Biotechnology Letters Volume 27, Number 17

Laboratory Scale Bioremediation of Acid Mine Water Drainage from a Disused Tin Mine

(1251 - 1257)

Lawrence Darkwah, Neil A. Rowson, Christopher J. Hewitt DOI: 10.1007/s10529-005-3201-z

Inhibition of Platelet Aggregation of a Mutant Proinsulin Chimera Engineered by Introduction of a Native Lys-Gly-Asp-containing Sequence

(1259 - 1265)

Jian Jing and Shan Lu DOI: 10.1007/s10529-005-3202-y

Construction of an Effective Protein Expression System Using the tpl Promoter in Escherichia coli

(1267 - 1271)

Takashi Koyanagi, Takane Katayama, Ai Hirao, Hideyuki Suzuki, Hidehiko Kumagai DOI: 10.1007/s10529-005-0216-4

Purification and Properties of an N-acetylglucosaminidase from Streptomyces cerradoensis

(1273 - 1276)

Iderval da Silva Junior Sobrinho, Luiz Artur Mendes Bataus, Valéria Ribeiro Maitan, Cirano José Ulhoa DOI: 10.1007/s10529-005-0218-2

A Modified PCR System for Amplifying β-ketoacyl-ACP Synthase Gene Fragments with DNA from Streptomyces luteogriseus

(1277 - 1282)

Feng-Ming Yu, Xin Jiang, Jin-Chuan Wu, Ying-Jin Yuan DOI: 10.1007/s10529-005-3219-2

Molecular Cloning and Tissue Distribution of SF-1-related Orphan Receptors During Sexual Maturation in Female Goldfish

(1283 - 1290)

Cheol Young Choi and Hamid R. Habibi DOI: 10.1007/s10529-005-0220-8

Enhancement of Isoflavone Synthase Activity by Co-expression of P450 Reductase from Rice

(1291 - 1294)

Dae Hwan Kim, Bong Gyu Kim, Hyo Jung Lee, Yoongho Lim, Hor Gil Hur, Joong-Hoon Ahn

DOI: 10.1007/s10529-005-0221-7

Structural Characterization of β-glucans of Agaricus brasiliensis in Different Stages of Fruiting Body Maturity and their Use in Nutraceutical Products

(1295 - 1299)

Carla Maísa Camelini, Marcelo Maraschin, Margarida Matos Mendonça, Cezar Zucco, Antonio Gilberto Ferreira, Leila Aley Tavares

DOI: 10.1007/s10529-005-0222-6

Taxane Production in Suspension Culture of Taxus × Media var. Hicksii Carried Out in Flasks and Bioreactor

(1301 - 1304)

Katarzyna Syklowska-Baranek and Miroslawa Furmanowa DOI: 10.1007/s10529-005-0223-5

Regio- and Stereo-selective Hydroxylation of Abietic Acid Derivatives by Mucor circinelloides and Mortierella isabellina

(1305 - 1310)

Koichi Mitsukura, Takeshi Imoto, Hirokazu Nagaoka, Toyokazu Yoshida, Toru Nagasawa DOI: 10.1007/s10529-005-3224-5

Increased Conformational and Thermal Stability Properties for Phenylalanine Dehydrogenase by Chemical Glycosidation with End-group Activated Dextran

(1311 - 1317)

Reynaldo Villalonga, Shinjiro Tachibana, Yunel Pérez, Yasuhisa Asano DOI: 10.1007/s10529-005-3225-4

Production of Fungal Biomass Immobilized Loofa Sponge (FBILS)-discs for the Removal of Heavy Metal Ions and Chlorinated Compounds from Aqueous Solution

(1319 - 1323)

M. Iqbal, A. Saeed, R.G.J. Edyvean, B. O’Sullivan, P. Styring DOI: 10.1007/s10529-005-0477-y

A Metal Ion as a Cofactor Attenuates Substrate Inhibition in the Enzymatic Production of a High Concentration of

(1325 - 1328)

Kazuaki Yoshimune, Ai Hirayama, Mitsuaki Moriguchi DOI: 10.1007/s10529-005-0480-3

Characterization of an Extracellular Serine Protease Gene from the Nematophagous Fungus Lecanicillium psalliotae

(1329 - 1334)

Jinkui Yang, Xiaowei Huang, Baoyu Tian, Hui Sun, Junxin Duan, Wenping Wu, Keqin Zhang DOI: 10.1007/s10529-005-0482-1

Author’s Quick Check Checklist for Preparing Manuscripts for Submission (1335 - 1335)

DOI: 10.1007/s10529-005-1074-9

Laboratory scale bioremediation of acid mine water drainagefrom a disused tin mine

Lawrence Darkwah, Neil A. Rowson & Christopher J. Hewitt*Centre for Formulation Engineering, Biochemical Engineering, School of Engineering (Chemical Engineering),The University of Birmingham, Edgbaston, B15 2TT, UK*Author for correspondence (Fax: +44-121-414-5324; E-mail: c.j.hewitt@bham.ac.uk.)

Received 6 May 2005; Revisions requested 2 June 2005; Revisions received 13 June 2005; Accepted 14 June 2005

Key words: acid mine drainage, Acidithiobacillus ferrooxidans, bioremediation, laboratory scale, microbialcatalysis

Abstract

Real acidic mine-water drainage was seeded with Acidithiobacillus ferrooxidans to catalyse the removal ofiron contained therein. The addition of At. ferrooxidans increased metal precipitation kinetics anddecreased the water iron content by �70%. Supplementing non-sterile mine water with a bacterial growthmedium accelerated metal removal by indigenous micro-organisms both at the 500 ml shake-flask and 5 lbioreactor scale.

Introduction

Both alkaline and acid mine water drainage(ADM) is a growing world-wide problem forboth working and abandoned mines as well ascolliery spoil heaps due to the highly toxic prod-ucts generated in both underground and surfacemining (Johnson 2003). In particular, acidicdrainages originate from the exposure of sulphi-dic mineral surfaces to O2, which results in theformation of soluble sulphates. On contact withwater, these minerals (mostly with a high ferrousiron content) become oxidised, usually withmicrobial catalytic enhancement, producing ferricions and H2. These ions when leached intostreams cause the water to become more acidic,normally reaching pH values <3. Additionally,other metal ions, such as Cd, Cu, Mn, Al andAs, also leach into the AMD, at final concentra-tions far above permissible legal levels.

Wheal Jane, an important cassiterite (the mainmineral ore for tin) mine within the Carnon Valley(Cornwall, UK) provided the AMD for use in thisstudy. The workings that make up this disused

mine (work finished finally in c. 1991) extend to adepth of 450 m below ground being partially floo-ded and very wet. This is probably due to waterseeping from interconnected workings that histori-cally produced pyrite and arsenopyrite. During themines working life such water, partially treated,was actively pumped and discharged into the Riv-er Carnon (UK) but, when government fundingwas withdrawn, all treatment ceased so the risingmine water, with a pH of 2.8 and a high metalcontent, was passively discharged into the RiverCarnon (UK). Further attempts to treat the waterwith CaCO3 (lime) before it was pumped into theexisting Wheal Jane tailings dam were made but,when this stopped for technical reasons in January1992, 50 million litres of acidic metal laden waterwere accidentally released into the river. Since thiswater contained iron hydroxides at high levels, avery visual contamination of the local river causedsignificant public pressure for a long-term solutionto the problem of AMD in the UK (Banks et al.1997, Somerfield et al. 1994).

Many methods have been investigated,designed and used in remediating AMD. Most

Biotechnology Letters (2005) 27: 1251–1257 � Springer 2005DOI 10.1007/s10529-005-3201-z

common, are the passive (wetland) methods, con-ventional active methods of adding limestone,quicklime or NaOH (caustic soda) or soda ash topromote metal precipitation and other biologicalroutes such as biosorbents and rotating biologi-cal contactors (Wildeman 1993, Groudev et al.1999, Shutes 2001, Brown et al. 2002).

The bioremediation approach to AMD remedi-ation, which incorporates both biological and pas-sive chemical processes, is a proven alternative toconventional environmental cleanup technologies(Macaskie et al. 1995, Boswell et al. 1998, Bon-throne et al. 2000). The naturally occurring, acido-philic bacterium Acidithiobacillus ferrooxidans haslong been used for the bioleaching of copper fromchalcopyrite (CuFeS2) because it can oxidise fer-rous iron at a high rate, exhibits rapid growth andis able to tolerate high iron concentrations (Kelly& Wood 2000). For these reasons, it has also beenstudied extensively for use in solving environmen-tal problems involving iron rich AMD. Thereforein this work we seek to investigate and enhancethe capability of the indigenous mostly acidophiliciron and sulphur-oxidising bacteria (Hallberg &Johnson 2005, Johnson 2003) in the mine water tooxidise the ferrous iron in solution to ferric iron.The latter quickly precipitates at the low pH ofthe AMD used here and is easily removed usingconventional technologies such as sedimentation(enhanced by flocculation or the use of hydro-cy-clones), froth flotation or filtration. This was donefirstly, by supplementing the water with a bacterialgrowth medium and secondly by seeding the waterwith a culture of the bacterium At. ferrooxidans atboth the shake flask and laboratory scale.

Materials and methods

Acid mine water (AMD)

AMD was obtained from the Wheal Jane mine(Cornwall, UK). Fresh AMD was collected inpre-sterilized plastic bottles. This water was thenstored at 4 �C and used for experiments less than5 days from the date of collection.

Bacterial strain and growth medium

At. ferrooxidans (ATCC 19859) was used to inoc-ulate the AMD where appropriate. Inoculum

(average �1.7�107 cells/ml) was produced onATCC medium 64 (GM) and maintained at4 �C. GM is made up from two solutions. Solu-tion A, 0.4 g (NH4)2SO4, 0.2 g KH2PO4 and0.08 g MgSO4Æ7H2O made up in 400 ml distilledwater. Solution B, 10.0 g FeSO4Æ7H2O made upin 100 ml distilled water and acidified with 1 ml0.5 M H2SO4. Solution A was autoclaved at121 �C for 20 min whilst Solution B was filter-sterilised (0.2 lm) and the two combined asepti-cally. The pH of the GM was 2.8.

AMD experiments

Aliquots of AMD were either heat-sterilised(121 �C for 20 min) or not. These are referred toas heat-sterilised mine water (HSMW) and non-sterilised mine water (NSMW), respectively. Bothof these (NSMW and HSMW) were supplementedwith 50%, 80% and 90% (v/v) GM or distilledwater where appropriate. A total volume of 50 mlof each of the above solutions was put into dupli-cate 500 ml shake flasks and shaken on an orbitalshaker at 100 rpm and 30 �C. Similar flasks wereinoculated with 10% (v/v) At. ferrooxidans whereappropriate and also incubated at 30 �C and100 rpm. In a similar way, laboratory scale biore-actor studies were carried out in a 5 l cylindricalglass vessel, (157 mm diameter � 260 mm totalheight), with a working volume of 4 l. The vesselwas fitted with one 76 mm diameter, four bladedRushton turbine which was situated 30 mm abovethe bottom of the vessel. The vessel also had fourequally spaced vertical baffles, width 17 mm. Thevessel was equipped for the measurement of pHand temperature. Cultures were run at 30 �C withan impeller speed of 100 rpm.

Analytical methods

For all experiments, 10 ml samples were takenand filtered through a 0.45 lm cellulose acetatemembrane filter to remove precipitates (mostlyinsoluble ferric iron species) and then acidifiedwith 1.5 ll conc. HNO3 per 1 ml sample andkept at 4 �C for total soluble iron (mostly fer-rous iron species) content analysis using anatomic absorption spectrophotometer (Model751, Instrumentation Laboratory, USA) (Green-berg et al. 1992). For shake flask experiments,the remainder of the sample was used to measure

1252

temperature, pH and in all cases reduction–oxi-dation (redox) potential using a Water TestMeter calibrated as per the manufacturersinstructions (Hannah Instruments, UK).

Results and discussion

A series of experiments was carried out in 500 mlshake-flasks and a 5 l bioreactor where the AMDwas sterilised or not, supplemented with variousproportions of GM (50%, 80%, 90% v/v) or dis-tilled water or not and inoculated with a labora-tory strain of At. ferrooxidans or not. Sinceconventional techniques for following microbialactivity were not suitable for use in this case,because of the very complex particulate nature ofthe GM and mine water mixture used, metabolicactivity can be inferred from changes in redoxpotential (mV) (Nemati & Harrison 2000, Bhattiet al. 2001, Medrano-Roldan et al. 2001) changesin pH and a decreases in total (ferrous) iron insolution. Also, because AMD is an environmen-tal sample, its composition varies with the time

of year but, even with these different startingconcentrations of iron (also due to the differingproportions of GM used), reproducible measure-ments of pH, redox potential (mV), total iron insolution mg/l were made for duplicate experi-ments (not all data shown, Figures 1–5). Initialstudies showed that the At. ferrooxidans strainused here could not grow in undiluted AMD orAMD diluted with sterile distilled water (50% v/v) but exhibited a typical growth profile on theGM. This confirmed the work of others (Denni-son et al. 2001) that showed that the AMD needsto be supplemented with a phosphate source forAt. ferrooxidans to grow.

