\o

a

L.ttls._._...

D

,/ he

I00 0 0 v "1o

80

0

(3

0

3°

~, 02

-o 0£_ E

"?_,

60 40

L U

20

3 0

24

~-~-----~

48

72

96

120

o~

0

!

I

I

I

I

2

4

6

8

I

I0

12

~lme (h) Figure4

incubation

Repeated

(a) Structure of cephamycinC. (b) Capacity of Strepto-

(runs)

myces clavuligerus resting cells (A) tO produce cephalosporins as a function of culture age, as compared to the antibiotic formation pattern ('.) during growth (e). 3° [Reproduced from Freeman A. and Aharonowitz, Y. BiotechnoL Bioeng. 1981,23, 2747 by permission of John Wiley & Sons, Inc. @ ]

fermentation medium and incubated at 30°C for 4 h. The initial activity of the immobilized wttole cells was only 20-25% of that of an equivalent amount of washed cells (rate of production 1 3 - 1 8 U ml -] h -1). Again, polymerization reagents (especially acrylamide and ammoniunl persulphate) and limited diffusion apparently lowered the capacity to synthesize antibiotic. Successive use in 1% peptone or 0.5r~ meat extract as a reaction medium resulted in a gradual increase of the activity, reaching a steady state at 80-90% of the activity of freshly washed cells; this increase in activity seems to be caused by active growth of the cells in and on the gel, as could be observed with electron microscopy (Figure 5b). The apparent half-life of this bacitracin-synthesis catalyst was estimated to be at least l week. Continuous production of bacitracin has recently been achieved in an 'immobilized whole-cell fermenter'. 28,29 Maximum bacitracin levels were obtained after I day of fermentation, after which a gradual decrease in productivity was observed. This was caused mainly by cell growth at the surface of the gel blocks. This prevented diffusion of oxygen and peptone into the gel towards the immobilized cells. Periodic washing of the gel (flushing the chemostat with saline for 2 h once each day) solved this problem. Under those operational conditions, continuous bacitracin production remained high for 8 days. Continuous bacitracin production by this immobilized whole-cell fermenter displayed several advantages:

[

I00

i

I

'

80 v -0 Q~ U "0 0 L E U rt3

60

\

40 \~

% r0 c~

,0

20

0

0.2

0.4 Dilution

0.6 rate

0.8

1.0

(h-l)

Figure5 (a) Schematic structure of bacitracin A. (b) Repeated batch production of bacitracin in 1% peptone medium by Bacillus sp. K Y 4 5 1 5 : o washed cells;o immobilized cells. 2~(c) Continuous production of bacitracin: effect of dilution rate on productivity; • conventional fermentation; ~, fermentation with immobilized living cells 29 [Reproduced from Morikawa, Y., Karube, I. and Suzuki, S. BiotechnoL Bioeng. 1980, 22, 1015 by permission of John Wi Iey and Sons, Inc. © ]

Enzyme Microb. Technol., 1983, vol. 5, November

407

Review

Antibiotic bioconversions with immobilized biocatalyst technology

HZN~

COOH Figure6 Schematic structure of thepeptideantibioticnisin

bacitracin is produced from one simple nutrient, peptone: antibiotic recovery is easier since the ct, lture liquid does not contain cells; productivity (even at high dilution rates) is higher than in conventional chemostat culture (Figure 5c). Nisin (Figure 6) is another economically important polypeptide antibiotic compound, produced by Streptococcus lactis. 35 Its biosynthesis appears not to proceed according to the above-mentioned multienzyme thiotemplate mechanism and follows protein synthesis and mechanisms. It is active against Grant-positive organisms, including related streptococci. Nisin resembles a bacteriocin rather than an antibiotic. Among antibiotics, nisin has the unique function of being exclusively used as a biological food preservative. It is non-toxic and, being a polypeptide, any residues remaining in tbod are digested. In pH-controlled (pH 6.0) batch fermentations at 20 to 30°C withot, t aeration, maximum nisin synthesis is related to maximum biomass formation. Nisin is normally excreted into the culture medium, but when cells are grown at pH 6.7 they retain the synthesized nisin. Media are based on corn extract, yeast autolysate and dairy waste. Maximum production can attain 2000 U ml -t, or 50 ~g ml -t. Nisin production by polyacrylamide gel-immobilized Streptoeoectts lactis cells has been reported by Egorov el aL 32 but few details of the process are available. In view of its use in foods, cell immobilization systems based on agar, gelatin, alginate or carrageenan should be used for such immobilized-cell fermentation processes. Another antibiotic produced by immobilized cells is the polyene macrolide candicidin, produced by Streptomyces griseus. Innnobilization of the streptomycete in collagen membranes resulted in an antibiotic production level from glucose of 14~ of that obtained m a conventional batch fermentat ion process. 3 Other antibiotics recently reported to be produced with immobilized whole cell biocatalysts include colistin, patulin, thienamycin and nikkamycin. These systems are typical exampies of synthesis by fixed viable cells of secondary metabelites, normally non-growth associated complex fermentation products. It already seems that the physiological state, culture age and viability of the cells in the immobilized reactor are of prime importance for effecting such complex multienzyme reactions (whereas in the case of single-step reactions viability of the cells is not mandatory). Indeed, it might seem necessary to control cell reproduction at a certain low level within the immobilized matrix to obtain nmximum productivity. In this perspective, new bioreactor configurations, such as the immobilized whole-cell fermenter described by Morikawa et al.28 for bacitracin production, could be designed combining batch or continuous fermenter and immobilized-cell reactor characteristics, thus facilitating synthesis of both growthas:;ociated as well as non-growth-associated products.

408

Enzyme Microb. Technol., 1983, vol. 5, November

Immobilized enzyme or immobilized cell technology to convert antibiotics using single-step reactions into useful derivatives is largely focused on the production of 6-aminopenicillanic acid (6-APA), the penicillin nucleus, from penicillin G or penicillin V, and the production of 7.aminocephalosporanic acid {7-ACA) and 7-aminodeacetoxycephalosporanic acid (7-ADCA), both cephalosporin C-nucleus compounds, s'e The compounds 6-APA, 7-ACA and 7-ADCA can be acylated to yield senaisynthetic antibiotics with improved performance. Stereospecific side chains to be coupled to these compounds can also be pro,"h, ced using immobilized enzyme or cell technology Apart from the penicillins and cephalosporins, several other/3-1actam antibiotics (thienamycins, noca,-dicins) as well as the novel monocyclics, the monobactams and the bicyclic/3-1actams produced by bacteria might serve as substrates for enzymic conversion. All these antibiotic conversions are simple single-step enzyme reactions, amenable to immobilized biocatalyst technology. So far, only 6-APA is produced enzymically on a large scale, but several other conversions (7-ADCA production and side chain resolution) are on the verge of large-scale Operation.

