Implementation of a rapid microbial screening procedure for biotransformation activities

JOURNAL OF BIOSCIENCEAND BIOENGINEERING Vol. 89, No. 4, 367-371. 2000

Implementation of a Rapid Microbial Screening Procedure for Biotransformation Activities SARAH STAHL, * RANDOLPH GREASHAM, AND MICHEL CHARTRAIN Department of Bioprocess R&D, Merck Research Laboratories, Rahway, NJ, USA Received 28 July 1999/Accepted 27 January 2000

A rapid and efficient microbial screening procedure was developed utilizing a 24-well plate format in conjunction with au automated liquid handling system and an HPLC. For the evaluation of this miniaturized and automated screening system, we selected the bioreduction of 6-bromo-,&etralone to 6-bromo-ptetralol. This procedure employed both yeast and rhodococci libraries, representing a culture collection comprised of several hundred strains, from which to screen for desirable bioconversion activity. Most of these strains had demonstrated bioreducing activity during previous screensto insure a “hit rate” as high as possible. The cultivation of microbes in the plate format was facile, time saving, and eilicient compared to the standard method of screening utilizing larger volumes, such as test tubes or shake flasks. This improved method of screening for bioconversion activity, employing pre-selected microbial libraries based on microtiter plates and a fully roboticized analytical system, proved to rapidly yield valuable leads which compared advantageously with a more classical approach. A total of 192 yeast strains and 48 rhodococci strains were screened using this procedure. Analytical data revealed that 78% of the strains tested bioconverted the tetralone to the desired alcohol. [Key words:

microbial screening, robotics, automation, biotransformation]

Traditionally, microbial screeningsfor bioconversion activities have been quite labor intensive (1). Specifically, most screeningcan be summarized as beginning with the thawing of several hundred cryovials each containing an individual strain. This step is immediately followed by the inoculation and cultivation of each strain in a tube or flask. At some time during the incubation period, the substrate for bioconversion is added to each tube or flask which is then returned to its previous environmental conditions. Bioconversion activity is then evaluated by sampling each cultivation vessel and by manually preparing each sample for analytical purposes. Upon consideration of all these combined activities, it is understandable why microbial screeningsare vastly time consuming. The literature teaches us that most successful screeningsfor bioconversion activities, such as bioreductions, usually require the evaluation of up to severalhundred microbial strains (2-14). Therefore, having access to a better technology which would both minimize the labor input needed for such a microbial screening and concurrently reduce lead identification time would greatly improve the efficiency of screeningactivities. The objective of this present work was two fold. Firstly, we wanted to establish a microbial library in a format that would both limit manual handling and also greatly reduce the length of the screeninginitiation (the thawing of hundreds of cryovials and the corresponding inoculation of flasks). Secondly, our objective was to reduce the length of time of the analytical operations (sample preparation and assay) by taking advantage of existing analytical high throughput technology (15-17). To concurrently addressthese two points, we developed and implemented the cryostorage and activity evaluation of microbial strains in a 24-well template format. This microbial library was linked with the development of a robotic

system, which has the capability

to perform

both

plates used for the cultivation of the microbes. For proof of concept, we selectedthe bioreduction of 6-bromo-P-tetralone (1) to 6-bromo-/3-tetralol (2) (Fig. I), which had been previously handled in our laboratory via a classicalmicrobial screen(9). The present communication clearly shows that this screeningtechnique, when co-implementedwith a simple robotic system, can generate reliable data in a very short period of time. To the best of our knowledge, such an approach has not been reported in the biotransformation literature. MATERIALS

Chemicals All chemicals used were of reagent grade and were purchased from either Fisher Scientific (Springfield, NJ, USA) or Sigma Chemical Co. (St. Louis, MO, USA). Sabouraud dextrose broth (SDB) and tryptic soy broth (TSB) were purchased from Difco (Detroit, MI, USA). The 6-bromo+tetralone and 6bromo-fi-tetralol (for analytical standards)were prepared at Merck ResearchLaboratories (ProcessResearch,Rahway, NJ, USA) according to a previously published procedure (18). Microbial screening Master template construction

Yeast and rhodococci strains were obtained from the Merck Microbial Resources Culture Collection (Merck & Co., Rahway, NJ, USA). Most yeastshad beenpreviouslyprocuredby the Merck Microbial ResourcesCulture Collection from the Brm”

microbe-

(1)

sample preparation and assaysdirectly from the 24-well

AND METHODS

BrmoH

(2)

FIG. 1. Microbial bioreduction of 6-bromo-p-tetralone (1) to 6bromo-a-tetralol (2).