The effect of both heat-sterilisation and sup-plementing the AMD with 90% (v/v) of GMcan be seen in Figure 1. In the case of theHSMW, the total iron in solution remainedhigh (3300–4500 mg/l), the redox potential wasalways below 355 mV and the pH remainedabove 2.3 indicating little microbial activity andferrous iron removed from solution through-out. However, in the case of the NSMW, theincrease in redox potential to a maximum of

Time (days)

0 2 4 6 8

Hp

2.0

2.2

2.4

2.6

2.8

3.0

3.2

Red

oop

xet

itnal

m()

V

200

250

300

350

400

450

500

550

Tota

l iro

n in

sol

utio

n m

g/l

0

1000

2000

3000

4000

5000

pH HSMWpH NSMWRedox potential (mV) HSMWRedox potential (mV) NSMWTotal iron in solution mg/l HSMWTotal iron in solution mg/l NSMW

Fig. 1. pH, redox potential (mV) and total iron in solution mg/l shake flask profiles for HSMW and NSMW both supplementedwith 90% GM.

1253

520 mV with a concomitant decrease in pH to2.2 indicated indigenous microbial activity(Nemati & Harrison 2000, Bhatti et al. 2001,Medrano-Roldan et al. 2001), which was capableof oxidising ferrous iron to ferric iron hencelowering the total iron in solution by �70%after 6 days’ incubation.

The effect of inoculating the HSMW andNSMW supplemented with 90% (v/v) GM withAt. ferrooxidans is illustrated in Figure 2. In bothcases the redox potential had risen above 550 mVand the pH had dropped to <1.8 by the end ofday 2. This indicated sufficient microbial activityto reduce the total iron in solution to below1300 mg/l (>70% reduction) in both cases duringthe same time period. Essentially identical experi-ments were carried out in a 5 l stirred (100 rpm)tank bioreactor and similar results were obtained.

The effect of inoculating the NSMW supple-mented with 80% (v/v) GM with At. ferrooxidansis illustrated in Figure 3. In the case of the NSMWsupplemented with 80% (v/v) GM withoutinoculation, the total iron in solution decreased

only slightly (5% by day 5), the redox potentialrose steadily from 200 to 320 mV and the pH fellfrom 2.9 to 2.4 indicating diminished microbialactivity and ferrous iron oxidtion throughout.However, in the case of the inoculated NSMWsupplemented with 80% (v/v) GM, the redox po-tential increased to a maximum of 600 mV by day2 and a steady decrease in pH to 1.7 throughout.This indicated microbial activity which was capa-ble of lowering the total iron in solution by �45%after 2 days’ incubation increasing to 56% by day5. Essentially, identical experiments were carriedout in a 5 l stirred (100 rpm) tank bioreactor andsimilar results were obtained.

The effect of adding 50% (v/v) GM to bothNSMW and NSMW inoculated with At. ferroox-idans in a 5 l stirred (100 rpm) tank bioreactor isshown in Figure 4. In the case of the NSMWsupplemented with 50% (v/v) GM without inoc-ulation, the total iron in solution decreased onlyslightly initially but there was a greater second-ary decrease starting on day 6 resulting in a 50%drop in total iron in solution by day 8. A similar

Time (days)

0 2 4 6 8

pH

1.6

1.8

2.0

2.2

2.4

2.6

odeR

netopx

tm(lai

)V

300

350

400

450

500

550

600

650o

Tnorilat

losni

unoit

l/gm

500

1000

1500

2000

2500

3000

3500

4000

4500

pH HSMWpH NSMWRedox potential (mV) HSMWRedox potential (mV) NSMWTotal iron in solution mg/l HSMWTotal iron in solution mg/l NSMW

Fig. 2. pH, redox potential (mV) and total iron in solution mg/l shake flask profiles for HSMW and NSMW both inoculated withAt. ferrooxidans and supplemented with 90% GM.

1254

trend was observed for pH but the redox poten-tial increased gradually throughout. In the caseof the NSMW supplemented with 50% (v/v) GMwith inoculation, the total iron in solution haddecreased by �80% by day 2. This was followedby a concomitant decrease in pH and increase inredox potential. Essentially identical experimentswere carried out in 500 ml Erlenmeyer shakeflasks and similar results were obtained.

The effect of adding 10% (v/v) GM to bothNSMW and NSMW inoculated with At. ferroox-idans in a 5 l stirred (100 rpm) tank bioreactor isshown in Figure 5. In the case of the NSMWsupplemented with 10% (v/v) GM withoutinoculation, the total iron in solution graduallydeclined with an �80% drop by day 6. A similartrend was observed for pH but redox potentialincreased gradually, throughout. In the case ofthe NSMW supplemented with 10% (v/v) GMwith inoculation, similar results were obtainedand the total iron in solution had decreasedby �80% by day 6. This was followed by a con-comitant decrease in pH and increase in redoxpotential. Essentially identical experiments were

carried out in 500 ml Erlenmeyer shake flasksand similar results were obtained.

Conclusions

Supplementing non-sterile mine water (NSMW)with growth medium (GM) at varying propor-tions, enhances natural microbial activity andfacilitates the removal of total iron from solutionfrom real acidic mine water such as that obtainedfrom the Wheal Jane mine even when its compo-sition varies according to the time of year and,not unsurprisingly, that this effect increases withincreasing proportions of the GM. Further thatthe addition of a pure culture of At. ferrooxidansto either heat-sterilized mine water or NSMWfurther enhances the ferrous iron oxidation kinet-ics. The subsequent hydrolysis of the resultantferric iron causes it to precipitate very rapidlyand it can then be removed from suspension byconventional techniques, in this case, filtration.Process performance seems to be independent ofthe scale of operation examined here (shake-flask

Time (days)

0 2 4 6 8

pH

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

Redox p

ote

ntial (m

V)

100

200

300

400

500

600

700

Tota

l ir

on in s

olu

tion m

g/l

500

1000

1500

2000

2500

3000

3500

4000

4500

pH NSMW not inoculated

pH NSMW inoculated

Redox potential (mV) NSMW not inoculated

Redox potential (mV) NSMW inoculated

Total iron in solution mg/l NSMW not inoculated

Total iron in solution mg/l NSMW inoculated

Fig. 3. pH, redox potential (mV) and total iron in solution mg/l shake flask profiles for NSMW and NSMW inoculated withAt. ferrooxidans, where both were supplemented with 80% GM.

1255

Time (days)0 2 4 6 8

Hp

1.8

2.0

2.2

2.4

2.6

2.8

laitnetopxode

R(m

)V

100

200

300

400

500

600

700rilato

Ts

nino

ol/g

mnoitul

0

200

400

600

800

1000

1200

1400

pH NSMW not inoculatedpH NSMW inoculatedRedox potential (mV) NSMW not inoculatedRedox potential (mV) NSMW inoculatedTotal iron in solution mg/l NSMW not inoclulatedTotal iron in solution mg/l NSMW inoculated

Fig. 4. pH, redox potential (mV) and total iron in solution mg/l 5 l laboratory scale bioreactor profiles for NSMW and NSMWinoculated with At. ferrooxidans where both were supplemented with 50% GM.

Time (days)0 2 4 6 8

pH

2.0

2.2

2.4

2.6

2.8

3.0

3.2

3.4

xodeR

optnet

Vm(lai

)

200

250

300

350

400

450

500

550

600

riTo

tal

nino

itulosno

l/gm

0

100

200

300

400

500

600

pH NSMW not inoculatedpH NSMW inoculatedRedox potential (mV) NSMW not inoculatedRedox potential (mV) NSMW inoculatedTotal iron in solution mg/l NSMW not inoculatedTotal iron in solution mg/l NSMW inoculated

Fig. 5. pH, redox potential (mV) and total iron in solution mg/l 5 l laboratory scale bioreactor profiles for NSMW and NSMWinoculated with At. ferrooxidans where both were supplemented with 10% GM.

1256

and 5 l laboratory scale) and demonstrates thepotential of using such a system for treating realacid mine water drainage containing a highferrous iron content.

Acknowledgements

The authors are grateful to United Utilities(Cornwall, UK) for allowing samples to betaken from the Wheal Jane mining facility. Thiswork was funded by the Ghana ScholarshipAgency through the Commonwealth ScholarshipCommission.

References

Banks D, Younger PL, Arnesen RT, Iversen ER, Banks SB(1997) Mine-water chemistry: the good, the bad and the ugly.Environ. Geol. 32: 157–174.

Bhatti TM, Bigham JM, Tuovinen OH (2001) Bacterial andchemical oxidation of marcasite and pyrite. In: CiminelliVST & Garcia JR, eds. International BiohydrometallurgySymposium Proceedings, Process Metallurgy, Part A, Am-sterdam: Elsevier, pp. 617–625.

Bonthrone KM, Quarmby J, Hewitt CJ, Allan VJM, Paterson-Beedle M, Kennedy JF, Macaskie LE (2000) The effect of thegrowth medium on the composition and metal bindingbehaviour of the extracellular polymeric material of a metal-accumulating Citrobacter sp. Environ. Technol. 21: 123–134.

Boswell CD, Hewitt CJ, Macaskie LE (1998) An application ofbacterial flow cytometry: Evaluation of the toxic effects offour heavy metals on Acinetobacter sp. with potential forbioremediation of contaminated wastewaters. Biotechnol.Lett. 20: 857–863.

Brown M, Barley B, Wood H (2002) Minewater Treatment:Technology, Application and Policy, Dorchester, UK:IWAPublishing.

Dennison FD, Sen AM, Hallberg KB, Johnson DB (2001)Biological versus abiotic oxidation of iron in acid mine

drainage waters: An important role for moderately acido-philic, iron-oxidising bacteria. In: Ciminelli VST & GarciaJR, eds. International Biohydrometallurgy Symposium Pro-ceedings, Process Metallurgy, Part A, Amsterdam: Elsevier,pp. 493–501.

Greenberg AE, Clesceri LS, Eaton AD (1992) Standard methodsfor the examination of water and wastewater, 18USA:Amer-ican Public Health Association.

Groudev SN, Bratcova SG, Komnitsas K (1999) Treatment ofwaters polluted with radioactive elements and heavy metalsby means of a laboratory passive system. Miner. Eng. 12:261–270.

Hallberg KB, Johnson DB (2005) Microbiology of a wetlandecosystem constructed to remediate mine drainage from aheavy metal mine. Sci. Total Environ. 338: 53–66.

Johnson DB (2003) Chemical and microbiological characteris-tics of mineral spoils and drainage waters at abandoned coaland metal mines. Water Air Soil Poll. 3: 47–66.

Kelly DP, Wood AP (2000) Reclassification of some species ofThiobacillus ferrooxidans to the newly designated generaAcidithiobacillus, Halothiobacillus and Thermithiobacillus.Int. J. Syst. Bacterial. 50: 511–516.

Macaskie LE, Hewitt CJ, Shearer JA, Kent CA (1995) Biomassproduction for the removal of heavy metals from aqueoussolutions at low pH using growth-decoupled cells of aCitrobacter spp. Int. Biodeter. Biodegr. 73–92.

Medrano-Roldan H, Chavez-Gonzalez BP, Solis-Soto A,Morales-Castro J, Ochoa-Martinez LA, Rocha-Fuentes M,Pereyra-Alferez B, Gala-Wong JL, Ramirez-Rodriguez GD,Davila-Flores RT (2001) Growth of a native Thiobacillusferrooxidans strain in mine tailings. In: Ciminelli VST &Garcia JR, eds. International Biohydrometallurgy Sympo-sium Proceedings, Process Metallurgy, Part A, Amsterdam:Elsevier, pp. 587–593.

Nemati M, Harrison STL (2000) A comparative study onthermophilic and mesophilic biooxidation of ferrous iron.Min. Eng. 13: 19–24.

Shutes RBE (2001) Artificial wetlands and water qualityimprovement. Environ. Int. 26: 441–447.

Somerfield PJ, Gee MJ, Warwrick RM (1994) Benthiccommunity structure in relation to an instantaneousdischarge of waste water from a tin mine. Mar. Pollut.Bull. 28: 363–369.

Wildeman TR (1993) Passive bioremediation of metals fromwater using reactors or constructed wetlands. In: Means J &Hinchee R, eds. Emerging Technology for Bioremediation ofMetals, Boca Raton: Interpharm/CRC, pp. 13 –25.

1257

Inhibition of platelet aggregation of a mutant proinsulin chimeraengineered by introduction of a native Lys-Gly-Asp-containing sequence

Jian Jing* & Shan LuDepartment of Biochemistry and Biotechnology, Laboratory of Biotechnology and Protein Engineering,Beijing Normal University, 100875, Beijing, China*Author for correspondence (Fax: +86-10-58807365; E-mail: jjing@bnu.edu.cn)

Received 13 January 2005; Revisions requested 2 February 2005; Revisions received 14 June 2005; Accepted 14 June 2005

Key words: glycoprotein IIb/IIIa receptor, inhibition of platelet aggregation, KGD (Lys-Gly-Asp) motif,proinsulin mutant

Abstract

An eight amino acid sequence, CAKGDWNC, from disintegrin barbourin, was introduced into an inactivehuman proinsulin molecule between the B28 and A2 sites to construct a chimeric, anti-thrombosis re-combinant protein. The constructed Lys-Gly-Asp (KGD)-proinsulin gene was expressed in Escherichia coliand then purified. The KGD-proinsulin chimera protein inhibits human platelet aggregation, induced byADP, with an IC50 value (molar concentration causing 50% inhibition of platelet aggregation) of 830 nMand demonstrates also specific affinity to glycoprotein IIb/IIIa receptor. Its insulin receptor binding activityremaines as low as 0.04% with native insulin as a control.