Enzj'mic Jbrmation and reacylation o f 6-APA Bioconversion of penicillins into 6-aminopenicillanic acid (6-APA), the penicillin nucleus, and its side-chain acid (Figatre 7) is a very important bioconversion reaction. The compound 6-APA is the starting material for the industrial production of the semisynthetic penicillins with superior therapeutic action. 4,6,36,37 The worldwide demand for these semisynthetic penicillins has brought 6-APA into a central position as a major pharmaceutical product. As many as 16 different semisynthetic penicillins, all derived from 6-APA, are in widespread clinical use today. 37 Bulk 6-APA production can now be achieved either by chemical 38,39 or enzymic hydrolysis. 6 The advantage of either process has depended so far upon technok)gical advances in individual companies, and each system has its merits, ltowever, with the introduction of methods of immobilization of enzymes or cells that promote stability and prok)nged high activity that can result in reuse and continuous bioconversion, and with high energy prices, the enzymic splitting currently appears to display better economics. The microbial enzymes that hydrolyse penicillins into 6-APA or acylate 6-APA have been named penicillin acylases (EC 3.5.1.1 1). Compared to other industrial enzymes, information about penicillin acylase is less available. It is sold on the industrial enzyme market (Novozym 217®). 40 but it is mainly used in-house by the manufacturers. Furthermore, 6-APA tormation from penicillin is, in fact, a side reaction of this acylasc enzyme, since its basic physiological function in the producer cells is not clearly understood. 36 It is indeed a rare exalnple of an unconventional catalytic action of an enzyme industrially exploited for many years without knowing its conventional (natural) function. Penicillin acylase activity has now been demonstrated to occur in a wide range of bacteria, actinomycetes, yeasts and moulds. 6'36 Penicillin V, penicillin G and ampicillin acylase are the three types of acylase now clearly recognized (Figure 7). At the moment, highly productive (genetically engineered) penicillin acylase strains are used to produce 6-APA on an industrial scale, starting from either penicillin

Peptide antibiotic production: Erick J. Vandamme l~co t;ll

s

CH3 H.

S

CH]

•

RC(X)H

- H20 o~

• --"

~C(X)H

6-APA

Penicillin

Carboxylic acid

Benzvlpenicillin (G)

R

-

~'~H Z ~k

Phenoxyme'thvIpenicillln (V)

(-) a - A m i n o b e n z y l p e n i c i l l i n (ampicillin)

R = HI3

CH--

p-Hydroxy-a-aminobenzylpenicillin (amoxycillin)

Figure 7 F o r m a t i o n and reacylation of 6-aminopenicillanic acid (6-APA)

G (benzylpenicillin) or penicillin V (phenoxymethylpenicillin). Fermentation processes for penicillin acylase production have been reviewed recently. 6'41 Characterization and properties of these enzymes have recently been summarized by Vandamme. 4'6'36 Penicillin V acylases are highly specific enzymes with practically no activity towards penicillin G or other compounds, whether found in moulds (Fusarium semitectum, Pleurotus ostreatus, P. chrysogenum) 6 or in bacteria (Erwinia aroMeae,6 Pseudomonas acidovorans, 42 Bacillus sphaericus43). Penicillin V acylases produced by yeasts and actinomycetes have not yet been purified, but they all display preferential hydrolysis of the penicillin V molecule. Penicillin V acylases are mainly intracellular enzymes, with optimal hydrolytic activity at pH 7 - 8 , with a few exceptions known. None of the known purified fungal or bacterial penicillin V acylases displays synthetic activity. Penicillin G acylases, in contrast to the penicillin V acylases, display a rather broad substrate specificity, and these enzymes can be considered as aspecific deacylating enzymes.6 The Escherichia coil ATCC 9637 and ATCC 11105 acylases, which are the best studied, are specific for phenyl-acetylated compounds. Penicillin G acylases from actinomycetes and fungi have not yet been fully characterized but, in addition to penicillin G, they all hydrolyse a range of N-phenylacetyl-t,-a-amino acids. Penicillin G acylases are mainly intracellular enzymes, with optimal hydrolytic action at pH 7 - 9 . A few strains of Streptomyces and Bacillus megaterium are known to produce the enzyme extracellularly. Many bacterial penicillin G acylases also act at acidic pH values (4.5-5.5) to catalyse the synthesis of penicillins from 6-APA and phenylacetic acid or its derivatives. To take advantage ot" this synthetic activity of penicillin G acylases, the enzymic synthesis of many semisynthetic penicillins (and cephalosporins) has been achieved .6,44,45 Ampicillin acylases have so far been found only in Pseudomonas melanogenurn and P. ovalis strains. They neither hydrolyse nor synthesize penicillin G, penicillin V or acylated conlpounds.46

The industrial enzymic route for 6-APA production was developed around 1960. Initially, cell suspensions of active strains could only be used once and the productivity was very low, about 0.5 to 1 kg of 6-APA per kg t:'. coli suspension. Novel methods to immobilize E. coil cells increased productivity up to 50 kg of 6-APA per kg of immobilized catalyst. Beechams applied a poly(methacryl) glutaraldehyde method, while Pfizer used binding to glucidyl methacrylate polymers. Bayer used a cyanogen bromide-activated dextran method and Astra a cyanogen bromide-activated Sephadex method. Current productivities are within the range 100-250 kg 6-APA kg-~; however, no data have been published by the companies involved. It can only be assumed that ~60% of all 6-APA produced today is made by tile immobilized catalyst route. This means that about 3000 ton 6-APA is produced by 15-30 ton immobilized penicillin acylase catalyst.47 Industrial 6-APA producers include Bayer, Gist-Brocades, Beechams, Glaxo. Astra, Hoechst, Biochemie, Antibioticos and Snare Progetti in Europe; Pfizer, Bristol-Myers, Squibb, Wyeth in the USA; Kyowa Hakko Kogyo and Toyo Yozo (Japan) and Yung Jin Pharmaceuticals (S. Korea) in the Far East. Numerous immobilization techniques for penicillin acylase enzyme or cells have been tried out on a laboratory scale. Adsorption, crosslinking, covalent and physical attachment and entrapment methods have been used. Both soluble and insoluble carriers have been tested. A comprehensive survey of these aspects has been compiled recently by the author. 7 Industrial application of immobilized penicillin acylases is still confronted with problems. Upon penicillin hydrolysis into 6-APA at pH 7.0-8.0 at 37°C, the side-chain acid is liberated, which causes a drop in the pH and this pH change results in a slower reaction rate. A higher starting pH is not desired because of 13-1actam ring hydrolysis and penicillin inactivation. A strict pH control is thus necessary during this bioconversion process, which is therefore difficult to run in a continuous packed-bed reactor. In industry, a batch reactor is generally used. Such a process is then faced with another aspect of the catalytic properties of the enzyme, namely, end-product inhibition.