* Corresponding author. 367

368

J. BIOSCI.BIOENG.,

STAHL ET AL.

University of California-Davis Yeast Culture Collection assembled by Dr. H. Pfaff. This collection represents a very extensive and ecologically diverse yeast library. The rhodococci were selected to accommodate intra-species diversity. All strains were preserved as frozen suspensions in 20% glycerol at -70°C. A l-ml volume of the frozen suspension was used to inoculate a 25ml test tube containing 5 ml of Sabouraud dextrose broth (SDB) (30 g/f) or tryptic soy broth (TSB) (30 g/l) for the cultivation of yeast and rhodococci, respectively. After 48 h of incubation at 29°C on a shaker operated at 220rpm (2in throw), 5 ml of 50% glycerol were added to each tube. From each tube, 2ml of culture were transferred into one well in a polystyrene 24-well cell culture plate (3 ml total volume capacity; Costar Corp., Cambridge, MA, USA). This procedure was repeated for each culture and led to the creation of master templates, which were stored at -70°C until use (Fig. 2). Screening Using a multichannel pipettor (Finnpipette Digital, Fisher Scientific), a 200-/11 aliquot of each master template well was used to inoculate a corresponding well in a sterile polypropylene 24-well plate (each well consisting of a lOm1 total volume capacity; Whatman Polyfiltronics, Clifton, NJ, USA) containing 2ml of SDB (3Og/Z) or TSB (3Og/f) and a 2mmx7mm micro stir bar (Fisher Scientific). After 48 h of incubation at 29°C on a stir plate, 2mg of ketone in 40 ~1 of ethanol (2% v/v ethanol solution) were added to each well of the plate, again utilizing a multichannel pipettor. The 24-well plates were returned to the same incubation conditions as described above (Fig. 2). System description and analytical methods Sample processing and HPLC analyses were performed employing a Gilson autosampler system model 215 (Gilson, Inc., Middleton, WI, USA). This system automatically diluted, filtered, and performed HPLC analyses on samples directly from the 24-well cultivation/bioconversion plates. Essentially, the Gilson liquid handler is composed of an XYZ robot arm mounted on a linear track which is linked to system hardware and instrumentation (PC, pumps, filter apparatus, and HPLC system). The plates are processed as follows: an equal volume of methanol (2ml) is automatically added to the well containing whole broth. The system mixes the whole broth and methanol together in the well by aspirating and dispensing the mixture three times at a flow rate of 5 ml/min. The cell components are removed by automatically passing 1 ml of the mixture through a 0.2 /* filter (model A-332 semi-prep filter, Upchurch Scientific, Oak Harbor, WA, USA) into a collection plate (each well consisting of a 3 ml total volume capacity; Costar Corp.). From the collection plate, 20~1 of the filtrate is automatically loaded onto a reverse phase HPLC column. The filter is then backwashed for 1.5 min with 100% methanol at a flow rate of 4ml/min prior to filtration of the following sample. Analyses of samples (standards and actual samples) followed by the injection of blanks ensures the absence of carry-over between samples, and therefore validates the back-washing program used. A reverse phase assay consisting of a Gilson HPLC system equipped with a Zorbax RX-C8 column (4.6 m m x 25 cm) (Mac-Mod Analytical, Chadds Ford, PA, USA), was employed for the separation of the 6-bromo-b-tetralone and the 6-bromo-,Q-tetralol. Separation was achieved by employing a mobile phase comprised of acetonitrile

TABLE 1. Yeaststrainsutilized in creatingthe microbiallibrary Yeast strain Aciculoconidium Ambrosiozyma Arxiozyma (2) Brettanomyces (2) Bullera (2) Candida (55) Clavispora (2) Cryptococcus (8) Cystofilobasidium Debaryomyces (8) Dipodascuc (2) Filobasidium Geotrichum Guilliermondella Hanseniaspora (2) Hansenula (17) Zssatchenkia(2) Kloeckera (2) Kluyeromyces (11) Lipomyces Malassezia

Metschnikowia (6) Moniliella Nadsonia Pachysolen Pichia (23) Rhodosporidium Rhodotorula (6) Saccharomyces(6) Schwanniomyces (2) Sporidiobolus Sporobolomyces Sterigmatomyces Sterigmatosporidium Sympodiomyces Taphrina Torulaspora (4) Trichosporon (6) Trigonopsis Wingea Zygosaccharomyces