Introduction

The functional motif of Arg-Gly-Asp/Lys-Gly-Asp (RGD/KGD) exists in many adhesion pro-teins, such as fibrinogen, fibronectin and vonWillebrand factor. Platelet aggregation is animportant step in thrombotic events and requiresthe binding of fibrinogen to glycoprotein IIb/IIIareceptor on activated platelets. Most proteins orpeptides from snake and leech venom (Musial &Niewiarowski 1990, Scarborough et al. 1991),known as the disintegrin family, containing RGDmotifs or KGD motifs, are potential antagonistsof platelet aggregation. The RGD motif occursin many disintegrins, but the KGD motif hasonly been found in the disintegrins barbourin(Scarborough et al. 1991) and ussuristatin-2 (Ki-yotaka & Shigeyuki 1999). RGD and KGD mo-tifs have almost the same biological activity toinhibit platelet aggregation but the KGD motif isglycoprotein IIb/IIIa receptor-specific as com-pared with the RGD motif. If KGD-containing

proteins or peptides can simulate the conforma-tion of native functional KGD motif, they maybe also glycoprotein IIb/IIIa receptor-specific andhave high potency to inhibit platelet aggregationby interacting specificially with glycoprotein IIb/IIIa receptor on the platelet surface (Wittig et al.1998). Minoux et al. (2000) have revealed thatfunctional KGD motif is usually located at thetop of solvent-accessible, highly flexible loopstructure and always tends to assume a b-turnconformation (Minoux et al. 2000).

Insulin is produced by cleavage of the C-pep-tide of proinsulin, the precursor of insulin hor-mone. The C-peptide, exposed on the surface ofproinsulin, is a loose loop structure (Weiss et al.1990) which is similar to the KGD motif of na-tive KGD-containing disintegrins. To obtain anovel platelet aggregation inhibitor, an eightamino acid peptide, CAKGDWNC, originatedfrom the functional motif of the disintegrin, bar-bourin (Scarborough et al. 1991), was selected toreplace the C-peptide of human proinsulin to

Biotechnology Letters (2005) 27: 1259–1265 � Springer 2005DOI 10.1007/s10529-005-3202-y

construct a chimeric anti-thrombosis peptide.The structure of chimeric KGD-proinsulin mole-cule was simulated, inhibitory activity of plateletaggregation, glycoprotein IIb/IIIa receptor bind-ing specificity and insulin hormone activity weredetermined.

Materials and methods

Simulation of the KGD-proinsulin structure

The structure of the KGD-proinsulin was simu-lated with sgi workstation using Insight II soft-ware. Structural simulation was carried out withthe data of NMR structure of native humaninsulin (Protein Data Bank, Brookheaven, CA).The Biopolymer module in Insight II was em-ployed to add the CAKGDWNC sequence be-tween the B28 and A2 sites of insulin structure.At the same time, the A6Cys and A11Cys ofproinsulin were replaced by Ser to delete theintra-A chain disulfide bond. The Builder modulewas employed to carry out structural optimiza-tion (iterations: 1000; derivative: 0.01).

Construction of the KGD-proinsulin mutant gene

The standard polymerase chain reaction methodwas employed to construct mutant KGD-con-taining proinsulin gene. The proinsulin gene withA6 and A11 Cys to Ala mutations was used asthe template (Dai & Tang 1996). The sequencesof the primers were as follows: AGA-AAGACGTTGGAG (primer l, 5¢ primer); GAC-TAATATTACGTCGACTCCCAAATAACCAATATTCCCCCAGCACTGAAGCTGCTACGG-TGGTAGAA (primer 2, mutant primer); ATTA-GCTAGGTGGCC (primer 3, complementary tothe mutant primer); GTCTGATCCCCGGCA(primer 4, 3¢ primer). Primer 1 and 2 were usedto obtain upstream DNA fragment, and primer 3and 4 were used to obtain downstream DNAfragment. Then, the above two DNA fragmentswere mixed and annealed with the primer 1 and4 to obtain full-length KGD-proinsulin gene. Therecombinant gene was confirmed by DNAsequencing and cloned into an expression vectorpET21a at BamHI and HindIII sites under thecontrol of a T7 promoter. The constructed re-combinant expression vector was named as

pEK214 and used to transform Escherichia coliBL21(DE3)pLysS for further expression andpurification of the KGD-proinsulin protein.

Expression and purification of the KGD-proinsulin

The procedures for expression and refolding ofthe KGD-proinsulin were as described previouslywith some modifications (Jing & Tang 2000).After Sephacryl S200 separation, the pH of theKGD-proinsulin fraction was adjusted to 3.5with 2 M HCl, and solid NaCl to 12.5% (w/v) tosalt-out the KGD-proinsulin. The pellet was col-lected by centrifugation and dissolved in 0.05 M

Tris–HCl and 40% (v/v) 2-propanol, pH 7.0, andwas further purified by DEAE Sephadex ion-exchange chromatography.

Characterization of the KGD-proinsulin

Amino acid composition and circular dichroism(CD) analysis were done as described (Dai &Tang 1996). The molecular weight of the KGD-proinsulin was determined by mass spectrometry.Inhibition of platelet aggregation was determinedas follows. Fresh human blood was anticoagulat-ed with 0.01 M sodium citrate (10% v/v), pH 7.4,and centrufuged at 150 g at 25 �C for 10 min.The supernatant was collected as platelet-richplasma. The protein samples were dissolved in0.9% NaCl. Platelet-rich plasma, 300 ll, wasincubated with 25 ll of various concentrations ofprotein samples at 37 �C for 2 min before addi-tion of 25 ll of 100 lM adenosine 5¢-diphosphate.The inhibition of platelet aggregation was deter-mined by light transmission measurement. TheIC50 value was calculated by linear regression.Cell attachment assay was carried out to deter-mine glycoprotein IIb/IIIa receptor binding speci-ficity. The human melanoma cell line K2 wascultured in modified RPM1640 medium contain-ing 12% fetal calf serum and harvested at sub-confluency with 0.4% trypsin and 1 mM EDTA.Wells (2 cm2) of six-well dishes were treated with50 ll of fibrinogen (50 lg ml)1) or fibronectin(50 lg ml)1) in PBS overnight at 4 �C. A 500 llaliquot of cells (1.5� 105 cells ml)1) was added toeach well. The cells were incubated for 1.5 h at37 �C, then non-adherent cells were removed byaspiration, and adherent cells were fixed andstained with 2% Giemsa solution for 60 min. The

1260

total number of cells in each well was countedmicroscopically. Insulin receptor binding assaywas done as following. Insulin receptor was par-tially isolated as crude membranes from humanplacenta. The proteins were diluted into a seriesof concentrations in KRB buffer (0.114 M NaCl,1.2 mM MgSO4, 0.03 M HEPES, 0.05 M KCl,1.2 mM KH2PO4, 1.3 mM CaCl2, 0.01 M NaH-CO3, pH 7.4). To 50 ll of protein samples, anequal volume of properly diluted 125I-insulin inKRB buffer and 6-fold diluted insulin receptor inKRB buffer in 2% of bovine serum albumin wereadded and the samples were mixed. The mixtureswere incubated overnight at 4 �C and centrifugedat 10,000 g for 5 min. The radioactvitiy of thepellets was determined by gamma ray detection.The concentration of native insulin producing50% inhibition of 125I-insulin binding to receptorwas set as 100%, and the activity of the KGD-proinsulin was compared.

Results and discussion

Design of the KGD-proinsulin molecule

Intra-A chain disulfide bond-deleted proinsulinand insulin mutant shows almost no insulinreceptor binding activity, but whole immuneactivity is retained (Dai & Tang 1996). The struc-ture of the intra-A chain disulfide bond-deletedproinsulin mutant is quite similar to that of na-tive proinsulin (Hua et al. 1996). That suggeststhat the intra-A chain disulfide bond-deleted pro-insulin can be adopted as a promising scaffoldfor foreign functional motif to exhibit its activeconformation. The mutant proinsulin with intra-A chain disulfide bond deletion is human originand inactive in vivo, it can also be prepared easilyas recombinant protein. Spatial distance betweenB28 site of the B-chain and A2 site of the Achain is about 5–10 A which is suitable forinserting a small peptide. Also, the proinsulinand native KGD-containing platelet antagonistsare all small proteins rich in disulfide bonds.

The chimeric KGD-proinsulin has the possi-bility to specifically inhibit platelet aggregationand shows no insulin hormone activity as a mu-tant proinsulin scaffold. To reduce further thepossibility of degradation of proinsulin to insulin,B29Lys, a possible protease digestion site was

deleted. A1Gly and B30Thr were replaced by twoCys residues on the two ends of the eight aminoacid KGD-containing peptide. The two Cysintroduced with the KGD-containing peptidemay form a disulfide-bond to improve plateletaggregation inhibitory activity of the KGD-pro-insulin protein. The structural simulation andhomology modeling of the KGD-proinsulin chi-mera was carried out with the sgi-workstationusing Insight II-software based on detailed struc-tural information of mutant proinsulin (Huaet al. 1996) and the disintegrin barbourin(Minoux et al. 2000). The simulated structure(Figure 1) of the KGD-proinsulin chimera pro-tein demonstrates that the CAKGDWNC-peptideintroduced between the B28 site and A2 site ofthe mutant proinsulin with intra-A chain disul-fide-bond deletion exhibits its functional motif tointeract with the glycoprotein IIb/IIIa receptor,know as fibrinogen receptor. The structural prop-erty of the KGD-proinsulin provides necessaryrequirements of recognition and interaction with

Fig. 1. Simulation of three-dimensional structure of theKGD-proinsulin. The structure was simulated with an sgiworkstation using Insight II software. Peptide backbone isshown as a long slender coil shaded in black, with the KGDsequence (including the side-chains) highlighted. The KGD-motif lies on the surface of mutant proinsulin molecule andthe side-chains of the KGD sequence are extended to the out-side of loop structure. The KGD-motif is far away from thecore of mutant proinsulin and exhibit a native-like conforma-tion. Replacement of the C-peptide with KGD-containing se-quence has little influence on the proinsulin structure.

1261

the glycoprotein IIb/IIIa receptor on platelet cellsurface. So there is great possibility to obtain anovel potent anti-thrombosis protein with thespecificity of the glycoprotein IIb/IIIa receptor.

Expression, purification, and characterizationof the KGD-proinsulin

Routine recombinant DNA techniques were usedto construct the chimeric gene and the gene wasover-expressed in Escherichia coli BL21(DE3)-pLyS with constructed expression vectorpEK214. The expression and purification of theKGD-proinsulin is mainly processed accordingto the method described previously (Jing & Tang

2000). The expressed protein was analyzed, asshown in Figure 2a. The expression level wasabout 17% of total cellular proteins based on thespectrophotometric analysis. After isolation ofthe inclusion bodies, unfolding and refolding ofthe KGD-proinsulin protein, Sephacryl S200chromatography was used to separate the re-folded KGD-proinsulin chiemra protein, asshown in Figure 2b. (Peak 2 stands for theKGD-proinsulin chimera protein fraction.) TheKGD-proinsulin was further purified by aDEAE-Sephadex ion-exchange chromatography(Figure 2c). High homogeneity of purified KGD-proinsulin recombinant protein can be deter-mined by both SDS-PAGE (Figure 2a) and

Fig. 2. Purification of the KGD-proinsulin chimera protein. (a) Analysis of expression of recombinant clones by 15% SDS-PAGE.Lane 1, protein molecular weight marker; lane 2, total cellular proteins of cells transformed by pET21a vector; lanes 3 and 4, totalcellular proteins of cells transformed by recombinant expression plasmid pEK214; lane 5, purified KGD-proinsulin after DEAE-Sephadex chromatography. (b) Sephacryl S200 chromatography separation. The column (1� 50 cm) was eluted with 0.05 M

Gly–NaOH buffer (pH 10.8). Peak 2 represents the KGD-proinsulin. (c) DEAE-Sephadex ion-exchange chromatography to furtherpurify the KGD-proinsulin. The column (1� 5 cm) was equilibrated with 0.05 M Tris–HC1 and 40% 2-isopropanol at pH 7.0. TheKGD-proinsulin was eluted out with a linear NaCl gradient from 0 to 0.30 M in a total volume of 75 ml. Peak 2 represents theKGD-proinsulin. (d) Electrophoretic analysis of purified KGD-proinsulin on 12% (v/v) polyacrylamide gel. Lane 1, human proin-sulin; lane 2, purified KGD-proinsulin after DEAE-Sephadex ion-exchange chromatography. The gels were stained with CoomassieBrilliant Blue R-250.

1262

native PAGE (Figure 2d) analyses. The predictedpI of the KGD-proinsulin chimera is about 5.3,so our previous procedure for the purification ofrecombinant human mutant proinsulin (Jing &Tang 2000) can be adopted to purify the mutantKGD-proinsulin chimeric protein. The final re-sults were shown by Figure 2 which is in agree-ment with our expectation.

Amino acid composition analysis of theKGD-proinsulin chimera was carried out alsowith native human proinsulin peptide as a con-trol and the data agrees well with the amino acidresidues numbers calculated by its primary struc-ture (data not shown). CD spectra analysis indi-cated the KGD-proinsulin contains very closesecondary-structure contents compared withhuman insulin (Table 1). This determinationsuggests that our structural simulation of theKGD-proinsulin is very similar to its actualstereo-structure.

The inhibitory activity of adenosine 5¢-diphos-phate-induced human platelet aggregation by the

KGD-proinsulin was tested, as shown in Figure 3.Light transmission of ADP-induced plateletaggregation is illustrated in Figure 3a. TheKGD-proinsulin shows high platelet aggregationinhibitory activity. The IC50 value calculated isabout 830 nM as shown in Figure 3b, which ishigh compared with that of native disintegrins(Gan 1988, Musial & Niewiarowski 1990, Scar-borough et al. 1991). Under the same conditions,both human proinsulin and insulin peptidesshowed no inhibitory activity of the plateletaggregation (data not shown).