Enzyme Microb. Technol., 1983, vol. 5, November 409

Review PerlcdhnG Alkahfrom SupPlytank ('1%

=I

=%,

ent PeClC,U,nG

Alkalifrom~a,pplyla~k : ..... pH ¢onlroli

Watero,lt

i ~, ~

~ Waterout

,(,

Ar n •

L ......

t,o,

[] pH

I emperature

~

'$,;2,'L

tO

6-APAextracIio~

,

il

Figure 8 Industrial immobilized biocatalyst reactors for 6-aminopenicillanic acid (6-APA) production. (a) Stirred tank reactor with external recovery of immobilized enzyme. (b) Recirculation reactor: (i) high aspect ratio column; (ii) low aspect ratio column (reproduced by kind permission of T. A. Savidge)

So, in this particular case, it is in fact the nature of the biochemical reaction that determines the type of reactor and reactor controls to be used. Conventional stirred-tank reactors are used with adaptations for immobilized biocatalyst retention or recuperation; a less shear-intensive agitation system is usually also needed to avoid mechanical damage to the enzyme or cell carrier as well as an efficient systeln for the addition and distribution of alkali to avoid local alkaline inactivation of enzyme and penicillin, and to obtain strict pH control (Figure 8a). If this reactor type still fails because of biocatalyst-carrier limitations, a recycle reactor-column combination could be used (Figure 8b). In such a system, pH adjustment occurs separately from the immobilized enzyme. A low aspectratio column is preferred to minimize the residence time in the column to prevent an excessive pl] fall. The column can be packed with bead-entrapped or fibre-entrapped penicillin acylase enzyme or cells. Relatively little information is available on the acylation of 6-APA to produce semisynthetic penicillins using immobilized biocatalysts,v The penicillin G acylases from k: coli ATCC 9637, Bacillus circulans, B. me~aterium and the ampicillin acylase from Pseudomonas melanogenum are able to synthesize ampicillin or amoxycillin from 6-APA and D-phenylglycine methyl ester, or p-hydroxyphenylglycine methyl ester, respectively. 49-s2 These acylases display a distinct preference, although not an absolute requirement, for the D-configuration of the side-chain molecule. The development of a commercially feasible penicillin acylase process for enzymatic synthesis of semisynthetic penicillins from 6-APA is still unlikely due to the low conversion ratios, difficult product separation, high cost of the side-chain and the availability of relatively simple chemical acylation procedures.

7-ACA nucleus thus depends upon removal of the ouaminoadipic acid side chains. This can be accomplished with reagents such as PCIs or nitrosyl chloride. Enzymic hydrolysis to yield 7-ACA has also recently been claimed (Figure 9). Chemical as well as enzymic acylation procedures can be used to couple other side chains to 7-ACA. Acylation of 7-ACA with 2-thiopheneacetic acid derivatives yields the therapeutically important cephalothin; acylation with D-phenylglycine methyl ester yields cephaloglycm. There has indeed long existed doubt about the occurrence of microbial acylases capable of hydrolysing cephah)sporin C specifically into 7-aminocephalosporanic acid (7-ACA) and D~-aminoadipic acid. s3-s6 However, recent Japanese patents claim the direct enzymatic hydrolysis of cephalosporin C into 7-ACA and (~-aminoadipic acid with Pseudomomzs putida, sT"s8 An interesting two-step enzymic process for 7-ACA production has also been proposed recently by Fujii e¢ al. ;s9 cephalosporin C is first transformed by Trigonopsis variabilis CBS 4095 into glutaryl 7-ACA, which is then hydrolysed by Comamonas or Pseudomonas sp. into 7-ACA (Figure 10). These researchers, from Toyo Jozo Cy in Japan, also reported in detail on the isolation and properties of

Pseudomonas (Pseu&mlonas putida, Pseudomonas SY-77-1) strains, able to deacylate 7-/3-(4-carboxybutaneamido)cephalosporanic acid (glutaryl 7-ACA) into 7-ACA. These strains specifically hydrolysed cephalosporin compounds having aliphatic dicarboxylic acid in the acyl side chain, 6° cephalosporin C was not hydrolysed. The substrate. glutaryl 7-ACA, can be obtained by oxidative deamination from cephalosporin C by D-amino acid oxidase from fungi

H00C ~-CH (CHz)3CONH " l ~

HN

Cephah)sporin-related bioconversions Formation and reaeylation of 7-ACA. Cephalospirins are penicillin analogues that can be synthesized by direct fermentation with Acremonium chrysogenum (Cephalosporium acremonium) or by chemical ring expansion of penicillin. 2° The major fermentation product, cephalosporin C, contains the 7-aminocephalosporanic acid (7-ACA) nucleus and the side chain a-aminoadipic acid (Figure 9). 7-ACA derivatives with other side chains are not easily obtained by fermentation. The synthesis of semisynthetic cephalosporins containing the

410

Enzyme Microb. Technol., 1983, vol. 5, November

S Cephalosporln S ""1 acylose HzN.,_r/ -h

I L l - -

o'-'y%oco% COOH

"

o

1 d

l

COOH 7-ACA

Cepholospcrln C +HOOC~.cH(CH2) COOH H2N a ct- Aminoodipic acid Figure 9 Deacylation of cephalosporin C. Conversion of cephalosporin C into 7-aminocephalosporanic acid (7-ACA) and e-aminoadipic acid by cephalosporin C acylase

Peptide ant~biotic production: Erick J. Vandamme

HOC'C-,:;;n2)s .CO-n H]----~-'S'-] O"~'~ N . , ~ C

0

H2OCOCH3

()/~N ..~C

COOH

i is

ao,,la~-"

COOH

Glutaryl - 7- ACA 7- ~:,- (4 - cor boxylb ulanomido ) cepholosporonic ocid

7- Phenylacetamido. amitlodeacetox~cephalo.

sp(}[ anl( acid

i ...~ .CH.COOH tt:N i ~ / / S " ,

ococ.

+

~

o ~ , _ _ ~ c .

'

(':0OH

COOH

Phen~I aCCtl+,acid

7-Aminocepholospomnic acid (7-ACA)

7- Ammodea(etox'!t~ph Io-

sporaru( a(ld

O

Figure 10 B i o c o n v e r s i o n o f 7 - ~ - ( 4 - c a r b o x y b u t a n a m i d o ) c e p h a l o sporanic acid ( g l u t a r y l 7 - A C A ) i n t o 7 - A C A b y g l u t a r y l 7 - A C A acylase

a,.la,.

