Journal of Bioscience and BioengineeringVOL. 112 No. 4, 409–414, 2011

www.elsevier.com/locate/jbiosc

Strain selection and scale-up fermentation for FR901379 acylase production byStreptomyces sp. no. 6907

Satoshi Ueda,1,⁎ Masayoshi Kinoshita,1 Fumihiro Tanaka,1 Masaru Tsuboi,1, § Shiho Shimizu,1,§§

Nobutaka Oohata,1 Motohiro Hino,1 Masato Yamada,2 Yasuhiro Isogai,2 and Seiji Hashimoto2

⁎ CorrespondE-mail addPresent add

178 KojinmachManagement, A8411, Japan.

1389-1723/$doi:10.1016/j

Fermentation and Biotechnology Laboratories, Astellas Pharma Inc., 156 Nakagawara, Kiyosu, Aichi 452-0915, Japan1 and Biotechnology Research Center,Faculty of Engineering, Toyama Prefectural University, 5180 Kurokawa, Imuzu, Toyama 939-0398, Japan2

Received 11 April 2011; accepted 9 June 2011Available online 12 July 2011

Micafungin (FK463) is a widely used treatment for life-threatening, deep-seated fungal infections. It is an echinocandin-like lipopeptide derived from the chemical modification of deacylated FR901379, a type of lipopeptide antibiotic produced byColeophoma empetri F-11899. The palmitoyl moiety of FR901379 is deacylated by FR901379 acylase produced by Streptomycessp. no. 6907. In this study, our goal was to generate an improved strain of Streptomyces sp. no. 6907 capable of hyperproducingthe FR901379-acylase enzyme. To accomplish this goal, modified strains of Streptomyces sp. no. 6907 were generated using UV-irradiation mutagenesis, and strain selection was performed using an agar-plate screening method to efficiently select anacylase-hyperproducing strain. Three marker indices were shown to correlate with elevated acylase production: decreasedcandidacidal activity of FR901379, decreased proteolytic activity on skim milk, and phenotypic characteristics. Cloning andsubsequent sequencing of the acylase gene from the hyperproducing mutant revealed no mutations in either the acylasestructural gene or the 5′-flanking region required for gene expression. The growth medium was also modified to maximizeacylase production. We successfully increased acylase activity approximately 65-fold, compared with the original growthconditions (wild strain cultured in the original unmodified medium). To minimize formation of excess foam during thefermentation process, we optimized the parameters of agitation speed, as calculated from the discharge flow rate. Using ourimproved strain and the optimized medium and growth conditions, we have developed an improved and highly reproduciblemethod for stable large-scale production of FR901379-acylase.

© 2011, The Society for Biotechnology, Japan. All rights reserved.

[Key words: FR901379 acylase; Strain breeding; Medium improvement; Scale-up; Bioconversion]

The improvement of a microbial strain, in order to hyperproduce itstargetproduct, is anessential component for establishing its commercialfermentation process. This improvement helps reduce costs relatedto raw materials, facility usage, labor costs, and other expenditures. Italso reduces production of chemical waste (such as organic solvents),decreases carbon dioxide emissions, and reduces energy consumption.In addition, optimization of the production medium and culture con-ditions helps to maximize productivity and establishes a reproduciblecommercial-scale manufacturing procedure.

Micafungin (FK463), developed by Astellas Pharma, Inc., is usedworldwide in chemotherapy for life-threatening, deep-seated fungalinfections. It is a semi-synthetic product derived from acylated cyclichexapeptide FR901379, a natural product from the fungus Coleophomaempetri F11899, via enzymatic deacylation of FR901379, followed by

ing author. Tel.: +81 52 400 2611; fax: +81 52 409 5927.ress: satoshi-ueda@jp.astellas.com (S. Ueda).resses: § Toyama Technology Center, Astellas Pharma Tech. Co. Ltd., 2-i, Toyama-City, Toyama 930-0809, Japan and §§ Project and Qualitystellas Pharm Inc., 2-3-11 Nihonbashi-Honcho, Chuo-Ku, Tokyo 103-