The result suggests that the CAKGDWNC-se-quence inserted between the B28 site and the A2site of the mutant proinsulin with intra-A chaindisulfide-bond deletion exhibits its functionalconformation and can interact directly with cor-responding glycoprotein IIb/IIIa receptor on theplatelet cell surface. The potent inhibitory activ-ity of the KGD-proinsulin agrees well with ourexpectation based on the result of structural sim-ulation of the KGD-proinsulin chimera protein.Insulin receptor binding assay showed that theinsulin receptor binding activity of the KGD-proinsulin is only 0.04% of native human insulin,as shown in Figure 4. This agrees with ourexpectation that the KGD-proinsulin would notexhibit insulin hormone activity, so the KGD-proinsulin chimera is a safe and specific anti-thrombosis agent and the mutant proinsulin withintra-A chain disulfide bond deletion only acts as

Table 1. Secondary structure content (%) of the KGD-proin-sulin chimera with CD spectra analysis.

KGD-proinsulin Insulin

a-Helix 15.1 22.8

b-Sheet 22.6 17.1

Random structure 63.3 60.1

Fig. 3. Inhibitory activity analysis of adenosine 5¢-diphosphate-induced platelet aggregation by the KGD-proinsulin. (a) Lighttransmission measurement of platelet aggregation. Three hundred microliters of platelet-rich plasma were incubated with 25 ll ofvarious concentrations of the KGD-proinsulin at 37 �C for 2 min before addition of 25 ll 100 l M ADP. Samples 1–5 representcontrol buffer, 0.1, 0.5, 1.0 and 10 lm KGD-proinsulin, respectively. (b) Inhibition curve of adenosine 5¢-diphosphate-inducedplatelet aggregation of the KGD-proinsulin.

1263

a molecular scaffold. The cell attachment assayshowed that the KGD-proinsulin could inhibitthe attachment of K2 human melanoma cells tofibrinogen-coated wells but not block the cells’attachment to fibronetin-coated wells. This phe-nomena suggests that the KGD-proinsulin exhib-ited high binding activity with glycoprotein IIb/IIIa receptor known as the fribrinogen receptorand low activity with alpha5beta1 receptorknown as the fibronectin receptor (Figure 5).This receptor specificity indicated that the KGD-containing sequence introduced into the mutantproinsulin scaffold exhibited its full activity andthe KGD-proinsulin not only inhibits plateletaggregation, but also recognizes and interactsspecificially with the glycoprotein IIb/IIIa recep-tor. The detailed structure determination of theKGD-proinsulin chimera is carried out currently.

Proinsulin molecule is human origin and hasalmost no immunogenicity to humans. The con-structed KGD-proinsulin with the scaffold of hu-man proinsulin molecule should also notdemonstrate immunogenicity when adopted as atherapeutic agent in vivo. In this respect, the KGD-proinsulin has a considerable advantage over somenaturally occurring antagonists of platelet aggre-gation purified from snake and leech venom (Gan1988, Musial & Niewiarowski 1990, Scarboroughet al. 1991). The KGD-proinsulin chimera is a first

reported KGD-containing recombinant mutantproinsulin with glycoprotein IIb/IIIa receptorspecificity. The KGD-proinsulin is easy to obtainthrough large-scale fermentation as compared withother large recombinant anti-thrombosis proteins(Lee et al. 1993, Nie & Tang 1998).

We conclude from the above result thatthe replacement of the C-peptide with theCAKGDWNC-peptide in mutant proinsulinyields a novel anti-thrombosis agent with potentanti-platelet aggregation activity and the glyco-protein IIb/IIIa receptor binding specificity, butno insulin hormone activity and immunogenicityto human body. This advantage enables theKGD-proinsulin chimera possibility to be adop-ted as a novel therapeutic agent during the treat-ment of thrombosis disease.

Acknowledgments

We thank especially Prof Qun Wei, for helpfuldiscussion about this work; Dr Yin-Ye Wang,for the cell attachment assay; Prof Jian-Xing Ma,from Oklahoma University Health Science Cen-ter U.S.A., for suggestion about this manuscript.This work was supported by a grant from Nat-ure Science Foundation of China (30171100,

0 1 2 3 4 50

20

40

60

80

100

Bin

ding

of

125 I-

insu

lin(%

)

Log concentration(ng ml-1) of KGD-proinsulin 6

Fig. 4. Insulin receptor binding assays of the KGD-proinsu-lin. Vertical axis indicates the binding of 125I-insulin to insulinreceptor, and horizontal axis is logarithm of the native insu-lins or KGD-proinsulin concentrations. The binding of the125I-insulin in the absence of native insulin was set at 100%.(n): native insulin, (d): KGD-proinsulin.

0

20

40

60

80

0 1 10 100 1000

Concentration (nM) of KGD-proinsulin

100

120

Adh

esio

n o

f K

2 m

elan

oma

cell

s(10

3 )

×

%

Fig. 5. Effects of the KGD-proinsulin on adhesion of K2melanoma cell to (n) fibrinogen and (d) fibronectin. TheKGD-proinsulin inhibits the attachment of K2 cell to fibrino-gen-coated wells, while it does not block cell attachment tofibronectin-coated wells. All determinations were made inquadruplicate and each sample was examined a minimum ofthree times.

1264

2001) and a grant from Funds of DistinguishedYoung Scholars of Beijing Normal University toDr Jian Jing (104977, 2003).

References

Dai Y, Tang JG (1996) Characteristic, activity and conforma-tional studies of [A6-Ser, A11-Ser]-insulin. Biochim. Biophys.Acta 1296: 63–68.

Gan ZR (1988) A potent platelet aggregation inhibitor from thevenom of viper, Echis carinatus. J. Biol. Chem. 263: 19827–19832.

Hua QX, Hu SQ, Frank BH, Jia W, Chu YC, Wang SH, BurkeGT, Katsoyannis PG, Weiss MA (1996) Mapping thefunctional surface of insulin by design: structure and functionof a novel A-chain analogue. J. Mol. Biol. 264: 390–403.

Jing J, Tang JG (2000) Platelet aggregation inhibitory activityof mutant proinsulin with C-peptide replaced by CRGDSCsequence. Biotechnol. Lett. 22: 47–52.

Kiyotaka O, Shigeyuki T (1999) Ussuristatin 2, a novel KGD-containing disintegrin from Agkistrodon ussuriensis venom.J. Biochem. 125: 31–35.

Lee G, Chan W, Hurle MR, Desjarlais RL, Waston F, SatheGM, Wetzel R (1993) Strong inhibition of fibrinogen binding

to platelet receptor alpha IIb beta 3 by RGD sequencesinstalled into a presentation scaffold. Protein Eng. 6: 745–754.

Minoux H, Chipot C, Brown D, Maigret B (2000) Structuralanalysis of the KGD sequence loop of barbourin analphaIIbbeta3-specific disintegrin. J. Comput. Aided Mol.Des. 14: 317–327.

Musial J, Niewiarowski S (1990) Inhibition of platelet adhesionto surfaces of extra-corporeal circuits by disintegrins: RGD-containing peptide from viper venoms. Circulation 82: 261–273.

Nie X, Tang JG (1998) RGD-containing trypsin with bothplatelet aggregation inhibitory activity and proteolyticactivity. Biochem. Mol. Biol. Int. 45: 1149–1154.

Scarborough R, Rose JW, Hsu MA, Phillips DR, Fried VA,Campbell AM, Nannizzi L, Charo IF (1991) Barbourin. AGPIIb/IIIa-specific integrin antagonist from the venom ofSistrurus m barbouri. J. Biol. Chem. 566: 9359–9362.

Weiss MA, Frank BH, Khait I, Pekar A, Heine R, Shoelson SE,Neuringer LJ (1990) NMR and photo-CIDNP studies ofhuman proinsulin and prohormone processing intermediateswith application to endopeptidase recognition. Biochemistry29: 8389–8401.

Wittig K, Rothe G, Schmitz G (1998) Inhibition of fibrinogenbinding and surface recruitment of GPIIb/IIIa as dose-dependent effects of the RGD-mimetic MK-852. Thromb.Haemost. 79: 625–630.

1265

Construction of an effective protein expression system using the tplpromoter in Escherichia coli

Takashi Koyanagi1, Takane Katayama2, Ai Hirao1, Hideyuki Suzuki1 &Hidehiko Kumagai22,*1Division of Integrated Life Science, Graduate School of Biostudies, Kyoto University, Sakyo-ku Kyoto,606-8502, Japan2Research Institute for Bioresources and Biotechnology, Ishikawa Prefectural University, Nonoichi-machiIshikawa, 921-8836, Japan*Author for correspondence (Fax: +81-76-227-7557; E-mail: hidekuma@ishikawa-c.ac.jp)

Received after revision 14 June 2005; Accepted 14 June 2005

Key words: Escherichia coli, protein expression system, the tpl promoter, transcriptional regulator TyrR,tyrosine transporter TutB

Abstract

An effective protein expression system was constructed in Escherichia coli using the promoter of thetyrosine phenol-lyase (tpl) gene of Erwinia herbicola. This system involves a mutant form of the TyrRprotein with an enhanced ability to activate tpl and the TutB protein with an ability to transport LL-tyrosine(an inducer of Tpl). The highest expression level obtained for this system was more than twice that obtainedfor the tac system, although it was lower than the level obtained for the T7 system, as revealed with the lac-reporter assay and SDS-polyacrylamide gel electrophoresis.

Introduction

Expression of the tyrosine phenol-lyase (tpl) geneis tyrosine-inducible (Kumagai et al. 1970, Antsonet al. 1993, Suzuki et al. 1993), the basis of whichis the TyrR-mediated transcriptional activation inthe presence of LL-tyrosine (Suzuki et al. 1995, Bai& Somerville 1998, Katayama et al. 1999, Pittardet al. 2005). In the previous study on TyrR ofErwinia herbicola, we obtained the mutant tyrR5allele (tyrRV67A Y72C E201G), the product of whichactivated tpl expression even without the additionof L-tyrosine to the medium (Katayama et al.2000). Since the ability of this mutantTyrRV67A Y72C E201G to activate the tpl promoterwas significant (Koyanagi et al. 2005), we consid-ered that it could be applied to creating a heterol-ogous protein expression system in Escherichiacoli. The construction of the tpl-expression systemand evaluation of the system using the

lac-reporter assay and SDS-PAGE analysis aredescribed in this paper.

Materials and methods

Bacterial strains

The bacterial strains used in this study werederivatives of E. coli K-12. HMS174 (kDE3)(Campbell et al. 1978), JM101 (Yanisch-Perronet al. 1985), TK743 [F) ara D(lac-pro) thiDtyrR::cat+] (Katayama et al. 2000), and deriva-tives of TK743 were used as host strains.

Media

LB broth (Miller 1992) was used as the growthmedium. Ampicillin, chloramphenicol, kanamy-cin, and tetracycline were used at 100, 30, 30,and 15 lg/ml, respectively.

Biotechnology Letters (2005) 27: 1267–1271 � Springer 2005DOI 10.1007/s10529-005-0216-4

Genetic techniques

Standard genetic techniques were used essentiallyas described by Sambrook & Russell (2001). Thelac-reporter gene (lacZ) was amplified by high-fidelity PCR using KOD-Plus polymerase (Toy-obo). Site-directed mutagenesis was carried outby the method of Kunkel et al. (1987). The entirefragment to be used for manipulation was se-quenced to ensure that no base change otherthan those planned had occurred. The lysogeni-zation of phage kDE3 was performed accordingto the protocol supplied by Novagen.

b-Galactosidase assay

An overnight culture was added to fresh LB mediumat 1%(v/v). The culture was incubated at 37 �C withshaking at 120 rpm. Samples were withdrawn at spe-cific time points and subjected to a b-galactosidaseassay according to the method of Miller (1992). Foreach strain, assays were performed for separate cul-tures of five independently isolated transformants.

SDS-PAGE analysis

The cells grown in LB medium were harvested,suspended in 10 mM sodium phosphate buffer(pH 7.2), disrupted by sonication, and then cen-trifuged at 10,000 g for 10 min to separate thesoluble fraction from the insoluble fraction. Tenmicrograms of protein from the soluble fractionwere loaded on a 12% SDS-PAGE. The concen-tration of protein was determined by the Lowrymethod. The insoluble fraction was washed oncewith 10 mM sodium phosphate buffer (pH 7.2),suspended and boiled in cracking buffer (60 mM

Tris/HCl (pH 6.8), 140 mM 2-mercaptoethanol,35 mM SDS, 1 M glycerol, and 150 lM Brom-ophenol Blue), and then centrifuged at 10,000 gfor 10 min. The supernatant, equivalent to 0.7 lg(wet wt) of cells, was loaded on the same gel.After the electrophoresis, the gel was stainedwith Coomassie Brilliant Blue R-250.

Results and discussion

Construction of the vector

First, we made a suitable expression vector withmultiple cloning sites as shown in Figure 1. Step 1:

An NdeI site was introduced at the initiationcodon of the tpl gene on pTK304 (Katayama et al.1999) by site-directed mutagenesis. Step 2: The

Fig. 1. Scheme of the vector’s construction. The numberingof pAH423 starts at the first T in the sequence GAATTC(EcoRI), and the unique restriction sites are indicated. Themultiple cloning sites are shown below the map.