COOH 7- Phcnoxxa{etanndo-

and pig kidney. 59,61 Glutaryl 7-ACA has also been isolated from Cephalosporium fermentation broths. 62 Beta-lactamase deficient and acylase constitutive mutants were derived from the Pseudontottas SY-77-1 strain with improved activity.63 Ichikawa et aL 64 crystallized the glutaryl 7-ACA-acylase from cell-free extracts from Pseudomonas. The molecular weight of this periplasmic enzyme was estimated to be 130000 by Sephadex G-IO0 gel fihration. Maximum activity was found in the broad pHrange from 6.5 to 10.0 at 37°C. Glutaryl 7-ACA acylase has also been claimed to be produced by Bacilh/s, A rthrobacter and A lcaligenes species.6s-68 Immobilized enzyme or cell technology should soon be applied to this new type of acylase and it might compete with conventional chemical methods fbr 7-ACA production, which can, in turn, be converted into the clinically important cephaloglycin and cephalothin (see Figure 12). These coupling reactions have already been performed with the Celite-adsorbed penicillin G-acylase from B. megatcrh#n. 69,7° An imnaobilized cell system has been described by Fukushima et al. ,71 where cells o f Psettdontonas sp. or G)mamoltas sp. entrapped in cellulose acetate capsules hydrolysed glutaryl 7-ACA into 7-ACA. Formation and reacylation of 7-ADCA. Cephalosporins can also be produced from the precursors penicillin G or penicillin V by a series of chemical reactions that expand the five-membered thiazolidine ring of the pcnicillins into the six-membered dihydrothiazine ring of the cephalosporins. The cephalosporins obtained contain the penicillin G or V side-chain, which can then be removed chemically or enzymically by conventional penicillin G or V-acylase action. The cephalosporin nucleus obtained is named 7-aminodesacetoxycephalosporanic acid (7-ADCA) (Figure 11). Acylation of 7-ADCA with D-phenylglycine produces the useful antibiotic cephalexin (l~'igure 12). In contrast to enzymic acylation, chemical acylation procedures need blocked derivatives of the reactants. This fact, combined with the advantages of immobilized enzyme or cell technology, could favour the use oJ"enzymic procedures in semisynthetic cephalosporin synthesis. So far, most studies of enzymic deacylation of cephalosporins are directed towards producing the 7-ADCA nucleus (Figure 11). The substrates for this reaction, 7-phenyl and 7-phenoxyacetamidoacetoxyccphalosporanic acid are readily obtained by ring expansion of the precursors penicillins G or V and are easy substrates for penicillin G

annnodea( etoxv( ephalo -

sporanic acid

O-"~'-N - x ~ C H 3 COOH Phenoxx aceu( "aod

}'-Ammodeacemx',(ephalosporam( actd

Figure 11 H y d r o l y s i s o f cephalosporins i n t o 7 - a m i n o d e a c e t o x y cephatosporanic acid ( 7 - A D C A )

O " ~ N ' - . ~ C H,

•

NH 2

acyla~" ~

COOH

D- Phenylglycine

7 -Aminodea(etoxycephalosporanic acid

medoI ester

H~N

O

o~N",.~CH, COOH Cephalexin + H ' N" ~ " ~ % sr~cH

O

~C~:cH~

2- Th lop henacetic

attd methvl ester

COOH 7-Ammocephalo. sporanic acid

COOH

Cephalothin

[ ~'~-CI.IC

I

~O

""OCH

NH~ D- Phen~Aglycme meth'~l ~ter

COOH 7-Amino<¢phalosporanic acid H~N O

] II

[~'---CHCNH ' ~ S h

O

COOH Cephalog|'.on

Figure 12 Synthesis of or 7 - A D C A

semisyntheticcephalosporinsf r o m

7-ACA

Enzyme Microb. Technol., 1983, vol. 5, November

411

Review

or V acylases. Several penicillin acylases have been described that readily deacylate 7-ADCA derivatives and reacylate 7-ADCA into cephalexin. hnmobilized biocatalysts have been used for the deacylation (or acylation) of such cephalosporin compounds. Extracellular penicillin G acylases from E. colt, B. megaterium and P. rettgeri, adsorbed onto Celite, activated carbon, carboxymethylcellulose or an Amberlite ion-exchange resin hydrolysed 7-phenylacctamido-ADCA.69'7°'72 The Celite-adsorbed enzyme was selected for pilot-plant production of 7-ADCA, yielding 89.2% hydrolysis of the substrate. Upon addition of toluene, to minimize bacterial contamination, the immobilized system had a useful life of ~10 days. The partially purified penicillin-V acylase as well as intact cells of 1:] aroideae 73 were entrapped in cellulose triacetate fibres at Glaxo Laboratories, UK; a fibre-packed column hydrolysed 7-phenoxyacetamidoADCA to 7-ADCA in 58% yield after 3 h of substrate circulation at 37°C. 74 Several imnmbilized cell systems for the synthesis of the clinically useful antibiotic cephalexin have been developed 7s (Figure 12). Acetone-dried Achromobacter, Beneckea hyperoptica, A lcaligenes faecalis and Flavobacterium aquatile cells, adsorbed on to DEAE, TEAE or carboxymethylcellulose could synthesize cephalexin from 7-ADCA and i)-phenylglycine methyl ester. The highest yields (80-85%) were obtained with the Beneckea DEAE-cellulose complex. The fibre-entrapped F,. coil acylase used by Marconi et aL 49 to acylate 6-APA could also be applied to cephalexin synthesis from 7-ADCA and D-phenylglycine methyl ester, with a yield of 75(~,. The extracellular acylase from B. megaterium adsorbed onto Celite, and a cell-free Achromobacler acylase adsorbed onto hydroxyapatite performed the same reaction, in 80 -85% yield. 76

Deacylation of 7-ACA and cephalosporins with acetyl esterase. In view of the synthesis of new cephalosporin derivatives, deacylated cephalosporins are useful intermediates. Indeed, the antibk)tic properties of cephalosporins are determined not only by the side chain at the C-7 position but also by the substituent at the C-3 position. After removal of the acetate group, many substituents can be added to the C-3 position of 7-ACA. Acetate can be chemically removed, but yields are low due to unwanted side reactions, such as lactone formation and doublebond migration. At the moment, enzymic hydrolysis is the only efficient method to prepare deacetyl 7-ACA (Figure 13). The enzyme involved has been named acetyl esterase and occurs in mammalian tissue7 v plant tissue, 78,79 in bacteria, actinomycetes and yeasts 8°'8t and in fungi, s6 Several processes that use cell suspensions or soluble enzyme preparations have been described for the deacetylation of cephalosporins. The production, isolation and purificat ion of Bacillus subtilis NR R L B-558 cephalosporm acetylesterase, splitting 7-ACA into acetate and deacetyl 7-ACA. has been described by Abbott and Fukuda. 82,83 The enzyme (tool. wt 190 000) proved to be extremely stable in solution and very resistant to thermal inactivation. The extracellular esterase was absorbed onto bentonite H2N~ -7"8 "~ '~: .. 0 0/2--==-N.~.S')-.. CHtO~C Hs C'OOH

cephalosporln

HtN. _._/-S. I

mO • C'H~C~oH

COOH

Figure 13 Conversions of 7 - A C A i n t o d e a c e t y l 7-ACA b y c e p h a l o sporin acetylesterase