- see front matter © 2011, The Society for Biotechnology, Japan. All.jbiosc.2011.06.002

chemical reacylationwith the optimizedN-acyl side chain (1–3) (Fig. 1).In the manufacturing process, FR901379 acylase is used to deacylateFR901379, resulting in the intermediate product FR179642. FR901379acylase is produced by Streptomyces sp. no. 6907, which we previouslyisolated while screening for novel acylase-producing microorganisms(4). Because acylase productivity in thewild strain cultured in flaskswasvery low, it was necessary to develop an acylase-hyperproducingstrain,using strain-improvement methods, before establishing the man-ufacturing process in a large-scale fermentor. Although modernbiotechnology methods for evolutionary engineering (such as generecombination or gene shuffling) are available for use in strainbreeding (5), the use of these technologies in commercial productionis limited by strict legal regulations and high costs of constructingfacilities appropriate for use of genetically engineered strains. In thisstudy, we report a method for strain improvement that can be easilyapplied to commercial production through the successful selection ofan FR901379-acylase-hyperproducingmutant. The accompanyingmedium optimization and parameter setting for large-scale fermen-tation of the Streptomyces sp. no. 6907 mutant are also shown to beeffective for FR901379 acylase production.

rights reserved.

FIG. 1. Manufacturing process of micafungin.

410 UEDA ET AL. J. BIOSCI. BIOENG.,

MATERIALS AND METHODS

Microorganisms Streptomyces sp. no. 6907 (FERM BP-5809) and its variantswere used throughout this study. Candida albicans Yu-1200 was also used in thescreening to select the acylase-hyperproducing mutant.

Chemicals FR901379 was prepared at the Fermentation and BiotechnologyLabs., Astellas Pharma Inc. Other chemicals were purchased fromWako Pure Chemicals(Osaka, Japan).

Mutagenesis The mutants of Streptomyces sp. no. 6907 were developed usingUV-irradiation mutagenesis. After Streptomyces sp. no. 6907 was grown on a nutrientagar plate at 30°C for 5 days, the spores were suspended and diluted in 0.05% Tween 20solution to obtain a final concentration of at least 1×106 cfu/ml. The spore suspensionwas then exposed to a UV lamp (254 nm) for more than 60 s, yielding a survival rate ofless than 1%. The UV-irradiated spores were plated onto agar medium (e.g., corn mealagar or Hickey–Tresner agar) and cultured at 30°C for 5 days.

Culture conditions and media A double-layer agar method was developed forselection of the FR901379-acylase hyperproducing mutants. The bottom layerconsisted of 1% glucose, 0.4% bouillon, 0.2% yeast extract, 0.4% peptone, 2.5% sodiumchloride and 1.5% agar supplemented with 1 mg/l of FR901379. The upper layerconsisted of 0.5% glucose, 1.0% Bacto-peptone, 0.2% NaNO3, 0.05% KCl, 0.05% MgSO4.7-H2O and 0.7% agar inoculated with C. albicans Yu-1200 after sterilization. The bottomlayer, on which the mutants grew to the appropriate colony size, was overlaid with themolten upper layer, and plates were incubated overnight at 30°C.

For the initial fermentation of FR901379-acylase-producing strains in thelaboratory, a seed medium (20 ml), consisting of 6% soluble starch, 4% defatted soyflour, and 0.5% CaCO3, was poured into a 100 ml Erlenmeyer flask and sterilized at121°C for 20 min. A loopful of the slant culture of Streptomyces sp. no. 6907 or mutantstrain was inoculated into the medium and cultured under shaking conditions of250 rpm at 25°C for at least 20 h. A production medium (Medium A, 20 ml) consistingof 6% starch, 2% potato protein, 2% dried yeast and 0.5% CaCO3 was poured into a 100 mlErlenmeyer flask and sterilized at 121°C for 20 min. The seed culture (0.4–1 ml) wasthen transferred to the production medium, and fermentation was performed undershaking conditions of 250 rpm at 30°C for at least 3 days. Acylase activity in the culturebroth was measured as described below. In the scale-up experiments, the workingvolumes in 30 l and 1 kl fermentors were 20 l and 700 l or 800 l, respectively. Anti-foaming agent was added to the production medium and sterilized at 121°C for 30 min.