1268

1.5-kb NdeI-PstI fragment containing the codingregion of tpl was replaced with a similarly digestedshort fragment that was made by annealing twocomplementary oligonucleotides containing multi-ple cloning sites, 5¢-CCATATGGATCCGCGGC-CGCCATGGCTGCAGG-3¢ and 5¢-CCTGCAGCCATGGCGGCCGCGGATCCATATGG -3¢.Step 3: The resulting plasmid was digested withSphI, blunt-ended, and ligated with the 0.7-kbXmnI fragment containing a strong transcrip-tion terminator (rrnBT1T2) excised frompTrc99A (Amersham). Step 4: Next, the regioncontaining the TyrR binding sites, the tpl pro-moter, the ribosome-binding site, the multiplecloning sites, and the transcription terminatorin this order was excised by digestion withSmaI and HindIII (blunt-ended), and insertedinto a blunt-ended NdeI site of pBR322. Step5: Finally, the tetracycline resistance gene wasdeleted by removing the 0.8-kb EcoRV-NruIfragment to generate pAH423.

Evaluation of the tpl-expression system usingthe lac-reporter gene

The efficiency of the tpl-expression system wasevaluated using the lac-reporter assay. Theb-galactosidase (lacZ) gene was inserted into theNdeI-NotI sites of pAH423, and the resulting re-porter plasmid (pAH444) was introduced intoTK743 [F) ara D(lac-pro) thi DtyrR::cat+] carry-ing either the wild-type tyrR gene or the mutanttyrR5 gene on a pSC101-derived plasmid. Thestrain was grown in LB medium and then sub-jected to the b-galactosidase assay.

As shown in Figure 2, the strain carrying thetyrR5 allele had significantly increased reporteractivity as compared to the strain carrying thewild-type allele. The level of expression was fur-ther elevated on introduction of the tyrosinetransporter tutB gene (Katayama et al. 2002). Inthe presence of a pACYC-derived plasmid carry-ing tutB, approximately 1.3-fold more activitywas obtained, due to the increased intracellularconcentration of L-tyrosine (an inducer of Tpl).However, no significant effect was observed onthe expression level even when 0.1% L-tyrosinewas added to the medium, maybe because LBbroth originally contains a considerable amountof L-tyrosine (data not shown). Thus, the level of

expression is maximized when the strain carryingtyrR5 and tutB is used as a host and incubatedin LB medium for 32 h.

The highest value obtained for the tpl-expres-sion system was compared to that obtained forthe tac- and T7-expression systems. The lac-reporter gene was inserted into the NdeI-PstI(blunt-ended) sites of pMAL-c2E (for the tac sys-tem; New England BioLabs) to generatepAH349, and the NdeI-BamHI (blunt-ended)sites of pET-3a (for the T7 system; Novagen) togenerate pAH471. The resulting reporter plas-mids pAH349 (tac system) and pAH471 (T7 sys-tem) were introduced into strain AH359 [F’traD36 lacIq D (lacZ)M15 proA+B+/ara D(lac-pro) thi DtyrR::cat+] and kDE3-lysogenizedAH359, respectively. The reporter plasmid forthe tpl system, pAH444, was introduced intostrain AH359 carrying the tyrR5 and tutB geneson separate plasmids. These strains were grownin LB medium and, as for the tac and T7

Fig. 2. Evaluation of the tpl-expression system using the lac-reporter assay. The plasmid pAH423 carrying the lac-reportergene under the control of the tpl promoter (pAH444) wasintroduced into strain TK743 [F) ara D(lac-pro) thiDtyrR::cat+] carrying either the wild-type tyrR gene (pSC101replicon tet+tyrRE. herbicola

+) ()) or the tyrR5 gene (pSC101replicon tet+ tyrRE. herbicola

V67A Y72C E201G) (s) on the plas-mid, or strain TK743 carrying an empty vector (pSC101 repli-con tet+) (h). The strain carrying the tyrR5 allele was furthertransformed with a plasmid carrying tutB (p15A repliconkan+ tutB+) (n). These strains were grown in LB medium.Samples were withdrawn at the indicated time, and subjectedto a b-galactosidase assay. The values are averages of at leasttwo independent experiments.

1269

systems, 0.5 mM IPTG was added at the mid-growth phase to induce expression. Samples werewithdrawn at different times, and subjected to ab-galactosidase assay. The inhibitory effect ofIPTG (the substrate analogue) on b-galactosidaseactivity was negligible (3.6% of total Miller U).

As a result, the highest expression level ob-tained for the tpl system (20,000 ± 3,000 MillerU) was found to be 2.5-fold higher than that forthe tac system (7,900 ± 800 Miller U, 2 h afterinduction). This was also the case when a differ-ent E. coli strain, JM101, was used as a host(21,000 ± 2,000 Miller U for the tpl system vs.8,600 ± 500 Miller U for the tac system).Although JM101 possesses the wild-type tyrR al-lele on its chromosome, it did not affect the ac-tion of TyrRV67A Y72C E201G on the tpl promoter,indicating that the tyrR) genotype is not neces-sary for the host genetic background.

Compared to the T7 system, the tpl systemexhibited a comparatively low level of efficiency ofexpression. The highest value obtained for the T7system using AH359 (kDE3) as a host was33,000 ± 2,000 Miller U, about 1.7-fold higherthan that obtained for the tpl system. WhenHMS174 (kDE3) (Novagen) was used as a hostcell, a maximum of 28,000 ± 4,000 Miller U wasobtained. As for strain HMS174 (kDE3) harboringthe reporter plasmid pAH471, the net reporteractivity was calculated by subtracting the b-galac-tosidase activity of the host (Lac+) (9,000 ± 1,000Miller U) from the total activity of the straincarrying pAH471 (37,000 ± 4,000 Miller U).

Next, the solubility and functionality of thesynthesized LacZ protein were evaluated by SDS-PAGE analysis. The cells harvested when the re-porter activity was at a maximum were disruptedby sonication, and then centrifuged at 10,000 g for10 min to separate the soluble fraction from theremaining insoluble fraction. Each fraction wassubjected to electrophoresis, and the gel wasstained with Coomassie Brilliant Blue R-250. TheLacZ protein was detected with a molecular size ofapproximately 120 kDa as shown in Figure 3(indicated by an arrow). Consequently, it was re-vealed that the differences among the three systemsas to the amounts of soluble forms of LacZ wellreflected the results of the reporter assay. The tplsystem (lanes 4 and 5) produced a soluble b-galac-tosidase at two- to three-fold higher level of the tacsystem (lanes 2 and 3), but at about a two-fold

lower level than the T7 system (lane 6 and 7). Theresults suggested that all soluble protein moleculesexpressed by the tpl system should be functional.

Apparently all LacZ molecules were expressedin a soluble form when its gene was placed underthe control of the tpl promoter (lane 10), while asignificant amount of LacZ was synthesized in aninsoluble form when the T7 system was em-ployed (lane 11).

The results presented here indicate that wesucceeded in constructing an effective proteinexpression system by using the tpl promoter, themutant TyrR protein, and the tyrosine trans-porter TutB. One important advantage of thissystem is that, essentially, any strain is availableas the host, in contrast to the T7 system whichrequires a certain genetic background for theinfection and/or lysogenization of phage kDE3when one wants to use a specific E. coli strain as

Fig. 3. Comparison of the amounts of soluble and insolubleforms of the LacZ protein expressed by the tac-, tpl-, and T7-expression systems. The cells grown in LB medium were har-vested, suspended in 10 mM sodium phosphate buffer (pH7.2), disrupted by sonication, and then centrifuged. The solu-ble fractions (equivalent to 10 lg of protein; lanes 2–7) andinsoluble fractions (equivalent to 0.7 lg of wet weight of cells;lanes 9–11) were loaded on a 12% SDS-polyacrylamide gel,and the gel was stained with Coomassie Brilliant Blue R-250.b-Galactosidase is indicated by an arrow (approximately120 kDa). Lanes 1 and 8, Prestained Protein Marker, BroadRange (New England BioLabs); lanes 2 and 3, the tac systemusing AH359 and JM101 as hosts, respectively; lanes 4 and 5,the tpl system using AH359 and JM101 as hosts, respectively;lanes 6 and 7, the T7 system using AH359 (kDE3) andHMS174 (kDE3) as hosts, respectively; lanes 9–11, the tac-,tpl-, and T7-expression systems using AH359 (for tac and tpl)and AH359 (kDE3) (for T7) as hosts.

1270

an expression host. Thus, although the tpl-expression system is not applicable to proteinstoxic to E. coli, it could serve as a good alterna-tive to commonly used protein expressionsystems.

Acknowledgements

This work was partly supported by a Grant-in-Aid for Scientific Research (B), No. 14360056,from the Ministry of Education, Culture, Sports,Science, and Technology, Japan. T. Koyanagiwas supported by the 21st Century COE Pro-gram of the Ministry of Education, Culture,Sports, Science and Technology.

References

Antson AA, Demidkina TV, Gollnick P, Dauter Z, von TerschRL, Long J, Berezhnoy SN, Phillips RS, Harutyunyan EH,Wilson KS (1993) Three-dimensional structure of tyrosinephenol-lyase. Biochemistry 32: 4195–4206.

Bai Q, Somerville RL (1998) Integration host factor and cyclicAMP receptor protein are required for TyrR-mediatedactivation of tpl in Citrobacter freundii. J. Bacteriol. 180:6173–6186.

Campbell JL, Richardson CC, Studier FW (1978) Geneticrecombination and complementation between bacteriophageT7 and cloned fragments of T7 DNA. Proc. Natl. Acad. Sci.USA 75: 2276–2280.

Katayama T, Suzuki H, Koyanagi T, Kumagai H (2000)Cloning and random mutagenesis of the Erwinia herbicola

tyrR gene for high-level expression of tyrosine phenol-lyase.Appl. Environ. Microbiol. 66: 4764–4771.

Katayama T, Suzuki H, Koyanagi T, Kumagai H (2002)Functional analysis of the Erwinia herbicola tutB gene and itsproduct. J. Bacteriol. 184: 3135–3141.

Katayama T, Suzuki H, Yamamoto K, Kumagai H (1999)Transcriptional regulation of tyrosine phenol-lyase genemediated through TyrR and cAMP receptor protein. Biosci.Biotechnol. Biochem. 63: 1823–1827.

Koyanagi T, Katayama T, Suzuki H, Nakazawa H, YokozekiK, Kumagai H (2005) Effective production of 3,4-dihydr-oxyphenyl-LL-alanine (LL-DOPA) with Erwinia herbicola cellscarrying a mutant transcriptional regulator TyrR. J. Biotech-nol. 115: 303–306.

Kumagai H, Yamada H, Matsui H, Ohkishi H, Ogata K (1970)Tyrosine phenol lyase. I. Purification, crystallization, andproperties. J. Biol. Chem. 245: 1767–1772.

Kunkel TA, Roberts JD, Zakour RA (1987) Rapid and efficientsite-specific mutagenesis without phenotypic selection.Meth-ods in Enzymology 154: 367–382.

Miller JH (1992) A Short Course in Bacterial Genetics. ALaboratory Manual and Handbook for Escherichia coli andRelated Bacteria. Cold Spring, Harbor, NY:Cold SpringHarbor Laboratory.

Pittard J, Camakaris H, Yang J (2005) The TyrR regulon. Mol.Microbiol. 55: 16–26.

Sambrook J, Russell DW (2001) Moleculer Cloning: A Labo-ratory Manual, 3rd edn. Cold Spring Harbor, NY:ColdSpring Harbor Laboratory.

Suzuki H, Katayama T, Yamamoto K, Kumagai H (1995)Transcriptional regulation of tyrosine phenol-lyase gene ofErwinia herbicola AJ2985. Biosci. Biotechnol. Biochem. 59:2339–2341.

Suzuki H, Nishihara K, Usui N, Matsui H, Kumagai H (1993)Cloning and nucleotide sequence of Erwinia herbicolaAJ2982 tyrosine phenol-lyase gene. J. Ferment. Bioeng. 75:145–148.

Yanisch-Perron C, Vieira J, Messing J (1985) Improved M13phage cloning vectors and host strains: nucleotide sequencesof the M13mp18 and pUC19 vectors. Gene 33: 103–119.

1271

Purification and properties of an N-acetylglucosaminidasefrom Streptomyces cerradoensis

Iderval da Silva Junior Sobrinho1, Luiz Artur Mendes Bataus2, Valeria Ribeiro Maitan2

& Cirano Jose Ulhoa1,*1Laboratorio de Enzimologia, Universidade Federal de Goias, 74001-970, Goiania, GO, Brazil2Laboratorio de Bioquımica e Engenharia Genetica, Universidade Federal de Goias, 74001-970, Goiania, GO,Brazil*Author for correspondence (E-mail: ulhoa@icb.ufg.br)

Received 22 March 2005; Revisions requested 19 April 2005; Revisions received 14 June 2005; Accepted 15 June 2005

Key words: N-acetylglucosaminidase, characterization, production, Streptomyces cerradoensis

Abstract

An N-acetylglucosaminidase produced by Streptomyces cerradoensis was partially purified giving, by SDS-PAGE analysis, two main protein bands with Mr of 58.9 and 56.4 kDa. The Km and Vmax values for theenzyme using p-nitrophenyl-b-N-acetylglucosaminide as substrate were of 0.13 mM and 1.95 U mg)1

protein, respectively. The enzyme was optimally activity at pH 5.5 and at 50 �C when assayed over 10 min.Enzyme activity was strongly inhibited by Cu2+ and Hg2+ at 10 mM, and was specific to substratescontaining acetamide groups such as p-nitrophenyl-b-N-acetylglucosaminide and p-nitrophenyl-b-D-N,N¢-diacetylchitobiose.