412

EnzymeMicrob. Technol., 1983, vol. 5, November

(I) $ynthests of cepholexm

COOH

l

COOH

(2) Hydrolysis of phenylglycine methyl ester

NH2

~ N H

/ (3) Hydrolysis of ~ cepholexin

2

0 ~ 1 ~

CH3

C00H Figure 14 Reactionscatalysed by a-amino acid ester hydrolasein the case o f cephalexin

and used for muhiple batch reactions to deacetylate 7-ACA solutions. During the reaction, the enzyme dissociated from the bentonite particles. The complex could be partially stabilized by addition of aluminium hydroxide gel. The physical characteristics of the enzyme also made it ideally suited to immobilization by containment within an ultrafiltration device. With this technique, the enzyme was reused 20 times over an 11 day period at pH 7.0 and 25°C, in 94% yield. Spores and vegetative cells of mycelium of Fusarium oxysporum transformed cephalothin, 7-phenylacetamidocephalosporanic acid (benzylcephalosporm) and phenoxymethylcephalosporin at pH 7.5-8.0 into the corresponding deacetylcephalosporins,s6

a-Amino acid ester hydrolase bioconversions. In addition to penicillin acylases, enzymes called ~-amino acid ester hydrolases are able to synthesize cephalosporins, starting from 7-ACA or 7-ACDA and an appropriate side chain ~-amino acid ester, usually O~-D-phenylglycinemethyl ester and analogues. 84'8s The enzyme involved in these synthetic reactions utilizes a-amino acid esters are acyl donors and transfers these acyl groups to both water (hydrolysis) and 7-aminocephem compounds or 6-APA (transfer). It has been found chiefly in microorganisms belonging to the Pseudomonadaceae (Pseudomonas melanogenum, P. maltophila, Xanthonumas orvzae, X. citri, Acetobacter pasteurianus, A. turbidans, A. xylinus and Ghwonobacter. The enzymes f'rom Acetobacter turbidans ATCC 9325 and Xanthomonas citri IFO 3835 have been studied in detail.86-91 The X. citri IFO 3835 enzyme catalysed both hydrolysis of a-anaino acid esters (such as D-a-phenylglycine methyl ester) and cephalexin, but it could also synthesize cephalexin from 7-ADCA and D-phenylglycine methyl ester in high yield (Figure 14). 7-ACA and 6-APA can also be utilized as acyl acceptors. The optimum pH for these reactions was 6.4 at an optimum temperature of 35°C. The K m for cephalexin deacylation was 2.99 raM. With washed cells, the conversion rate of 7-ADCA to cephalexin was ~5% within 2 h. Amoxycillin could be synthesized from D-~-p-hydroxyphenylglycine methyl ester and 6-APA in a similar fashion, although the rate of synthesis was slower. Yields could be further improved by t,sing a penicillinase-deficient mutant strain K24 and by the addition of alcohols (s-butanol) to the reaction mixture, which lowered the rate of hydrolysis of the ester and increased that of the acylation of 6-APA. The conversion rate of 6-APA to amoxycillin was ~93% (at pH 6.8 and 20°C after 15 h incubation).

Peptide antibiotic production: Erick J. Vandamme

The enzyme from A. turbidans ATCC 9325 has been immobilized to curdlan and used for continuous synthesis of cephalexin. 87 The crude enzyme preparation, obtained after sonication of washed cells, and CNBr-activated curdlan 13140 were mixed at pH 8 and 5°C and after 20 h 93% of the enzyme activity was immobilized. For cephalexin synthesis, a solution (at pH 7.0) containing 7-ADCA (0.5%) and D-phenylglycine methyl ester hydrochloride (1.5%) was passed (240 ml/day) through a column packed with the immobilized enzyme. The reaction temperature was kept at 5°C to aw)id bacterial contamination and inactivation of the enzyme. At the start of the reaction, the conversion rate of 7-ADCA to cephalexin was ~85%; this ratio did not change substantially even after 70 days of continuous reaction. Enzymatic coupling of side chains to the cephalosporin nucleus is as yet not competitive with chemical procedures; however, the use of the a-amino acid ester hydrolase rather than the conventional acylase looks promising. The development of efficient immobilized biocatalysts might accelerate the industrial exploitation of this enzyme activity.

chemical transformation of penicillins G or V into the corresponding 7-ADCA derivatives and subsequent side chain hydrolysis, should also receive attention. An overall reaction scheme of real and potential bioconversions of penicillins and cephalosporins is given in Scheme 1.

Penicillin and cephalosporin side chain preparation with immobilized biocatalyst technology Microbial enzymes are now used not only for bioconversion of penicillins and cephalosporins, but also in the manufacture of certain side-chain acids to be coupled to 6-APA or to the cephalosporin nucleus. Enzymic resolution processes have been developed for producing n-phenylglycine used to prepare ampicillin, cephalexin and cephaloglycin, and D-p-hydroxyphenylglycine to prepare amoxycillin. Chemical procedures to resolve the racemic mixtures are also known, but enzymatic processes now attract industrial attention and have recently reached the production stage. Immobilized E. coli penicillin G acylase stereospecifically hydrolysed only the L-isomer of N-phenylacetyl-DL-hydroxyphenylglycine. Partially purified enzyme was immobilized onto an acrylic ester resin, Amberlite XAD7, and packed into a column reactor through which a 5% substrate solution was continuously passed. The unhydrolysed n-derivative was recovered by solvent extraction and subsequently chemically hydrolysed into D-p-hydroxyphenylglycine.9z A Pseudomonas putida amidase stereospecifically hydrolysed

Interconversions o f 6-APA, 7-ACA and 7-ADCA Interesting potential bioconversions would comprise the hydrolysis of 7-ACA into 7-ADCA or the formation of 7-ACA through coupling of 7-ADCA with an acetoxy derivative. However, enzymes able to carry out these reactions have not yet been reported. Direct (chemical) conversion of 6-APA into 7-ADCA or 7-ACA, rather than

Glucose etc.

E.C

Penicillin G or V

Conventional fermentation

C E

~ 6-APA zz S

~-- Semisynthetic penicillins (ampicillin, amoxycillin etc.)

/ ,/ C

) / / //

Immobilized living cell fermentation

/

/

/

7-ADCA

E.C

Semisynthetic cephalosporins (cephalexin)

A I Immobilized cellparticle (vesicle) fermentation I I I I I I

Total (immobilized) enzymic synthesis

I I

,I Cephalosporin C

C E

~- 7-ACA

Glutaryl 7-ACA

E.C

Deacetyl 7-ACA

~.- Semisynthetic cephalosporins (cephaloglycin, cephalothin)

....... , Hypothetical reactions --C, Chemical reaction --E, Enzymic reaction Scheme 1 Interconversion of penicillin and cephalosporin compounds