Measurement of FR901379 acylase activity Because FR901379 acylase is acellular or membrane-bound enzyme that can be released into the medium, its activitymust be measured in the culture broth during the fermentation process. Acylaseactivity was routinely assayed by estimating the amount of FR179642 (the finalacylation-reaction product) produced in a reaction (1 ml) containing 10 mg/ml ofFR901379, 50 mM Tris-HCl buffer (pH 6.0), 0.7 ml of culture broth or diluted cultured

broth and 10% methanol to disperse FR901379 in the reaction mixture. Reactions wereincubated with shaking at 40°C for 15 min and were terminated by the addition of 4%acetic acid solution. Thereafter, the reaction mixture was diluted with an equal volumeof methanol and filtered through a cellulose acetate membrane filter (0.45 μm). Thefiltered reaction mixture was analyzed for FR179642 content by measuring UVabsorption at 215 mm using an HPLC system equipped with a Kaseisorb LC PO Supercolumn (4.6 mm i.d.×250 mm, Tokyo Kasei Kogyo Co. Ltd., Japan) at a columntemperature of 50°C. A mobile solvent consisting of 4% methanol buffered with 0.1%phosphoric acid was used at a flow rate of 1 ml/min.

RESULTS AND DISCUSSION

Strain improvement As we described in a previous report (4),C. albicans Yu-1200 is sensitive to FR901379 but insensitive toFR179642 (the final product of the FR901379 deacylation reaction).Based on this characteristic sensitivity, a screening method of double-layer agarwasused for selecting the FR901379-acylase-hyperproducingstrain. The growth of C. albicans around a colony of an hyperproducingmutant is better than that around a colony of the parental strain in theupper layer, because FR901379 in the bottom layer around theFR901379-acylase-hyperproducing mutant is more deacylated toFR179642 (Fig. 2A). This method was simple to test but difficult toperform after completing several sequential rounds of strain improve-ment. As a result of the increased levels of acylase produced by themutant strain, a shift in the relative levels of FR901379 and FR179642allowed C. albicans to grow over the entire upper layer. Additionally,spores of Streptomyces sp. no. 6907 or mutants tended to spread andgrow on the surface of the upper layer. These factors hamperedour ability to accurately select acylase-hyperproducing colonies.Therefore, we developed amethod to improve themedium formulationused in the agar plates as follows: (i) The amount of FR901379supplemented in the bottom layer was increased to compensate forincreased acylase activity, thereby preventing the complete deacylationof FR901379 in the medium, allowing the growth of C. albicans onlyaway from the colony of the acylase-hyperproducing mutant;

FIG. 2. Three markers were used for selecting the FR901379-acylase-hyperproducing strains on agar plates. (A) Candidacidal activity on double-layer agar. (B) Proteolytic activity onskim milk agar. (C) Sporulation on ISP2 agar. The colony indicated with an arrow was selected by each marker in the experiment of strain improvement.

IMPROVED FERMENTATION FOR FR901379 ACYLASE PRODUCT 411VOL. 112, 2011

(ii) Tween 80 (0.1%) was added to the upper agar medium to suppressthe overgrowth of C. albicans, and Streptomycin (10 μg/ml) was alsoadded to inhibit overgrowth of Streptomyces sp. no. 6907 or themutants;(iii) The volume of the upper-layer agar medium was decreased from15 to 10 ml, which was also effective for diffusing residual FR901379 inthe lower agar. Thus, by continually optimizing themedium formulation,we were able to obtain an FR901379-acylase-hyperproducing strain bythis screening method after several rounds of strain improvement.