Introduction

Chitin, a b)1, 4-linked polymer of N-acetylglu-cosamine, is a structural component of the arthro-pod exoskeleton and is a common constituent offungal cell walls. The complete degradation ofchitin is performed by the chitinolytic systemcomposed of chitinases (EC 3.2.1.14) and N-acet-ylglucosaminidases (E.C. 3.2.1.30) (Sahai &Manocha 1993). Chitinases can hydrolase the sub-strate by two possible mechanisms, identified bythe products of hydrolysis: (a) endochitinasescleave internal bonds within chitin releasing chito-tetraose, chitotriose and chitobiose; (b) exochitin-ases catalyses the release of chitobiose without theformation of oligo or monosaccharides. TheN-acetylglucosaminidases cleave chitobiose, chito-triose and chitotetraose releasing N-acetylglucos-amine.

Among bacteria, the actinomycetes are animportant chitinase-producing group, especially

those belonging to the genus Streptomyces(Broadway et al. 1995, El Sayed et al. 2000,Gomes et al. 2000, Ueno & Miyashita 2000).There is less work about N-acetylglucosaminidasefrom other organisms and few researches havebeen done to characterize this type of enzymefrom Streptomyces. In this article we describe thepartial purification and characterization of a N-acetylglucosaminidase produced by Streptomycescerradoensis, isolated from Brazilian cerradosoils.

Materials and methods

Microorganism and enzyme production

Streptomyces cerradoensis obtained from solidculture media (ISP-2) were inoculated in Erlen-meyer flasks containing 50 ml YEME liquidmedia (Hopwood et al. 1985).

Biotechnology Letters (2005) 27: 1273–1276 � Springer 2005DOI 10.1007/s10529-005-0218-2

Samples of 50 ll of cells suspension fromYEME media were added to Erlenmeyer flasks(250 ml) containing 50 ml media: 0.2% (w/v)K2HPO4, 0.3% (w/v) NaCl, 0.3% (w/v) KNO3,0.05% (w/v) MgSO4 Æ 7H2O, 0.04% (w/v) CaCl2 Æ2H2O, 0.002% (w/v) FeSO4, 0.001% (w/v)MnSO4, 1% (w/v) chitin, pH 7.0. The flasks wereincubated at 30 �C with shaking (150 rpm) for9 days.

N-Acetylglucosaminidase assay (NAGase)

Enzyme activity was assayed using p-nitrophe-nyl-b-N-acetylglucosaminide (pNP-GlcNAc) assubstrate. The reaction mixture consisted of50 ll enzyme solution, 350 ll 50 mM sodiumacetate buffer (pH 5.5) and 100 ll 5 mM pNP-GlcNAc as described by Ulhoa and Peberdy(1993). One unit (U) of enzyme was defined asthe amount of enzyme that released 1 lM

p-nitrophenol per min.

Partial enzyme purification

Samples of the supernatant (150 ml) were appliedto a SP-Sepharose chromatography column(2.2� 15 cm), equilibrated with 50 mM sodiumacetate buffer (pH 5.5), at 1 ml min)1. The col-umn was washed with the same buffer and theproteins eluted with a linear gradient of 0–1 M

NaCl. Fractions containing NAGase activitywere pooled, dialyzed against 50 mM sodiumacetate buffer (pH 5.0), and applied to methyl-Sepharose column (1� 5 cm). The adsorbedproteins were eluted at 4 �C with a decreasinglinear gradient of (NH4)2SO4 (1.2 M to 0). Frac-tions containing NAGase activity were pooledand stored at )10 �C. SDS-PAGE with 10 %(w/v) polyacrylamide was by the standard meth-od of by Laemmli. The proteins were silverstained by the method of Blum et al. (1987).

Thin-layer chromatography

The end products of the enzymatic hydrolysis ofN,N¢-diacetylchitobiose (5 mM), N,N¢,N¢¢-tria-cetylchitotriose (5 mM) and colloidal chitin(0.2%, w/v) were analyzed by TLC according toChung et al. (1995).

Results and discussion

Production of N-acetylglucosaminidase

Recently, we isolated a chitin-degrading actino-mycete from soil sample, which was identified asbelonging to Streptomyces genus. Sequence anal-ysis of the gene encoding 16S rDNA (GeneBankAccession No. AY627277) showed that this acti-nomycete is a new species, named by us as Strep-tomyces cerradoensis. This strain showed inpreliminary tests, high chitinase and N-acetylglu-cosaminidase (NAGase) activity when grown inchitin-containing medium. A main peaks ofNAGase activity (0.061 U) was found after 120 hgrowth at 30 �C.

Partial purification of N-acetylglucosaminidase

The enzyme was partially purified from a culturesupernatant by ion exchange chromatographyand gel filtration. The fractions obtained afterthe procedure were free of chitinase activity. Atthe final steps, two main bands of protein weredetected by SDS-PAGE with apparent molecularmass of 58.9 and 56.4 kDa (Figure 1). This

Fig. 1. Protein profile on SDS-PAGE at a concentration of10% polyacrylamide. Lane M indicates the molecular massmarkers and N shows the N-acetylglucosaminidase partially.Phosphorylase b (99.4 kDa), serum albumin (66.0 kDa), oval-bumin (45.0 kDa) and carbonic anhydrase (29.0 kDa) wereused as molecular mass standards.

1274

molecular weight was in a similar range to thatproduced by S. thermoviolaceus (Tsujibo et al.1998) although, smaller NAGases (27 and49.5 kDa) also have been isolated and character-ized from Streptomyces plicatus (Tarentino et al.1978, Trimble et al. 1982).

Enzyme characterization

The lower Km (0.13 mM) indicates that the en-zyme has high affinity for the substrate when it iscompared with those reported for S. thermovio-laceus (0.43 mM) and for S. hygroscopicus (NA1:0.12 mM and NA2: 0.76 mM) (Irhuma et al. 1991,Tsujibo et al. 1998). Most bacterial NAGases areoptimally active in the range of pH 4–5 and

between 50 and 60 �C (Trimble et al. 1982,Irhuma et al. 1991, Saito et al. 1998, Tsujiboet al. 1998). Concurring with this, the NAGaseof S. cerradoensis displayed maximal activity atpH 5 and at 50 �C (Table 1). The enzyme,though, rapidly lost activity either at 40 or 55 �C(Table 1). Activity was not affected by Ca2+,Mg2+ and Zn2+ but was inhibited by Cu2+ andHg2+ at 10 mM. Almost all of the NAGasesfrom Streptomyces are inhibited by Hg2+ indi-cating the importance of indole amino acid resi-dues in the enzyme function (Irhuma et al. 1991,Trimble et al. 1982, Tsujibo et al. 1998).

Substrate specificity

The activity of the enzyme was observed onlyover substrates containing acetamide groups(Table 2). Similar results were observed for NAGasefrom other Streptomyces species, suggesting thatacetamide groups are important to the recogni-tion of chitin and its oligomers by these enzymes(Tsujibo et al. 1998). N-Acetylglucosaminidewas released from N,N¢-diacetylchitobiose andN,N¢,N¢¢-triacetylchitotriose after 24 h incuba-tion, while this monomer was released fromcolloidal chitin only after 48 h incubation. Sucha profile indicates that this enzyme acts at theextremity of the substrates releasing N-acetylglu-cosaminide and accords with the definition ofNGAse as proposed by Sahai and Manocha(1993).

Acknowledgements

This work was supported by a biotechnology re-search grant to C.J. Ulhoa from CNPq andFUNAPE/UFG. I.S. Jr. Sobrinho was supportedby CAPES/MEC Brazil.

Table 2. Specificity of N-acetylglucosaminidase to different substrates.

Substrate (5 mM) Enzyme activity (U) Relative activity (%)

p-Nitrophenyl-b-N-acetylglucosaminide 27±4 100

p-Nitrophenyl-b-D-N,N’-diacetylchitobiose 6.5±0.04 25

p-Nitrophenyl-b-D-glucopyranoside 0.07±0.01 0.3

p-Nitrophenyl-b-D-galactopyranoside 0 0

The reaction mixture consisted of 50 ll of enzyme solution, 350 ll 50 mM sodium acetate buffer (pH 5.5) and 100 ll of specific p-nitrophenyl substrate (5 mM). One unit (U) of enzyme was defined as the amount of enzyme that released 1 lM p-nitrophenol per min.

Table 1. Biochemical properties of the NAGase from Strepto-myces cerradoensis.

pH Optimum 5.0

Temperature optimum (�C) 50

Temperature stability pH 5.5 over 30 min

40 �C 30%

55 �C 12%

Km (mM) 0.13

Vmax (U mg)1 protein) 1.95

Inhibition by Hg2+ (10 mM) 100%

Inhibition by Cu2+ (10 mM) 78%

The effect of pH on the enzyme activity was determined byvarying the pH of the reaction mixtures using sodium acetate(pH 3.5–5.5) and potassium phosphate buffer (pH 6.0–7.5). Theeffect of temperature on the enzymatic activity was determinedat the pH optimum, in the range of 30–75 �C. The effect oftemperature on the enzyme stability was analyzed by previouslyincubating the enzyme at 40 and 55 �C for 30 min. Michaelis–Menten constant (Km) was determined by non-linear-regressionanalysis of data obtained by measuring the rate of qNP-Glc-NAc hydrolysis (from 0.02 to 0.6 mM). The inhibition of theNAGase activity by metal ions was determined through previ-ous incubation of the enzyme samples with 10 mM ZnSO4, KCl,MgSO4, CaCl2, CuSO4 and HgCl2 for 2 min. Results are meansvalues of three replicates.

1275

References

Blum H, Beier H, Gross H (1987) Improved silver staining ofplants proteins, RNA and DNA in polyacrilamide gels.Eletrophoresis 8: 93–99.

Broadway RM, Williams DL, Kain WC, Harman GE, LoritoM, Labeda DP (1995) Partial characterization of chitinolyticenzymes from Streptomyces albidoflavus. Lett. Appl.Microbiol. 20: 271–276.

Chung YC, Kobayashi T, Kanai H, Akiba T, Kudo T (1995)Purification and properties of extracellular amylase from thehyperthermophilic archeon Thermococcus profundus DT5432. Appl. Environ. Microbiol. 1: 1502–1506.

El Sayed EISA, Ezzat SM, Ghaly MF, Mansour M, El BoheyMA (2000) Purification and characterization of two chitin-ases from Streptomyces albovinaceus S-22. World J. Micro-biol. Biotechnol. 16: 87–89.

Gomes RC, Semedo LTAS, Soares RMA, Alviano CS,Linhares LF, Coelho RRR (2000) Chitinolytic activity ofactinomycetes from a cerrado soil and their potential inbiocontrol. Lett. Appl. Microbiol. 30: 146–150.

Hopwood DA, Bibb MJ, Chater KF, Kieser T, Bruton CJ,Kieser HM, Lydiate DJ, Smith CP, Ward JM, Schrempf H(1985) Genetic Manipulation of Streptomyces: A LaboratoryManual, New York:The John Innes Foundation.

Irhuma A, Gallagher J, Hackett TJ, McHale AP (1991)Studies on N-acetylglucosaminidase activity produced by

Streptomyces hygroscopicus. Biochim. Biophys. Acta 1074:1–5.

Sahai AS, Manocha MS (1993) Chitinases of fungi and plants:their involvement in morphogenesis and host–parasite inter-action. FEMS Microbiol. Rev. 4: 317–338.

Saito A, Fujii T, Yoneyama T, Miyashita K (1998) glk isinvolved in glucose repression of chitinase production inStreptomyces lividans. J. Bacteriol. 180: 2911–2914.

Tarentino AL, Trimble RB, Maley F (1978) Endo-b-N-acetyl-glucosaminidase from Streptomyces plicatus. Methods inEnzymology 1: 574–580.

Trimble RB, Tarentino AL, Aumick GE, Maley F (1982) Endo-b-N-acetylglucosaminidase L from Streptomyces plicatus.Methods in Enzymology 83: 603–610.

Tsujibo H, Hatano N, Mikami T, Hirasawa A, Miyamoto K,Inamori Y (1998) A novel b-N-acetylglucosaminidase fromStreptomyces thermoviolaceus OPC-520: gene cloning,expression, and assignment to family 3 of the glycosylhydrolases. Appl. Environ. Microbiol. 64: 2920–2924.

Ulhoa CJ, Peberdy JF (1993) Effect of carbon sources onchitobiose production by Trichoderma harzianum. MycologyRes. 97: 45–48.

Ueno H, Miyashita K (2000) Inductive production of chitin-olytic enzymes in soil microcosms using chitin, other carbon-sources, and chitinase-producing Streptomyces. Soil Sci.Plant Nut. 46: 863–871.

1276

A modified PCR system for amplifying b-ketoacyl-ACP synthase genefragments with DNA from Streptomyces luteogriseus

Feng-Ming Yu, Xin Jiang, Jin-Chuan Wu & Ying-Jin Yuan*Department of Pharmaceutical Engineering, Tianjin University, P.O. Box 6888, 300072, Tianjin, P. R. China*Author for correspondence (Fax: 86-22-27403888; E-mail: yjyuan@public.tpt.tj.cn, yjyuan@tju.edu.cn)

Received 27 April 2005; Revisions requested 18 May 2005; Revisions received 15 June 2005; Accepted 15 June 2005

Key words: b-ketoacyl-ACP synthase, gene amplification, modified PCR system, Streptomycesluteogriseus

Abstract

Streptomyces luteogriseus strain 099, producing a new type of macrolide antibiotic with anti-coxB6 virusand anti-HIV protease activities, was isolated from soil. PCR was optimized to amplify b-ketoacyl-ACPsynthase (KS) genes. The system was optimized around the use of higher concentrations of DMSO (15%vs. 10% v/v) and dNTP (500 lM vs. 50–200 lM) and a lower annealing temperature (55 �C vs. 60–70 �C)than the normal PCR method used to amplify high GC content DNA.