(A) and acidified water (0.1% phosphoric acid) (B). An isocratic gradient was utilized at a flow rate of 1.5 ml/ min and detection was performed at 220nm at 22°C. Under these conditions, the alcohol and the ketone eluted after 4.9 min and 6.5 min, respectively. RESULTS AND DISCUSSION Procedures and system development To ensure that this microbial collection would likely support the identification of a valuable biocatalyst for asymmetric bioreductions, most of the strains tested here had demonstrated bioreducing activity during previous screens conducted in our laboratory (7-14). The origin of the microbes included in this selective library is described in the Materials and Methods section, and Tables 1 and 2 list the microorganisms utilized to create the respective yeast and rhodococci libraries. The libraries were made up of 24-well templates, each containing a different strain (Fig. 2). A total of eight yeast and two rhodococci templates were created (containing 192 and 48 strains, respectively) and were made in quadruplicate so that a template would be readily available for use at any time. The templates were stored, after the addition of an equal volume of 50% glycerol, at -70°C. The templates were thawed to allow manual pipetting and were immediTABLE 2.

Rhodococcistrainsutilized in creatingthe microbiallibrary

Rhodococcus strain Rhodococcus australis Rhodococcus coprophilus Rhodococcus equi subsp. (3) Rhodococcus erythropolis subsp. (5) Rhodococcus fascians subsp. (4) Rhodococcus globerulus subsp. (5) Rhodococcus marinonascen Rhodococcus opacus subsp. (2) Rhodococcus rhodnii subsp. (2) Rhodococcus rhodochrous subsp.(13) Rhodococcus ruber Rhodococcus species(9) Rhodococcus zopfii

RAPID MICROBlAL SCREENING PROCEDURE

369

Add 5 ml 50% Glycerol 4 Incubate at 29°C For 48 h

U t rozen vlal Strain 1

2 ml

1esr 1uoe 5 ml Medium

Master Template Stored at -70°C

/

l000000

+

Assay Plate

’

000000 000000 ~000000~24

Frofl”ial0 lncu~~~t”““, 1 2m, 1

24

4

2 ml Medium, Microstir Bar

Test Tube

Strain 24

5 ml Medium

t Add 5 ml 50% Glycerol

FIG. 2. Creation of master templates and assay plates. Master template construction: strains were preserved as frozen suspensionsin 20% glycerol at - 70°C. A l-ml volume of the frozen suspension was used to inoculate a 25ml test tube containing 5 ml of Sabouraud dextrose broth (SDB) (30 g/[) or tryptic soy broth (TSB) (30 g/[) for the cultivation of yeast and rhodococci, respectively. After 48 h of incubation at 29°C on a shaker operated at 220 rpm (throw 2 in), 5 ml of 50% glycerol were added to each tube. From each tube, 2 ml of culture were transferred into one well in a polystyrene 24-well cell culture. This procedure was repeated for each culture and the completed plates were stored at - 70°C until use. Assay plate construction: from a master template, a 200 ~1 aliquot of each well was used to inoculate an autoclaved polypropylene 24-well plate containing 2 ml of SDB or TSB and a 2 m m x 7 m m micro stir bar. After 48 h of incubation at 29°C on a stir plate, 2 mg of ketone in 40 pl of ethanol (2% v/v ethanol solution) were added to each well of the plate utilizing a multichannel pipettor. The plates were returned to the same incubation conditions as above.

ately returned to the freezer. During the thawing, we observedthat the top of the well would thaw first while the bottom remained frozen. This allowed sampling without completethawing of the template wells. No loss of viability in the templates were observedduring this and other experimental work (representing several freeze thaw cycles). The autoclaved polypropylene 24-well plates which we utilized in the screening procedure were selected for severalreasons. The large diameter of the well (1.5 cm) enabled sufficient stirring of the cultivation medium with a m icro stir bar, which could be added to the wells before sterilization. This systemensuredthat adequatem ixing, which allowed for both oxygen transfer and sub-

FIG. 3. Photograph of an assay plate in which alternating wells were inoculated with the strain to test for cross-contamination.