A drastic decrease of FR901379 acylase activity in submergedculture after 4 days was one of the serious problems encounteredwhile establishing the manufacturing process (Fig. 3). We speculatedthat this decrease might be due to the degradation of acylase pro-tein by some proteolytic enzyme(s) co-produced in Streptomyces sp.no. 6907. According to the literature (6,7), a decrease in proteolyticdegradation can be achieved by optimizing growth conditions such aspH and temperature, particularly for the production of proteolyticallysensitive proteins. However, because it would be preferable to obtaina strain with intrinsically less proteolytic activity, we attempted toisolate mutants that showed less proteolytic activity, using with amodified procedure described in the literature (8). After a mutant wasspotted onto an agar plate containing 0.5% skimmilk and cultivated at30°C for 3–5 days, the halo size around the colony was measured toquantify its proteolytic activity (Fig. 2B). The halo size of theFR901379-acylase-hyperproducing strains tended to be smaller thanthose of low-producing strains, and a weak inverse correlation

FIG. 3. Change of FR901379 acylase activity in culture broth. Relative activity of 100%indicates the highest activity during cultivation. Wet cell volume was measured asfollows: broth was transferred into a centrifuge tube and centrifuged at 1640×g for10 min. Pellet volume was measured and used as wet cell volume. Symbols: opensquares, wet cell volume; closed squares, relative activity.

between acylase activity and the halo size was reproducibly obtained(Fig. 4). Our speculation was substantiated by the fact that, by se-lecting the mutants with smaller halo size on the skim milk agarplates, we were able to successfully select the mutants that showedno decrease in acylase activity at the late phase of submerged culture.

In the submerged culture, FR901379 acylase production started atthe late-logarithmic growth phase and continued to the stationaryphase (Fig. 3). A morphological change of mycelia concomitantlyoccurred, from the long filaments found at the middle- to late-logarithmic growth phase to the shortly fragmented filaments foundat the early stationary phase (Fig. 5). However, the mutants with lowFR901379 acylase activity grew in the formof long filamentousandnon-fragmented mycelia throughout the cultivation. These observationssuggested that the regulation of FR901379 acylase production mightbe similar to that of the morphological differentiation and mightbe controlled by the same regulatory pathway We assumed that thisregulation system would be morphologically expressed during thegrowth of Streptomyces sp. no. 6907 on the solid medium as theformation of aerial mycelium and sporulation, because these morpho-logical changes are known to correlate with secondary metabolite pro-duction in Streptomyces (9,10). A composition of solid medium wasoptimized in order to clearly differentiate the degree of aerial myceliaformation and sporulation between the parental strain and themutants.We were able to obtain an FR901379-acylase-hyperproducing mutantby selecting a mutant with abundant aerial mycelium and highlysporulated morphology on this optimized medium (Fig. 2C). Therelationship between morphological differentiation and secondarymetabolite production has been previously reported, primarily based

FIG. 4. Correlation between halo diameter and acylase activity. Relative activity of 100%indicates highest acylase activity of the mutant-culture broth.

FIG. 5. Microscopic photographs of Streptomyces sp. no. 6907 mycelia at growth phase (A) and stationary phase (B).

412 UEDA ET AL. J. BIOSCI. BIOENG.,

on the observation that morphologically defective mutants producedlower levels of the secondary metabolite (9–12). We used this rela-tionship to identify the hyperproducing mutants during the strainbreeding. To the best of our knowledge, there are few reports thatdescribe the successful application of this relationship in the breedingof an industrial strain.

The acylase genes from the genomes of the wild and hyperprodu-cing strains were cloned and sequenced as described previously (13).DNA sequences from both strains were completely identical (data notshown). This result suggests that the observed increase in acylaseactivity could result from the enhanced expression of acylase proteinby some up-regulation mechanism in the mutant.

Table 1 summarizes the results of the strain improvement, inwhich the acquisition rates were calculated from the total numbersof strains obtained in all rounds. We obtained the hyperproducingmutants with high emergence rates from the selection of eachmarker.Although the most effective marker was not identified, the marker forcandidacidal activity usually resulted in an hyperproducing strain thatled to an increase in acylase production, by an average of 2.7-fold(from 1.9 to 5.2 in medium B) in the first round (Tables 1 and 4).