E n z y m e M i c r o b . T e c h n o l . , 1983, vol. 5, N o v e m b e r

413

Review L-phenylglycinc amide in a racemic mixture, the remaining D-phenylglycine amide was then chemically hydrolysed into optically pure l)-phenylglycine.93 A protease, subtilisin. covalently immobilized onto methacrylate copolymers, also stereospecifically hydrolysed I 0%. N-acetylq)L-phenylglycine methyl ester to give N-acetyl-t.-phenylglycine. The problem is that these enzymes are specific for the L-isomer, which necessitates the chemical hydrolysis of the unaffected D-derivative into the desired compound. It wonld thus be advantageous if the D-isomer could serve as the direct enzyme substrate. Such n-aminoacylases have indeed been detected in Streptomyces olivaceus and S. tuirus.94'95 N-Acetyl-DL-phenylglycine (2%) was completely hydrolysed within 6 h at pH 7.0 and 30°C into optically pure macid. But in this case, the unaffected L-derivative has to be racemized before being reused. A novel process, developed and/or applied by Sna,n Progetti, Italy, Kanegafuchi Chemical Ind., Ajinomoto Cy, Japan. and others could avoid the above-mentioned problems. A dihydroxypyrimidinase (or hydantoinase) t'rom calf liver or from Pseudomonas strains was found to stereospecifically hydrolyse m~-p-hydroxyphenylhydantoin to a D-carbamoyl derivative, which upon chemical hydrolysis yielded the desired D-p-hydroxyphenylglycine: the remaining L-phenylhydantoin spuntaneously racemized such that the bL-mixture was finally totally converted to the D-compound.96 An alkalophilic Bacilhts sp. trapped in polyacrylamide gel hydrolysed stereoselectively DL-phenylhydantoin into N-carbamoylD-p-phenylglycine. which is further chemically hydrolysed to D-phenylglycine; under alkaline conditions, the remaining L-hydantoins racemize spontaneously. N-Carbanryl-Dthienylglycine and -p-hydroxyphenylglycine were produced in a similar fashion. 97 Further developments led to the isolation of strains (i.e. Agrobacterium radiobacter NRRLB 11291) with D-hydantoinase as well as D-carbamoylase activity, thereby avoiding the need for the complex chemical hydrolysis of the D-carbamoyl derivative.98 Such a process allows the complete conversion of the racemic substrate to the desired l)-amino acid in a single reactor. Resting cell suspensions could be used repeatedly up to six times without appreciable loss of activity (FigTtre 15). A Flaw~0

HO~

I .C - - NH

~CHC..

DL-p- Hydroxyphenylhydantoin

I

NH-- C%O Spontaneous

A~robocterium,

rocemizofion I Pseudomonos / O ~' Hydantoir~se' HO(~ /k(*)

CIOOH / C - - NH ~--CH.~ I + HO/~ ~_CH_NH_CO_ NH2 D(-)-- N - CarbQmoyl-p-hydroxyphenylglycme Chemical hydrolysis °'rcarbamoylose' A~0bocterium HO<>CHCOOH

+ CO2 + NH 3

2 o ( - ) - p - Hydroxyphenylglycine Figure 15 Enzymatic resolution of DLlO-hydroxyphenvIhydantoin into C)-p-hydroxyphenylglycine with hydantoinase and carbamoylase from Agrobacterium sp.

414

Enzyme Microb. Technol., 1983, voL 5, November

It

H

Nocardicin C

1

OIl

H2NI~ O"

3-AN/t Figure 16 Bioconversion of nocardicinCintoits nucleus3-aminonocardicinic acid (3-ANA)

bacter&m hydantohTophilum strain has been reported to carry out these reactions and also to couple the side chain thus formed to 6-APA, with formation of amoxycillin in one single run! It is clear that these important bioreactions are candidates [br immobilized enzyme or cell technology on an industrial scale. Bioconversion o J m ) v e l [3-lactam atttibiotics Several new natural/3-1retains have been described, including cephamycins, clavulanic acid, thienamycins, olivanic acids, nocardicins, carpetimycins and asparenomycins.99,1oo Many of these natural compounds have been reported to undergo transformation by microbial cell or enzyme preparations (Figure 16). I°l Pseudomonas schuylkillensis acylase hydrolysed the single/3-1attain nocardicin C at pH 8 and 37°C into 3-aminonocardicinic acid (3-ANA)) o2 The olivanic acid-related PS-5 antibiotic was deacylated by Pseudomonas sp. 1158 cells, immobilized in polyacrylamide gel 1°3 at pH 7.4 and 30°C and by L- and D-amino acid acylases on DEAE-Sephadex. An E coli acylase has been reported to acylate thienamycin reversibly. I°l Novel monocyclic/3-1actam antibiotics, sulfazecin and isosulfazecin, have been detected in acidophilic bacteria, identified as Pseudomonas acidophila and P. mesoacMophila.l°4 These compounds are now classified as nronobactams, l°s Similar compounds differing in side-chain structure at the 3-amino group, have been found in Ghwonobacter sp., Chromobacterium violaceum, and Agrobacterium radiobacter (Figure 1 7). In analogy with the well-known fi-lactam antibiotics, the nucleus of the monobactams, 3-aminomonobactamic acid (3-AMA), has been prepared chemically, but it should also be possible in an enzymatic way. This in turn might lead to immobilized enzyme or ceil technology to prepare scmisynthetic monobactams. Highly active/3-1actamase stable derivatives have already been synthesized chemically, l°s Recently, a bicyclic 13-1attain, SQ 27860, which shows activity in vitro against a broad spectrum of Gram-positive

Peptide antibiotic production: Erick J. Vandamme NH

CH .

o

.

,,

-7, ,

o 0

land, and to P. Poulson, Novo Industri A/S, Denmark, for providing pertinent information prior to publication. The secretarial assistance of A. Veys and of my wife Mireille is greatly appreciated.

"SO~ H

Sulfozecin

References

HN 2

1 2 S03H 3-AMA Figure 17 Structures of themonobactam sulfazecin and of its nucleus, 3-aminomonobactamic acid (3-AMA)

3 4 5 6 7

COOH

Figure 18 Structure of the bacterial bicyclic B-lactam, SQ 27 860 and Gram-negative bacteria, has been discovered in Serratia and Erwinia species. Its structure, represented in Figure 18, might also be particularly amenable to enzymic conversion, especially in view of its very unstable chemical nature. Io6

8 9 10 11

12 Conclusions In contrast with immobilized mono-step biosystems for antibiotic production already operational on a large scale (6-APA production, side chain resolution), industrial biocatalysis by multi-step enzyme systems or by immobilized cell fragments or living ceils remains a goal. The complexity of the pathways involved and the many basic aspects yet to be well understood hamper a quick translation of this concept into economic reality. Indeed, microbiological, biochemical, physiological and technical problems inherent to this immobilized biocatalyst principle, need further study if one is to arrive at an optimal performance of immobilized catalyst bioreactors, a,24,31,107 Physiological state, viability, lysis or growth phenomena. cell metabolism and maintenance energy at very low growth rate, cofactor utilization and regeneration, microbial contamination problems and prevention of unwanted side reactions need further study. Other important process design parameters, which determine overall cell reactor productivity, are enzyme (cell) loading factor, stability of the cells, biocatalyst packing density, oxygen transfer, mass transport and diffusion efficiencies and residence time distribution. For most complex multienzyme processes, there is as yet no economic incentive to replace well-established fermentation processes by immobilized-cell reactor technology, although emphasis is clearly focused upon switching from free or immobilized single-enzyme reactions to immobilized (living) cell processes. Complex fermentations, as for the production of antibiotics and other secondary metabolites, offer great potential for immobilized biocatalyst technology on an industrial level in the near future.