Introduction

Actinomycetes, especially streptomyces, are ma-jor producers of antibiotics and other secondarymetabolites and are industrially important. Theresearch on the functional genome is becomingdeeper and wider. The total sequence of Strepto-myces coelicolor and S. averimitilis genomes andpartial gene sequence of some other importantstrains producing antibiotics have been reported.This has resulted in a call for genetic reconstruc-tion of strains for producing industrially usefulmetabolites and the synthesis of new type ofantibiotics using combinatorial biosynthesis(Rodrigue et al. 2003, Weber et al. 2003). Thestudy of relative genes for antibiotic synthesis istherefore essential.

Genes of streptomyces are usually obtainedby a shot-gun method which is time-consuming.According to the homology among genes, frag-ments could be obtained by PCR (Izumikawaet al. 2003). For GC-rich template DNA fromstreptomyces PCR is often frustrated by inade-quate yield of the target DNA sequence and

undesired non-specific bonds (Chakrabarti &Schutt 2001). So normal PCR is not suitable foramplifying gene fragments of streptomyces anddifferent corrective actions have different resultsunder different conditions. It is essential that thePCR system and reaction conditions for strepto-myces be optimized.

Streptomyces luteogriseus 099 was isolatedfrom soil by our group. It can produce a sec-ondary metabolite having good anti-coxB6 virusand anti-HIV protease properties with the IC50

of 230 lg/ml and 40 lg/ml, respectively, show-ing a broad application spectrum (Wang 2004).By means of precursors and specific inhibitorsadded during experiments, together with NMR,UV and IR spectrum analyses, the screenedsubstance was speculated to be a macrolideantibiotic synthesized through the polyke-tide pathway. However, we found that a KS(b-ketoacyl-ACP synthase) fragment of its gen-ome could not be amplified perfectly by normalPCR. We have therefore attempted the to opti-mize the PCR system and reaction conditionsand report the results below.

Biotechnology Letters (2005) 27: 1277–1282 � Springer 2005DOI 10.1007/s10529-005-3219-2

Materials and methods

Strains and culture medium

Streptomyces luteogriseus 099 was isolated andpreserved in our laboratory. E. coli DH5a wasobtained from Prof. Lai-Jun Xing of NankaiUniversity, China. The medium for S. luteogri-seus contained 5 g glucose, 30 g soluble starch,4 g peptone, 1.5 g K2HPO4, 0.25 g NaCl, 0.5 gMgSO4 in 1 liter water, pH 7.0. Luria–Bertani(LB) medium was used for showing E.coli.

Reagents and materials

A PCR reagent kit, restriction enzyme EcoRI,HindIII, DNA Marker DL2000 and pMD-18Tvector were obtained from Takara BiotechnologyCorporation. The Ultra-sep Gel Extraction kitwas obtained from Omega Bio-Tek, USA. TheHybaid PCR express thermal cycler was obtainedfrom Hybaid, UK. The GDS-8000 System andUVP Bioimaging Systems were obtained fromUVP, Inc., USA. Other chemicals were ofreagent grade and obtained commercially.

Genome DNA extraction

A modified method of Hopwood et al. (1985)was used. After culturing S. luteogriseus for 24 h,the mycelium was collected, washed with 150 mM

NaCl/100 mM EDTA and treated with lysozymeat 37 �C. After addition of SDS followed byheating to 65 �C, the viscosity of the myceliumsolution declined. Equal volumes of phenol/chlo-roform/isoamyl alcohol (25:24:1, by vol.) wereadded to remove protein. After extracting twicewith chloroform, 0.9 vol. cold 2-propanol wasadded into the aqueous phase to precipitateDNA. DNA was washed twice with 70% (v/v)ethanol, dried at room temperature and then dis-solved in TE buffer (pH 8.0). RNA was degradedin the same manner by RNase. The purity andcontent of DNA were measured by absorbancies.

Primer designing

Forward primer (5¢-GCTCGTACTCGTCGCTCCCGGCCAG-3¢) and reversed primer (5¢-CGA-CATGGTCGCGACGGTCTCCTCG-3¢) were

designed according to the homology of KS genesequences in S. avermitilis, S. coelicolor and otherstrains published in the GeneBank. Primers wereprovided by Shanghai Sangon Biological Engi-neering Technology and Services Co., Ltd, China.

PCR system

PCR was carried out under the following condi-tions. The original reaction system was deter-mined by the preliminary experiment with thePCR solution containing 5 ll 10 � buffer,2.5 mM MgCl2, 500 lM dNTP, 500 ng DNA tem-plate, 1 lM each primer, certain organic additivesat different concentrations and 0.5 ll Taq poly-merase (5 U/ll) with the volume made to 50 llwith sterilize water. The reaction was conductedat 95 �C for 10 min, 94 �C for 30 s, 55 �C for30 s, 72 �C for 1 min (40 cycles) and 72 �C for10 min, respectively.

Electrophoresis

Electrophoresis of amplification products wasdone on 1% agarose gels in which 5 ll reactionproducts were loaded with 0.5 ll 10 � loadingbuffer. Gels were run at 80 V for 40 min, stainedwith ethidium bromide, visualized on a UVtransilluminator and documented by UVP Bioi-maging System.

Amplification product purified from gel

The Ultra-Sep Gel Extraction Kit was used topurify the amplification product according tomanufacturer’s recommendation.

Techniques for DNA manipulation

DNA manipulation techniques were carried outaccording to standard protocols and instructionkits, which include linkage of the amplificationproduct with the vector, transformation into E.coli, screening of the positive clone and plasmidpreparation, etc.

DNA sequencing

Amplification product was sequenced by ShanghaiSangon Biological Engineering Technology andServices Co., Ltd, China.

1278

Results and discussion

PCR enhancement by additives

GC content in the Streptomyces genome is over70%. Higher energy is needed to melt DNA

chains otherwise primers cannot be coupled withthe template DNA for two chains combined to-gether (Chenchik et al. 1996) and there will beno amplification product. To lower the annealingtemperature to make it suitable for the annealingbetween primers and template and favorable for

Fig. 1. Effect of additive types on amplification. (a), DMSO; (b), methanamide; (c), glycerol (Lane M, DL2000 DNA marker; lane1, PCR with no additive; lanes 2–6, PCR with additives at 1, 5, 10, 15 and 20% (v/v), respectively); (d), PCR at different pre-dena-turation times (Lane M, DL2000 DNA marker; lane 1, 10 min; lane 2, 5 min).

1279

DNA polymerase to go through the secondarystructure region, as well as to avoid the non-specific amplification and the formation of pri-mer dimer at the same time, some measures havebeen taken, such as improving the denaturingtemperature, adding organic solvents includingformamide (Sarkar et al. 1990) dimethylsulfoxide(DMSO) (Baskaran et al. 1996), glycerol (Pompand Medrano 1991) and betaine (Hengen 1997).However, these methods are not always effectivefor reasons that are still unclear. In our experi-ments, the pre-denaturation time was prolongedto 10 min and the annealing temperature was in-creased from 55 to 70 �C but there was no targetfragment amplified.

When glycerol and formamide were addedinto the reaction system, no target fragment wasamplified (Figure 1a–c). Formamide slightly sup-pressed the dimer formation and non-specificamplification. Glycerol only inhibited non-spe-cific amplification. Amplification was evidentwith DMSO at 5% (v/v) and was maximal at15% (v/v). Above 15% or under 5% (v/v),amplification was not observed. It is characteris-tic that the amount of the added DMSO washigher in our system than in the normal reactionsystem, and 15% (v/v) DMSO was used in thesubsequent experiments. Although 10% (v/v)DMSO is the commonly used concentration foramplifying high GC content DNA (Pomp andMedrano 1991, Carmody et al. 2004) it ischangeable and adjustable according to the prac-tical conditions.

When the pre-denaturation time was short-ened to 5 min, the amplified product was detectedand its quantity changed little (Figure 1d). There-fore the pre-denaturation time was set at 5 min.

Effect of primer concentration on PCR

The optimal primer concentration is between 0.1and 1.0 lM. If the primer concentration is toohigh, it would lead to a mismatch between pri-mer and template, non-specific amplification andan increased chance of dimer formation. In ourexperiments, a better yield and specificity ofamplified product were detected with each primerat 1 lM, but there was dimer formation (Fig-ure 2). When a pair of primers was used at0.2 lM, no dimer formation was observed but theamount of amplified product became less. The

suitable primer concentration was between 0.2and 1 lM so 1 lM was used in the subsequentexperiments.

Effect of dNTP amount on PCR

In the amplification reaction, increased precisionof the reaction was obtained with a lower con-centration of dNTP. When its concentration wasbelow 400 lM, the target amplification productwas not detected but reached its maximum at500 lM (Figure 3). A higher concentration ofdNTP was unfavorable for PCR due to its com-bination with Mg2+ resulting in the shortage offree Mg2+. In the normal PCR system, thedNTP concentration ranged from 50 to 200 lM

but, in our system, the dNTP concentration wasas high as 500 lM, dramatically exceeding thatconventionally used.

Influence of Mg2+ on amplification

Mg2+ can alter the amount of amplificationproduct by changing DNA polymerase activityand affect the reaction specificity by virtue ofchanging the primer annealing temperature. Theoptimal concentration of Mg2+ changes when

Fig. 2. Effect of primer concentration on amplification reac-tion. Lane M, DL2000 DNA marker; lanes 1–5, PCR withprimers at 0.2, 1, 2, 3 and 4 lM, respectively.

1280

different primers, templates and differentamounts of dNTP are used. A high concentra-tion of Mg2+ would increase the output andnon-specificity and decrease the fidelity. But theyield may be reduced at low concentrations. Inour experiments, the best result was obtained at2 mM Mg2+ (Figure 4).

Effect of template DNA concentration on PCR

In order to reduce the probability of mistakes pro-duced by Taq polymerase and achieve a higherquantity of amplified product, a higher concentra-tion of DNA was used. If DNA content is toohigh, it will increase contamination and reduce theamplification efficiency. With the template con-centration increased, the amount of amplificationproduct increased and peaked at 400 ng (Fig-ure 5). So 400 ng was selected as the optimizedDNA concentration in the subsequent tests.

Influence of cycles on amplification

Figure 6 shows that amplification was evidentwhen the cycles reached 25 and the yield wasmaximal at 35 cycles. In order to save energy,

reduce the possibility of mutant and ensure plen-tiful amplification product, 30–35 cycles are rec-ommended in this PCR system.

Fig. 3. Effect of dNTP concentration on amplification. LaneM, DL2000 DNA marker; lanes 1–5, PCR with dNTP at 100,200, 300, 400 and 500 lM, respectively.

Fig. 4. Effect of Mg2+ on amplification. Lane M, DL2000DNA marker; lanes 1–6, PCR with Mg2+ at 0.5, 1, 1.5, 2, 2.5and 3 mM, respectively.

Fig. 5. Effect of DNA amount on amplification reaction.Lane M, DL2000 DNA marker, lanes 1–5: PCR with DNAat 100, 200, 300, 400 and 500 ng, respectively.

1281

In addition to the conditions discussed above,the annealing temperature also influenced PCR.The annealing temperature must be low enoughto ensure efficient annealing and high enough toreduce non-specific combination between primerand template. The reasonable annealing temper-ature should be between 55 and 70 �C in thenormal PCR and should be higher in GC-richtemplate DNA system. In this study, there wasno necessity for PCR to be carried out at ahigher annealing temperature because no specificbonds were amplified at lower annealing temper-ature when DMSO was added.

In summary: the reaction system and condi-tions for amplification of this fragment in S.luteogriseus were set up as follows: the PCR solu-tion contained 5 ll 10 � buffer, 2 mM MgCl2,500 lM dNTP, 400 ng DNA template, each pri-mer 1 lM, 2.5 U Taq polymerase and 7.5 llDMSO, with the volume brought to 50 ll withsterilize water. The reaction was conducted at95 �C for 5 min, 94 �C for 30 s, 55 �C for 30 s,72 �C for 1 min (35 cycles) and 72 �C for 10 min.This protocol is unique in terms of the higher

concentrations of DMSO and dNTP and lowerannealing temperature.

Acknowledgement

We wish to thank the ‘‘863’’ Hi-Tech Researchand Development Program of China (Grant No.2001AA214081) and the National Natural Sci-ence Foundation of China (Grant No. 20425620)for financial support.

References

Baskaran N, Kandpal RP, Bhargava AK, Glynn MW, Bale A,Weissman SM (1996) Uniform amplification of a mixture ofdeoxyribonucleic acids with varying GC content. GenomeRes. 6: 633–638.

Carmody M, Byrne B, Murphy B, Breen C, Lynch S, Flood E,Finnan S, Caffrey P (2004) Analysis and manipulation ofamphotericin biosynthetic genes by means of modified phageKC515 transduction techniques. Gene 343: 107–155.

Chakrabarti R, Schutt CE (2001) The enhancement of PCRamplification by low molecular-weight sulfones. Gene 274:293–298.

Chenchik A, Diachenko L, Moqadam F, Tarabykin V,Lukyanov S, Siebert PD (1996) Full-length cDNA cloningand determination of mRNA 5¢ and 3¢ ends by amplificationof adaptor-ligated cDNA. Biotechniques 21: 526–534.

Hengen PN (1997) Optimizing multiplex and LA-PCR withbetaine. Trends Biochem. Sci. 22: 225–226.

Hopwood DA, Bibb MJ, Chater KF, Kieser T, Bruton CJ,Kieser HM, Lydiate DJ, Smith CP, Ward JM, Shrempf H(1985) Genetic Manipulation of Streptomyces: A LaboratoryManual. Norwich: John Innes Foundation.