Medium improvement In parallel with strain breeding, weconducted studies for the improvement of the culture medium at thescale of a 100 ml Erlenmeyer flask. Although many strategies forimproving culture medium are described in the literature (14), weinvestigated the effects of nitrogen and carbon sources on productiv-ity of FR901379 acylase.

Screening for a nitrogen source was initially performed byreplacing 2% potato protein with each nitrogen source (Table 2).Defatted soy flour showed higher production than any of the othernitrogen sources. When dried yeast, defatted soy flour, soybeanpowder and defatted wheat germ were used as nitrogen sources,higher acylase activity was observed. Overall, defatted soy flourproduced the greatest increase in acylase activity per wet cell volume

TABLE 1. Summary of strain improvement to select higher producing mutants.

Index Number of strains Acquisitionrate (%)

Screened Submerged cultured High producer

Candicidal 9826 677 17 2.5Proteolytic activity 1921 104 3 2.9Spore formation NC 671 13 1.9

This result was obtained from the 6 rounds of strain improvement. NC, not counted onthe agar plate on which UV-irradiated spores formedmany colonies. High producer, themutant strain which showed a more than 1.2-fold productivity compared with theparent strain. In the case of candidacidal index, each of 677 selected strains from the9826 mutants was cultured in a 100 ml Erlenmeyer flask. Each of 17 candidate strainswhich showed high productivity was cultured again in a plural number of 100 mlErlenmeyer flasks to assess the reproducibility of production.

(WCV) compared with the other nitrogen sources. Our results alsosuggested that FR901379-acylase production might be regulated by amechanism similar to the secondary metabolism that controlssporulation. Defatted wheat germ is generally considered superiorto defatted soy flour, because it contains not only a nitrogen sourcebut also more minerals and vitamins to be utilized in primarymetabolism. In the medium composed of defatted wheat germ, thelower production of FR901379 acylase may be due to the enhance-ment of primary metabolism related to cell growth (WCV). Thisenhancement of primary metabolismmay reduce the flow of nitrogensources to secondary metabolic pathways, resulting in lowerproduction of FR901379 acylase. In contrast, we speculate thatdefatted soy flour showed the highest acylase production, alongwith adequate cell growth, because it contains an optimal balancebetween the nutrients that are required for both primary andsecondary metabolism during the fermentation of Streptomyces sp.no. 6907.

Next, we screened for an optimal carbon source. Dextrin mediumproduced the highest level of acylase production (Table 3). However,the enzyme activity per WCV was highest in the medium containingfructose as a carbon source, due to the relatively lower WCV obtainedusing this medium. The maltose medium produced the highest WCVbut showed similar acylase activity to that of the fructose medium.The sucrose medium showed the lowest enzyme activity, most likelybecause Streptomyces. sp. no. 6907 could not utilize sucrose insubmerged culture. These findings provided us with the necessarydata to formulate an optimal medium allowing us to maximizeFR901379 acylase production. Using this data, we modified theoriginal medium A to produce medium D via mediums B and C.

Integrated result of strain andmedium improvement Table 4shows the integrated result obtained from strain- and medium-improvement studies. Through six rounds of strain breeding and

TABLE 2. Influence of nitrogen source on FR901379 acylase production.

Nitrogen source Acylase activity (U) Wet cell volume (%) Acylase activityper wet cell volume

Dried yeast 2.3 28 0.08Defatted soy flour 3.7 14 0.26Soybean powder 2.5 15 0.17Defattedwheat germ 2.3 23 0.10Cotton seed flour 1.5 18 0.08Polypeptone 1.5 12 0.13Ammonium sulfate 0.15 7 0.02Control 1.5 14 0.11

Two percent of each nitrogen source was added to the medium. Acylase activity andwet cell volume were monitored by periodical sampling from the individual flask andthe highest activity was indicated for each medium. One unit of acylase activitydeacylates 1 M of FR901379 in 1 h.