Acknowledgements The author is greatly indebted to T. A. Savidge, Beecham Pharmaceuticals, UK, to J. Konecny, Ciba-Geigy, Switzer-

13 14 15

16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

Demain, A. L. Science 1983, 219, 709 Biotechnology ofhzdustrialAntibiotics (Vandamme, E. J., ed.) Marcel Dckker Inc., New York, in press Berdy, J. Adv. Appl. MicrobioL 1974, 18,309 Vandamme, E. J. Adv. AppL Microbiol. 1977, 21, 89 Sebek, O. K. in Economic Microbiology (Rose, A. It., ed.) Academic Press, New York-London, 1980, vol. 5, p. 575 Vandamme, E. J. in Economic Microbiology (Rose, A. H., ed.) Academic Press, New York-London, 1980, vol. 5, p. 467 Vandamme, E. J. J. Chem. TechnoL BiotechnoL 1981,31, 637 Klibanov, A. M. Science 1983, 219, 722 Vournakis, J. N. and Elandcr, R. P. Science 1983, 219, 703 Hamilton, B. K., Montgomery, J. P. and Wang, D. I. C. in l:'nzyme Engineering (Pye, E. K. and Wingard, L. B.. eds) Plenum Press, New York, 1973, vol. 2, p. 153 Dcmain, A. L. and Wang, D. I. C. in Second International Symposium on the Genetics of Industrial Microorganisms (McDonald, K. D., ed.) Academic Press, New York -London. 1976, p. 115 Demain, A. L., Piret, J. M., Fricbel, T. O. E., Vandamme, E. J. and Matteo, C. C. in Microbiology (Schlessinger, D., ed.) American Society for Microbiology, Washington, D.C., 1976, p. 437 Kleinkauf, It. and Koischwitz, H. in Multifunctional Proteins (Bisswanger, It. and Schminke-Ott, t(., eds) J. Wiley and Sons, Inc., New York, 1980, p. 217 Vandamme, t-. J. in Topics in l:'nzyme and Fermentation Biotechnology (Wiseman, A., ed.) Ellis llorwood, Chichester, 1981, vol. 5, p. 185 Gardner, C. R., Colton, C. K., Langer, R. S., Ilamilton, B. K., Archer, M. C. and Whitesides, G. M. in Enzyme Engineering (Pyc, I-. K. and Wingard, I.. B., eds) Plenum Press, New York, 1974, vol. 2, p. 209 Pace,G. W., Yang, H. S., Tannenbaum, S. R. and Archer, M. C. Biotechnol. Bioeng. 1976, 18, 1413 Asada,M., Nakanishi, K., Matsuno, R., Kariya, Y., Kimuya, A. and Kamikubo, T. Agric. Biol. Chem. 1978, 42, 1533 Asada,M.. Morimoto, K., Nakanishi, K., Matsuno, R., Tanaka, A., Kamikubo, T. and Kimura, A. Agric. Biol. Chem. 1979,43, 1773 Butler, L. Biotechnol. Bioeng. 1977, 19,591 Abraham, E. P. Biosynthesis and Enzymatic Hydrolysis of Penicillins and Cephalosporins University of Tokyo Press, Tokyo, 1974 Kohsaka,M. and Demain, A. L. Biochem. Biophys. Res. Commun. 1976, 70, 465 Kurzatkowski, W., Kurylowicz. W. and Paszkiewics, A. Eur. J. Appl. Microb. Biotechnol. 1982, 5, 21 I Vandamme, E. J. Chem. Ind. 1976, 24, 1070 Chibata, I. and Tosa, T. Annu. Rev. Biophys. Bioeng. 1981, 10, 197 Abbott, B. J. Adv. Appl. Microbiol. 1976, 20, 203 Abbott, B. J. in Annual Reports on Fermentation Processes (Perhnan, D., ed.) Academic Press, New York, 1977, vol. 1, p. 205 Morikawa,U., Karube, I. and Suzuki, S. BiotechnoL Bioeng. 1979, 21,261 Morikawa,Y., Ochiai, K., Karube, 1. and Suzuki, S. Antimicroh. Agents Chemother. 1979, 15, 126 Morikawa,Y., Karubc, 1. and Suzuki, S. BiotechnoL Bioeng. 1980, 22, 1015 1.reeman,A. and Aharonowitz, Y. BiotechnoL Biocng. 1981, 23, 2747 Vcnkatasubramanian, K. and Vieth, W. R. in Prog. Ind. MicrobioL 1979, 15, 61 Egorov,N. S., Baranova, I. P. and Kozliova, Y. U. I. Anti. biotiki (Moscow) 1978, 23,872-874 Katz, 1". and l)emain, A. L. Bacteriol. Rer'. 1977, 41,449

Enzyme Microb. Technol., 1983, vol. 5, November

415

Review 34

35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70

416

Froyshov, O., Zimmer, T. L. and Laland, S. G. in International Review o f Biochemistry; Amino Acid and Protein Biosynthesis H (Arnstein, H. R. V., ed.) University Park Press, Baltimore, 1978, vol. 18, p. 49 Hurst, A. Adv. AppL MierobioL 1981, 27, 85 Vandamme, E. J. and Voets, J. P. Adv. AppL Mierobiol. 1974, 17,311 Rolinson, G. N. J. Antimierob. Chemother. 1979, 5, 7 5,7 Wcissenburger, H. W. O. and Vanderhoeven, M. G. Rec. Tray. Chim. Pays-Bas Belg. 1970, 89, 1081 Fosker, G. R., Hardy, K. D., Nayler, J. H. C., Seggery, P. and Stover, E. R.J. Chem. Soc. 1971, 10, 1917 Gestrelius, S. AppL Biochem. BiotechnoL 1982, 7, 19 Savidgc, T. A. and Cole, M. Methods EnzymoL 1975, 43, 705 Lowe, I). A., Romancik, G. and Elander, R. P. Dev. hid. Micro& 1981, 22, 163 Carlsen, F. and Emborg, C. BiotechnoL Lett. 1981,3,375 Svedas, V. K., Margolin, A. L. and Berezin, I. V. Enzyme Micro& Technol. 1980, 2, 138 Svedas, V. K., Margolin, A. L., Borisov, I. L. and Berezin, 1. V. t:'nzyme Microb. Technol. 1980, 2,313 Okachi, R.. Kawamoto, I., Yamamoto, M., lakasawa, S. and Nara, T. Agric. Biol. Chem. 1973, 37,335 Poulsen, P. B. Enzyme Microb. Technol. 1981, 3,271 Savidge, T. A. in Biotechnology o f Industrial Antibiotics (Vandamme, E. J., ed.) Marcel Dekker, Inc., New York, in press Marconi, W., Ceccre, I.., Morisi, F., Della Penna, G. and Rappuoli, B. J. Antibiot. 1973, 26,228 Dinelli, D. Process Biochem. 1972, 7, 9 Fujii, T., Hanamitsu, K., Isumi, R., Yamaguchi, T. and Watanabe, 1. Jpn. Pat. 7 399 393 (1973) Fujii, T., Matsumoto, K. and Watanabe, 1". Process Biochent. 1976, II, 21 Demain, A. L., Walton, R. B., Newkirk, J. I.. and Miller, i. M. Nature 1963, 199,909 Walton, R. B. Dev. lndust. MicrobioL 1964, 5,349 Walton, R. B. Science 1964,143, 1138 Singh, K., Sun, S. and Rakhit, S. Eur. J. AppL Micro& BiotechnoL 1980, 9, 15 Niwa, T., Nojiri, C., Goi, H., Miyado, S., Kai, F., Seki, M., Yamada, Y. and Niida, T. Jpn. Pat. 52-143 289 (1977) Goi, H., Niwa, T., Nojiri, C., Miyado, S., Seki, M. and Yamada. Y. Jpn. Pat. 53-94093 (1978) Fujii, T., Shibuya, T. and Matsumoto, K. Proc. Annu. Meet. Agric. Chem. Soc. Japan 1--4 April 1979 Shibuya, Y., Matsumoto, K. and Fujii, T. Agric. BioL Chem. 1981,45, 1561- 1567 Mazzeo, P. and Romeo, A. J. 01era. Soc. Perkin Trans. 1. 1972, 20, 2532 Kitano, K., Fujisawa, Y., Katamoto, K., Nara, K. and Nakao, Y. J. Ferment. TechnoL 1976, 54,712 Ichikawa, S., Mural, Y., Yamamoto, S., Shibuya, Y., Fujii, T., Komatsu, K. and Kodaira, R. Agric. BioL Chem. 1981, 45, 2225-2229 lchikawa, S.. Shibuya, Y., Matsumoto, K., Fujii, T., Komatsu, K. and Kodaira, R. Agrie. BioL Chem. 1981,45, 2 2 3 1 - 2 2 3 6 Takeda, H., Matsumoto, I., Matsuda, K. and Kawakami, T. Jpn. Pat. 52-128 293 (1977) Matsuda, K., lnoue, T., Tanaka, K., Fukuo, T. and Kawate, S. Abstr. Annu. Meet. Agric. Chem. Soc. Jpn Tokyo, April 1978, p. 1 I1 Takeda, H. and Matsuda, K. Jpn. Pat. 53-86 094 (1978) Inoue, T., Matsuda, K., Fukuo, T. and Kawate, S. Abstr. Annu. Meet. Agric. Chem. Soc. Jpn Tokyo, April 1979, p. 220 Toyo Jozo Co. Belg. Pat. 801 044 and 1 324 159 (1973) Toyo Jozo Co. Jpn. Pat. 7 588 694 (1975)