strate m ixing, was indeed achieved. The large volume capacity (10 m l) helped eliminate the possibility of crosscontamination since only 2 m l of medium were dispensed in each well. The lack of cross-contaminationwas indeed verified by cultivating three plates containing Sabouraud Dextrose Broth in which alternating wells were inoculated with Saccharomyces cerevisiae. After an incubation of 48 h at 29”C, the wells which were inoculated with S. cerevisiae were turbid and cloudy, indicating growth. The un-inoculated wells were clear, indicating a lack of cross-contamination(Fig. 3). System evaluation The speed and efficiency of this automated and m iniaturized m icrobial screeningmethod was evaluated by screeningfor m icroorganismscapable of stereoselectivelyreducing 6-bromo+tetralone (1) to 6-bromo-,%tetralol(2) (Fig. 1). The substrate6-bromo-ptetralone was chosen becauseit is non-water soluble and thus makes the bioconversion conditions more difficult than with a water soluble substrate. In addition, the data generatedfrom a previous screenutilizing a classicalapproach provides a useful comparison with this screening method (9). After screening192 yeast strains, data from the HPLC analysesindicated that a total of 143 strains successfully bioreduced the 6-bromo-/3-tetraloneto 6-bromo-ptetralol, representing a 74% positive hit rate. Of these 143 strains, 32 showed a “high” biotransformation activity (>600 mg/l p-tetralol) while 111 showed “low” biotransformation activity (between300-600mg/l p-tetralol) (Fig. 4). Of the total 48 rhodococci strains screened,46 successfullyperformed the desired bioreduction and of these,33 showeda “high” activity (Fig. 5). In comparison,

370

J.

STAHL ET AL.

FIG. 4. Histogram of yeast strains and resultant bioreduction activities. 192 yeast strains were screened utilizing the automated liquid handling system. Strains which showed high biotransformation activity produced >600 mg/l p-tetralol while strains which showed low biotransformation activity produced between 300-600 mg/l ,8tetralol. q , High bioreduction activity; , low bioreduction activity; w, absence of bioreduction activity.

a classical screen utilizing the same substrate evaluated 80 yeast strains cultivated in shake flasks and obtained a positive hit rate of SO%, with 9 strains producing substantial amounts of /3-tetralol (9). The higher “hit rate” observedwith the yeast strains should not be surprising. As previously mentioned, our m icrobial collection was comprised of many strains selected for their known bioreducing activity (7-14). In addition, the original screenutilized thin layer chromatography, a far less sensitive detection method when compared with the HPLC method used in the present study. In conclusion, we report here an automated m icrobial cultivation/bioconversion procedure, based on a 24-well plate format, which was facile, time saving, and efficient compared to the classical method of screeningutilizing larger volumes (test tubes or shake flasks) and manual sample preparation. This new screening procedure was time saving in that it yielded data in a relatively short amount of time (a few days versus a few weeks). Once the master template was constructed, cultivation plates were easily assembled using a multichannel pipettor; after cultivation, assayplates were createdin m inuteswith the use of the automated system. In comparison, classical methods of screeningrequired hours to manually prepare tubes or flasks for the cultivation and subsequent

FIG. 5. Histogram of rhodococci strains and resultant bioreduction activities. 48 rhodococci strains were screened utilizing the automated liquid handling system. Strains which showed high biotransformation activity produced > 600 mg/l ,9-tetralol while strains which showed low biotransformation activity produced between 300-600 mg/l ,E-tetralol. @, High bioreduction activity; duction activity; n , absence of bioreduction activity.