TABLE 3. Influence of carbon source on FR901379 acylase production.

Carbon source Acylase activity (U) Wet cell volume (%) Acylase activity perwet cell volume

Fructose 3.9 13 0.30Glucose 4.4 18 0.24Maltose 3.8 28 0.14Lactose 1.0 10 0.10Sucrose 0.9 8 0.11Galactose 3.0 16 0.19Dextrin 5.6 21 0.27Starch (control) 4.3 27 0.16

Four percent of each carbon source was added to the medium. Acylase activity and wetcell volume was determined by the same method described in Table 2.

IMPROVED FERMENTATION FOR FR901379 ACYLASE PRODUCT 413VOL. 112, 2011

concomitantly progressive medium improvement, acylase activity pervolume of broth was increased 65-fold compared with the originalgrowth conditions (wild strain cultured with medium A). Because theeffect of medium improvement was estimated to be at most 2.4-foldper cell, as shown in Tables 2 and 3 (defatted soy flour 0.26 vs. control0.11 in Table 2), or 3.5-fold per volume of broth, as calculated fromTable 4, the strain breeding greatly contributed to a the total enhancedactivity. The enhanced enzyme activity observed, along with the rawmaterials used for the growth medium, makes this a suitable methodfor establishing a commercial-scale process.

Conventional artificial mutagenesis has frequently been used forstrain breeding and has proven to be effective in the production ofcommercialized antibiotics. It is important to identify an effectivemarker to distinguish hyperproducing strains from the many colonieson a mutant-culture plate. We previously reported that a newlyidentified marker of colony size on the agar medium, supplementedwith sodium chloride, was sufficient for selecting an FR901379 high-producer (15). Although enzyme production may be regulated bydifferent mechanisms than the production of secondary metabolitessuch as antibiotics, the establishment of effective markers for pro-tein production is believed to be a key factor for the success of strainbreeding. We found three indices, deactivation of candidacidal activ-ity, decreased proteolytic activity on skim milk and morphologicalphenotype, which were effective markers for use in selectingFR901379-acylase-hyperproducing strains. The first marker wasdirectly related to enzymatic activity of the target protein,FR901379 acylase, which deacylates FR901379 to yield a productlacking candidacidal activity. The second marker was identified basedon the hypothesis that a drastic decline in acylase activity during late-phase culture might result from non-specific proteolytic degradationof FR901379 acylase by a co-produced protease(s). The third markerwas identified based on the hypothesis that the production ofFR901379 acylase might be regulated by a mechanism similar tothat of secondary metabolism. This conclusion was based on theobserved correlation between increased acylase activity and idio-

TABLE 4. Integrated result of strain and medium improvement for FR90139 acylaseproduction.

Selectionmarker

Relative productivity

Medium A Medium B Medium C Medium D

Original 1 1.9F1 Candidacidal 5.2 6.9F2 Proteolysis 9.3F3 Sporulation 19 26F4 Candidacidal 40F5 Sporulation 52F6 Candidacidal 65

A selected strain at each roundwas determined by culturing repeatedly in 100 ml flasksand 30 l fermentors to confirm the reproducibility of fermentation, and then used as thestrain to be mutated at the next round.

phase development of the microorganism. Using these markers ineach round of mutagenesis proved effective for selection of theenzyme-hyperproducing strain in the current study.