Enzyme Microb. Technol., 1983, vol. 5, November

71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99

Fukushima, M., Fujii, T., Matsumoto, K. and Morishita, M. Jpn. Pat. 7 6 7 0 8 8 4 (1976) Nys, P., Satarova, D. E., Podshibhakina, L. V., Korchasin, V. B. and Savitskaya, E. M. Antibiotiki {Moscow) 1980, 25,803 Vandamme, 1-. J., Voets, J. P. and Dhaese, A. Ann. lnst. Pasteur, Paris 1971, 121,435 Fleming, I. D., Turner, M. K. and Napier, E. J. Ger. Pat. 2422 374 (1974) Abe, J., Watanabe, T., Yamaguchi, 7". and Matsumoto, K. US Pat. 3 761 354 (1973) Toyo JozoCo. UK Pat. 1 347 665 (1974) O'Callaghan, C. and Muggleton, P. W. Biochem. J. 1963, 89,304 Jeffrey, J. D. A., Abraham, E. P. and Newton, (;. G. I. Biochem. J. 1961,91,591 Wick, W. E., Wright, W. E. and Kuder, H. V. AppL MicrobioL 1971, 21,426 Nishida, M., Yokota, Y., Okui, M., Mine, Y. and Matsubara, T J. Antibiot. 1968, 21, 165 Konecny, J. and Sieber, M. BiotechnoL Bioeng. 1980, 22, 2013 Abbott, B. J. and Fukuda, D. S. AppL MicrobioL 1975, 30, 413 Abbott, B. J., Cerimele, B. and l:ukuda, D. S. Bioteehnol. Bioeng. 1976, 19, 1033 Takahashi, T., Yamazaki, Y., Kato, K. and lsono, M. J. Am. Chem. Soc. 1972, 94, 4035 Takahashi, T., Yamazaki, Y. and Kato, K. Abstr. Annu. Meet. Agric. Chem. Soc. Jpn Tokyo, 1973, p. 287 Takahashi, T., Yamazaki, Y. and Kato, K. Bioehem. J. 1974. 137,497 Takahashi, T., Kato, K., Yamazaki, Y. and Isono, M. Jpn J. Antibiot. 1977,30, 130 Kato, K., Kawahara, K., Takahashi, T. and Igarasi, S. Agric. BioL Chem. 1980,44,821 Kato, K., Kawahara, K., Takahashi, T. and Kakinuma, A. Agric. Biol. Chem. 1980,44, 1075 Kato, K., Kawahata, K., Takahashi, T. and Kakinuma, A. Agric. BioL Chem. 1980,44, 1069 Kato, K.Agric. BioL Chem. 1980,44, 1083 Savidge, T. A., Powell, L. W. and Lilly, M. D. Br. Pat. 1 357 317 (1974) Niclsen, M. tt. in 13th International TNO Conference (Verbraeck, A., ed.) TNO, The Hague, 1980, pp. 4 1 - 5 8 Sugic, M. and Suzuki, H. Agric. Biol. Chem. 1978, 42, 107 Sugie, M. and Suzuki, H. Agric. BioL Chem. 1980, 44, 1089 Yamada, H., Takahashi, S., Kil, Y. and Kumagai, H. J. Ferment. TechnoL 1978, 5 6 , 4 8 4 - 4 9 1 Yamada, It., Shimizu, S., Shimada, I1., Tani, Y., Takahashi, S. and Ohashi, T. Biochimie 1980.62,395 Olivieri, R., Fascetti, E., Angelini, L. and Degen, L. Enzyme Micro& TechnoL 1979, 1,201 Aoki, H. and Okuhara, M. Annu. Rev. Microbiol. 1980,

34,159 100 101 102 103 104 105 106 107

Kitano, K. Prog. Indust. Microbiol. 1983, 17, 37 Okachi, R. J. Agric. Chem. Soc. Jpn 1979, 53, R169 Komori, T., Kunugita, K., Nakahara, K., Aoki, ti. and Imanaka, H. Agric. BioL Chem. 1978,42, 1439 Fukugawa, Y., Kubo, K., Ishikura, T. and Kouno, K. J. Antibiotics 1980, 33,543 lmada, A., Kitano, K., Kintaka, K., Muroi, M. and Asai, M. Nature 1981,289, 590-591 Sykes, R. B., Bonner, D. P., Bush, K., Georgopapadaku, N. I1. and Well, S. S.J. Antimicrob. Chemother. 1981,8, (Suppl. E), 1 - 1 6 Parker, W. L., Rathnum, M. L., Wells, J. S., Treso, W. H., Principc, P. A. and Sykes, R. B. J. Antibiotics 1982, 35,653 l"ukui, S. and Tanaka, A. Annu. Rev. Microbiol. 1982, 36, 145