BIOSCI. BIOENG.,

processingof the samplesfor analysis. The 24-well plate based m icrobial library was easy to manipulate by avoiding the manual labor involved with more conventional methods of screening.The processing and analysesof sampleswere facile due to the multiple capabilities and automation, such as sample processing, filtering, and HPLC analysis, of the Gilson liquid handling autosamplersystem. This improved method of screeningfor bioconversion activity, employing a m icrotiter plate-based m icrobial library and a fully roboticized analytical system, proved to rapidly yield valuable leads which compared advantageously with a more classical approach (9). This method also proved to be robust as it was found to outperform the conventional screeningmethod, even when employing a non-water soluble substrate. This semi-automated screening procedure renders m icrobial screeningfor bioreduction activity far lesstedious and allows for more rapid identification of the desired activity. Routinely utilizing such an approach to m icrobial screeningfor bioconversion activity, with the extension of the biodiversity base of the m icrobial libraries, should support the timely discovery of the desired activities and in turn reduce process development lead time. ACKNOWLEDGMENTS We acknowledge Robert Giacobbe (Merck), Joni Stevens (Gilson), and Bill Tuting (Gilson) for their time, help, and insightfulness with this screening system. REFERENCES 1. Goodhue, C.: The methodology of microbial transformations of organic compounds, p. 9-44. In Rosazza, J. P. (ed.), Microbial tranformations of bioactive compounds. (CRC Press, Boca Raton (1982). 2. Blacker, A, and Holt, R.: Development of a multi-stage chemical and biological process for an optically active intermediate for an anti-glaucoma drug, p. 245-261. In Collins, Sheldrake, and Crosby (ed.), Chirality in industry II. John Wiley and Sons Ltd., New York (1997). 3. Bel-Rhlid, E., Fauve, A., Renard, M., and Veschambre, H.: Microbiological reduction of carbonyl groupings: preparation of stereoisomeric acyclic chiral alpha-dials. Biocatalysis, 6, 319-337 (1992). 4. Pate], R., Banerjee, A., Liu, M., Hanson, R., Ko, R., Howell, J., and Szarka, L.: Microbial reduction of l-(Cfluorophenyl)4-[4-(5-fluoro-2-pyrimidinyl-l-piperazinyl]butan-l-one. Biotechnol. Appl. Biochem., 17, 139-153 (1993). 5. Peters, J., Zelinski, T., and Kula, M.: Studies on the distribution and regulation of microbial keto ester reductases. Appl. Microbial. Biotechnol., 38, 334-340 (1992). 6. Nassenstein, A., Hemberger, J., Schwartz, H., and Kula, M.: Studies on the enzymatic reduction of N-Boc-4S-amino-3-oxo5-phenylpentanoic acid methylester. J. Biotechnol., 26, 183201 (1992). 7. Chartrain, M., Armstrong, J., Katz, L., Keller, J., Mathre, D., and Greasham, R.: Asymmetric bioreduction of a p-ketoester to (R)+hydroxyester by the fungus Mortierella aIpina MF 5534. J. Ferment. Bioeng., 80, 176-179 (1995). 8. Katz, L., King, S., Greasham, R., and Chartrain, M.: Asymmetric bioreduction of ketosulfone to the corresponding transhydroxysulfone by the yeast Rhodotorulla rubra MY 2169. Enzyme Microbial. Technol., 19, 250-255 (1996). 9. Reddy, J., Tschaen, D., Shi, Y. J., Pecore, V., Katz, L., Greasham, R., and Chartrain, M.: Asymmetric bioreduction of a p-tetralone to its corresponding @ ‘)-alcohol by the yeast Trichosporon cupitatum MY 1890. J. Ferment. Bioeng., 81,

VOL.

89, 2000

304-309 (1996). 10. Chartrain, M., McNamara, J., and Greasham, R.: Asymmetric bioreduction of benzyl acetoacetate to its corresponding alcohol, benzyl (S)-(+)-3-hydroxybutyrate by the yeast Can&da schatuvii MY 1831. J. Ferment. Bioeng., 82, 507-508 (1996). 11. Char-train, M., Mathre, D., Reamer, R., and Greasha, R.: Asymmetric bioreduction of cyclohexylphenyl ketone to (+)cyclohexylphenyl alcohol by the yeast Candidu magnoliu MY 1785. J. Ferment. Bioeng., 83, 395-396 (1996). 12. Chat-train, M., Lynch, J., Choi, W. B., Churchill, H., Pate], S., Yamazaki, S., Volante, R., and Greasham, R.: Asymmetric bioreduction of a bisaryl ketone to use corresponding alcohol by the yeast Rhodotorula pilimanae. J. Mol. Catalysis. B. Enzym. (1998). (in press) 13. Stahl, S., Demoto, N., King, A., Greasham, R., and Chartrain, M.: Asymmetic direduction of 1,Zindandione to cis (lS,2R) indandiol by Thrychospora cutaneum MY 1506. J. Biosci. Bioeng. (1999). (in press) 14. Chartrain, M., Roberge, C., Chung, J., McNamara, J., Zhao,

RAPID MICROBIAL SCREENING PROCEDURE

15. 16. 17. 18.

37 1

D., Olewinski, R., Hunt, G., Salmon, P., Roush, D., Yamazaki, S., Wang, T., Grabowski, E., Buckland’, B., and Greasham, R.: Asymmetric bioreduction of (2-(4-nitro-phenyl)IV-(2-oxo-2-pyridin-3-yl-ethyl)-acetamide) to its corresponding (R) alcohol [(R)-N-(2-hydroxy-2-pyridin-3-yl-ethyl)-2-(4-nitrophenyl)-acetamide], employing Candida sorbophila MY 1833. Enz. Microb.Technol., 25, 489-496 (1999). Kenny, B., Bushfield, M., Parry-Smith, D., Fogarty, S., and Treherne, J. M.: The application of high-throughput screening to novel drug discovery. Prog. Drug Res., 51, 245-269 (1998). Lloyd, A.: High-throughput screening. Drug Disc. Today, 3. 566 (1998) Persidis, A.: High-throughput screening. Nat. Biotechnol., 16, 488-489 (1998). Tschaen, D., Abramson, L., Cai, D., Desmond, R., Dolling, U., Frey, L., Karady, S., Shi, Y.-J., and Verhoeven, T.: Asymmetric synthesis of MK-0499. J. Org. Chem., 60, 4324-4330 (1995).