Scale-up parameters When Streptomyces sp. no. 6907 wascultured in a fermentor, acylase activity in culture broth drasticallydecreased at the early proliferation phase, most likely due to a rise inpH. We found that decreasing the amount of antifoaming agent in themedium prevented a rise in pH; however, it caused severe foaming inthe later growth phase. Foam mitigation in fermentation has beenwidely studied for over 50 years (16). Adding antifoaming agent isone of the general methods used to avoid formation of excess foam(17); however, in this instance it was necessary to find a differentsolution for foam control during FR901379 acylase fermentation.Hoeks et al. proposed that foam disruption could be controlled bymodifying agitation speed in the fermentor, whether on a laboratoryscale or on a commercial scale (18,19). Our study conducted using 30 lfermentors revealed that agitation could be the most importantparameter for foam control during FR901379 acylase production.Therefore, the relationship between agitation speed and foaming wasinvestigated at the pilot scale, using 1 kl fermentors containingworking volumes of 700 l and 800 l. Under these conditions, foamingoccurred at an agitation speed of 170 rpm. However, foaming wassuppressed at 200 rpm in the fermentor containing 700 l of culturebroth (Fig. 6). Foam was not suppressed at any agitation speed in thefermentor containing 800 l of culture broth. The distances betweenthe upper impeller and the surface of the medium were 30% and 40%of the height of the medium, in the 700 l and 800 l working volumes,respectively. When the distance between the upper impeller and thesurface of the medium was less than 20% of medium height, a portionof the upper impeller and blade was visible at the surface of themedium, and no defoaming effect was observed during cultivation.Therefore, we determined that the optimal position for the upperimpeller was 25%−30% below the surface of the medium.

Foam formation is generally accelerated with increasing air flowduring fermentation; however, sufficient air supply is essential forsuccessful production of FR901379 acylase. We assumed that foamingcould be suppressed fairly easily using a vortex effect, by which thegenerated foams aremixed back into the broth and thereby defoamed.Presuming that the vortex effect is proportional to discharge flow ratefrom the upper impellers, we developed Eq. 1 by modifyingthe original equation described in the literature (20) to define the‘defoaming parameter’, which was calculated by dividing thedischarge flow rate by the air flow rate. We varied the value of the

FIG. 6. Effect of agitation speed on foam disruption in a 1 kl reactor using 700-l and800-l working volumes.100% of relative volume indicates the initial volume beforeagitation. As more foam is generated, the relative volume increases. Symbols: opencircles, working volume of 800 l; closed circles, working volume of 700 l.

TABLE 5.Usefulness of defoaming parameter as a tool for foam disruption in the reactorwith the working volume of 700 l.

Defoaming parameter(discharge flow rate/air flow rate)

Hold up rate (%) Foaming Relativeproductivity (%)

100 113 No nd.91 115 No 9881 116 No 10060 150 Yes 80

nd.: not determined.

414 UEDA ET AL. J. BIOSCI. BIOENG.,

defoaming parameter from 60 to 100 during scale-up experiments,using a fermentor with a working volume of 700 l. As a result, weconcluded that the defoaming parameter should be set to be greaterthan 80, in order to provide a better fermentation process with a lowhold-up rate, no foaming and satisfactory productivity (Table 5).

Defoamingparameter = dischargeflowrate�= air flowrate= agitation speedð Þ × impeller diameterð Þ3 = air flowrate�dischargeflowrate = discharged flow coefficientð Þ× agitation speedð Þ × impeller diameterð Þ3

ð1Þ

Discharge flow coefficient in each fermentor was set to equalvalues.

Thus, we overcame amajor hindrance of excess foaming, whichweencountered while scaling up the culture of Streptomyces sp. no. 6907.It is important to determine the appropriate parameters required toscale up the culture from laboratory flask to commercial fermentors.In this study, our defoaming parameter provided us with an optimalcondition for large-scale production.

References

1. Tomishima, M., Ohki, H., Yamada, A., Maki, K., and Ikeda, F.: Novel echinocandinantifungals. Part 2: optimization of the side chain of the natural product FR901379.Discovery of micafungin, Bioorg. Med. Chem. Lett., 18, 2886–2890 (2008).

2. Tomishima, M., Ohki, H., Yamada, A., Maki, K., and Ikeda, F.: Novel echinocandinantifungals. Part 1: novel side-chain analogs of the natural product FR901379,Bioorg. Med. Chem. Lett., 18, 1474–1477 (2008).

3. Hashimoto, S.:Micafungin: a sulfated echinocandin, J. Antibiot. (Tokyo), 62, 27–35(2009).

4. Ueda, S., Sakamoto, K., Oohata, N., Tsuboi, M., Yamashita, M., Hino, M., Yamada,M., and Hashimoto, S.: Screening and characterization of microorganisms withFR901379 acylase activity, J. Antibiot. (Tokyo), 63, 65–70 (2010).

5. del Cardayre, S. B.: Developments in strain improvement technology, pp. 107–125, in: Zhang, L. and Demain, A. L. (Eds.), Natural products: drug discovery andtherapeutic medicine. Human Press, Totowa (2005).

6. Baneyx, F., Ayling, A., Palumbo, T., Thomas, D., and Georgiou, G.: Optimization ofgrowth conditions for the production of proteolytically-sensitive protein in theperiplasmic space of Escherichia coli, Appl. Microbiol. Biotechnol., 36, 14–20(1991).

7. Zhou, X. S. and Zhang, Y. X.: Decrease of proteolytic degradation of recombinanthirudin produced by Pichia pastoris by controlling the specific growth rate,Biotechnol. Lett., 24, 1449–1453 (2002).

8. Mattern, I., van Noort, J., van den Berg, P., Archer, D., Roberts, I., and van denHondel, C.: Isolation and characterization of mutants of Aspergillus niger deficientin extracellular proteases, Mol. Gen. Genet., 234, 332–336 (1992).

9. Braun, S. and Vecht-Lifshitz, S.: Mycelial morphology and metabolite production,Trends Biotechnol., 9, 63–68 (1991).

10. Chater, K. F.: Genetics of differentiation in Streptomyces, Annu. Rev. Microbiol., 47,685–711 (1993).

11. Kang, D. K., Li, X. M., Ochi, K., and Horinouchi, S.: Possible involvement of cAMPin aerial mycelium formation and secondary metabolism in Streptomyces griseus,Microbiology, 145, 1161–1172 (1999).

12. Domain, A. L.: Induction of microbial secondary metabolism, Int. Microbiol., 1,259–264 (1998).

13. Ueda, S., Shibata, T., Ito, K., Oohata, N., Yamashita, M., Hino, M., Yamada, M.,Isogai, Y., and Hashimoto, S.: Cloning and expression of the FR901379 acylasegene from Streptomyces sp. no. 6907, J. Antibiot. (Tokyo), 64, 169–175 (2011).

14. Kennedy, M. and Krouse, D. J.: Strategies for improving fermentation mediumperformance: a review, Ind. Microbiol. Biotechnol., 23, 456–475 (1999).

15. Kanda, M., Tsuboi, M., Sakamoto, K., Shimizu, S., Yamashita, M., and Honda, H.:Improvement of FR901379 production by mutant selection and mediumoptimization, J. Biosci. Bioeng., 107, 530–534 (2009).

16. Junker, B.: Foam and its mitigation in fermentation systems, Biotechnol. Prog., 23,767–784 (2007).

17. Ghildyal, N. P., Lonsane, B. K., and Karanth, N. G.: Foam control in submergedfermentation: state of the art, Adv. Appl. Microbiol., 33, 173–222 (1988).

18. Hoeks, F. W., Van Wees-Tangerman, C., Gasser, K., Mommers, H. M., Schmid, S.,and Luyben, K. C.: Stirring as foam disruption (SAFD) technique in fermentationprocesses, Can. J. Chem. Eng., 75, 1018–1029 (1997).

19. Hoeks, F. W., Boon, L. A., Studer, F., Wolff, M. O., van der Schot, F., Vrabél, P., vander Lans, R. G., Bujalski, W., Manelius, A., Blomsten, G., Hjorth, S., Prada, G.,Luyben, K. C., and Nienow, A.W.: Scale-up of stirring as foam disruption (SAFD) toindustrial scale, J. Ind. Microbiol. Biotechnol., 30, 118–128 (2003).

20. Oldshue, J. Y.: Fermentation mixing scale-up technique, Biotechnol. Bioeng., 8,3–24 (1966).