Izumikawa M, Murata M, Tachibana K, Ebizuka Y, Fujii I(2003) Cloning of modular type I polyketide synthase genesfrom salinomycin producing strain of Streptomyces albus.Bioorg. Med. Chem. 11: 3401–3405.

Pomp D, Medrano JF (1991) Organic solvents as facilitators ofpolymerase chain- reaction. Biotechniques 10: 58–59.

Rodrigue E, Hu Z, Ou S, Volchegursky Y, Hutchinson CR,McDaniel R (2003) Rapid engineering of polyketide over-production by gene transfer to industrially optimized strains.J. Ind. Microbiol. Biotechnol. 30: 480–488.

Sarkar G, Kapelner S, Sommer SS (1990) Formamide candramatically improve the specificity of PCR. Nucleic AcidsRes. 18: 7465.

Wang ZP, (2004) Study on isolation, identification, bioactivityand biosynthesis of Maituolaimycin inhibiting HIV PR andCVB6. PhD Thesis. Tianjin, China: Tianjin University.

Weber T, Welzel K, Pelzer S, Vente A, Wohlleben W (2003)Exploiting the genetic potential of polyketide producingstreptomycetes. J. Biotechnol. 106: 221–232.

Fig. 6. Effect of cycles on amplification. Lane M, DL2000DNA marker; lanes 1–4, PCR with 25, 30, 35 and 40 cycles,respectively.

1282

Molecular cloning and tissue distribution of SF-1-related orphan receptorsduring sexual maturation in female goldfish

Cheol Young Choi1,* & Hamid R. Habibi21Division of Marine Environment & Bioscience, Korea Maritime University, 606-791, Busan, Korea2Department of Biological Sciences, University of Calgary, T2N 1N4, Calgary, Alberta, Canada*Author for correspondence (Fax: +82-51-404-3988; E-mail: choic@bada.hhu.ac.kr)

Received 18 April 2005; Revisions requested 2 May 2005; Revisions received 15 June 2005; Accepted 16 June 2005

Key words: cDNA cloning, goldfish, RT-PCR, SF-1, tissue distribution

Abstract

The steroidogenic factor (SF)-1 gene is one of a number of orphan nuclear receptors, which is a keytranscriptional regulator in vertebrate reproduction. We have isolated the SF-1 homologue cDNA from thegoldfish pituitary and designed primers for SF-1 on the basis of the highly conserved regions of variousknown SF-1 superfamily genes. SF-1 cDNA contained 1,948 nucleotides including an open reading framepredicted to encode a protein of 503 amino acids. The distribution pattern of SF-1 in a variety of tissuesduring sexual maturation in female goldfish was also examined by RT-PCR. Significant variations in therelative expression of SF-1 were observed in different tissues in immature and mature female goldfish. SF-1transcript in pituitary was significantly higher than other tissues tested in immature and mature femalegoldfish. Lower expression of SF-1 was observed in the liver but was not detected in brain and ovary of theimmature female goldfish. Presence of SF-1 was the predominant expression in the pituitary and brain ofmature female goldfish. Also, in the mature female goldfish, a weak transcript was detected in liver andovary. Interestingly, RT-PCR analysis revealed that the expression of SF-1 became higher in the brain andweaker in the liver in maturing female goldfish. Thus, SF-1 may be regulated in goldfish brain and/or liver.Thus is also tissue-specific distribution of SF-1 during sexual maturation in female goldfish.

Introduction

Many transcriptional factors in the orphan nucle-ar receptor superfamily have recently beencloned. Fushi tarazu transcription factor-1 (FTZ-F1), a member of an orphan nuclear receptor isimportant transcriptional regulator of the fushitarazu home box gene in Drosophila (Yu et al.1997). After the isolation of Drosophila FTZ-F1cDNA, many FTZ-1 homologues such as bovine(Honda et al. 1993), rat (Galarneau et al. 1996),chicken (Kudo & Sutou 1997), frog (Kawanoet al. 1998) and human (Galarneau et al. 1998)were reported. Recent sequence alignments alsosupport the existence of the orphan nuclear recep-tors in teleost fish (Liu et al. 1997, Watanabe

et al. 1999, Chai & Chan 2000). Vertebrate FTZ-F1 homologues were mainly classified into twosubgroups by phylogenetic tree analysis (Nakaj-ima et al. 2000) and/or based on function, tissuedistribution (Galarneau et al. 1996, Hofsten et al.2001). One is a subgroup of SF-1/Ad4BP (Steroi-dogenic factor-1/adrenal 4-binding protein), andthe other is that of LRH/FTF (Liver receptorhomologue protein/a1-fetoprotein transcriptionfactor).

The SF-1 is expressed in the adrenal cortex andgonads (Honda et al. 1993) in which steroid hor-mones are synthesized. SF-1 is essential for devel-opment of the steroidogenic organs, a role thathas been extended to all levels of the hypotha-lamic-pituitary-gonadal axis (Ingraham et al.

Biotechnology Letters (2005) 27: 1283–1290 � Springer 2005DOI 10.1007/s10529-005-0220-8

1994, Manglesdorf et al. 1995). Moreover, theexpression is maintained in male embryos duringtestis differentiation, but declines in female em-bryos just after the onset of ovarian differentia-tion, although it increases in females in lateembryogenesis (Ikeda et al. 1994). In chicken, SF-1 expression during ovarian differentiation in-creases, which is opposite to the pattern seen inmammals, where it decreases in females (Smithet al. 1999).

Mammalian LRH receptors regulate a1-feto-protein expression and are located in endodermalcells in pancreas and liver (Galarneau et al.1996).

Nuclear receptors are composed of severalhomologous modular domains. The members ofthis superfamily have a number of common fea-tures and their proteins can be divided into dis-tinct domains: an amino-terminal activationdomain, a central to the function of these pro-teins is the highly conserved DNA-binding do-main (DBD). The second most conserved regionis referred to as the ligand-binding domain(LBD) which mediates ligand-induced transacti-vation and participates in receptor dimerization(Mangelsdorf et al. 1995). Moreover, the region Ifunctions as the DBD contains the two zinc fin-ger motifs and has the activity of binding to hor-mone response elements, and the regions II andIII as the LBD/dimerization domain (Hondaet al. 1993).

RT-PCR analysis of various zebrafish tissuesindicated preferential expression of the SF-1mRNA in liver, followed by brain, testis and lowerexpression in ovary (Liu et al. 1997). Studies byHofsten et al. (2001) demonstrated the presence ofthe Medaka SF-1 and its predominant expressionin the ovary. However, no information is availableon the sequence and RT-PCR analysis of varioustissues indicated preferential expression of SF-1homologues mRNA in goldfish.

The objectives of this study were to character-ize the SF-1 homologues cDNA and investigateits tissue distribution in goldfish as a step to to-wards understanding the molecular mechanismsof SF-1 homologues action. The results also pro-vide for the first time a comparison of the tissuedistribution of SF-1 homologues mRNA duringsexual maturation in female goldfish.

Materials and methods

Animals

Goldfish (Carassius auratus) ranged from 10 to13 cm in length were purchased from AquaticImports (Calgary, Alberta, Canada), and kept at17–18 �C in a semi-recirculating tank. The lightregime was 16 h light, 8 h dark photoperiod.Goldfish were anesthetized with 3-aminobenzoicacid ethyl ester (Sigma) and killed in accordancewith the principles and guidelines of the Cana-dian Council of Animal Care. The gonads wereremoved and weighted for calculated of the go-nadosomatic index (GSI = gonad weight/bodyweight� 100). Tissue samples from female gold-fish at sexually immature (GSI: 4.5–6.0) and ma-ture stages (GSI: 18.5–22.0) which were removed,and stored at )80 �C until used.

Isolation of the SF-1 homologues cDNA

Two highly conserved regions of the chicken(Gallus gallus) FTZ-F1 (Kudo & Sutou 1997),frog (Rana rugosa) SF-1 (Kawano et al. 1998)and medaka (Oryzias latipes) FTZ-F1 (Watanabeet al. 1999) were design mixed primers for thepolymerase chain reaction (PCR).

One of the regions is located at the DBD do-main [primer 1: 5¢-ACAAGTTTGG(G/C)CCCATGTAC-3¢], and the other at the LBL domain[primer 2: 5¢-AGGTGCTTGTGGTA(C/G)AG-GTA-3¢] (Figure 1). Total RNA was extractedfrom each tissue using an RNAgents Total RNAIsolation System (Promega, Madison, WI),according to manufacturer’s instructions. One lgof total RNA was used for cDNA synthesis asdescribed by Choi and Habibi (2003). Amplifica-tion was performed as previously described (Choi& Habibi 2003) using Taq DNA polymerase(Promega) and PCR for 38 cycles of 45 s at94 �C, 45 s at 54 �C, and 1 min at 72 �C, exceptthat the first denaturation was carried out for3 min and the last elongation reaction for 5 min.After electrophoresis on 1% TAE-agarose gels,the DNA fragment was excised and ligated intothe pGEM-T Easy Vector (Promega) accordingto the manufacture’s instructions, and sequenced.

1284

Rapid amplication of cDNA 3¢ ends (3¢ RACE)

The 3¢ RACE technique was used to obtain se-quences downstream of the PCR product using3¢ RACE System (Version 2.0 kit, Gibco/BRL).First-strand cDNA synthesis was initiated at thepoly (A)+ RNA using the oligo (dT) anchor pri-mer [5¢-TGGAAGAATTCGCGGCCGCAGGAAT18-3¢]. The 3¢ RACE-PCR product was ampli-fied by PCR using gene specific primers [primer3: 5¢-TTACGTGGAGAGCGTGTACG-3¢] and[3¢ RACE adaptor primer: 5¢-TGGAAGAATTCGCGGCCGCAG-3¢] (Figure 1) under the fol-lowing conditions; 0.2 lg cDNA as template,10 lM primer 3 and 3¢ RACE adaptor primer,10 mM of each dNTP, and Taq DNA polymerase(5 U/ll, Promega) in 50 ll buffer. NestedPCR was performed using 35 cycles of 94 �C for45 s for denaturing, 57 �C for 45 s for primerannealing, and 72 �C for 1 min for extension,followed by final 1 cycle of 5 min at 72 �C forextension. The final PCR product was amplifiedand T-A cloned into pGEM-T Easy Vector, andsequenced.

Rapid amplication of cDNA 5¢ end (5¢ RACE)

The 5¢ RACE System (Version 2.0 kit, Gibco/BRL) was employed to obtain transcriptsequences upstream of the PCR product. Fivemicrograms of total RNA was reverse transcribedaccording to the kit protocol using a gene specificprimer [primer 4; 5¢-GTCGACGTATGTGTA-3¢]located within the coding sequence (Figure 1). Forthe 5¢ RACE-PCR, two gene specific primers, [pri-mer 5; 5¢-CTGGTGTAGTGCTCTCAGCTT-3¢]and [primer 6; 5¢-ACGCCGGGTACTGTG

CTGGCA-3¢], were designed (Figure 1). The 5¢RACE-PCR was carried out using primer 5 andthe oligo(dG) anchor primer [5¢-GGCCACGCGTCGACTAGTACGGGIIGGGIIGGGIIG-3¢]included in the kit. Second amplication was con-ducted using primer 6 and 5¢ RACE adaptor pri-mer [5¢-GGCCACGCGTCGACTAGTAC-3¢]under the same conditions as 3¢ RACE nested am-plication (Figure 1). Additional controls that am-plify dC-tailed cDNA using each primerindividually (either primer 6 or 5¢ RACE adaptorprimer) were used to identify nonspecific products.The final nested PCR product was amplified andcloned, and the DNA sequence was analyzedusing the GENETYX-WIN (Software Develop.Co., Japan) software package.

Reverse transcriptase-polymerase chain reaction(RT-PCR)

For the RT-PCR, two specific primers were usedto amplify SF-1 cDNA as follows: [primer 7;5¢-ACTACAGCTATGGCACGGAC-3¢] and[primer 8; 5¢-AGATGCAGGTTCTCTTGGCA-3¢] (Figure 1). First, cDNA was synthesized from1 lg total RNA from brain, pituitary, ovary andliver of goldfish. RT reactions were done with re-verse transcribed with an oligo(dT) primer andM-MLV reverse transcriptase (Gibco/BRL). Thereaction mixture was activated at 94 �C for3 min, 38 reaction cycles were conducted as fol-lows: 45 s denaturing at 94 �C, 45 s annealing at54 �C, and 1 min extension at 72 �C, followed by1 cycle of 5 min extension at 72 �C. Fifteen lg ofeach PCR products were electrophoresed on 1%TAE-agarose gels, with a 1 kb plus DNA ladder

Fig. 1. Cloning and sequencing strategy for the goldfish (Carassius auratus) orphan nuclear receptor steroidogenic factor (SF)-1cDNA using the 5¢/3¢ RACE and RT-PCR. The open bar indicates the open reading frame (ORF). Arrows indicate the relativelocation and direction of primers.

1285

(Gibco/BRL) used as a reference to estimate themolecular weights of the amplified fragments.

High-resolution scanner carried out quantifi-cation of PCR amplified fragment and the banddensities were estimated using NIH image soft-ware (NIH, Bethesda, MD). As a control forloading, in each case the loading was controlledby amplification of goldfish b-actin. The densi-tometry process from ethidium bromide stainedgel was optimized for linearity as described in aprevious study (Choi & Habibi 2003).

Statistical analysis

The results were presented as the mean± thestandard deviations (SD), and were analyzed bya one-way ANOVA followed by the Duncan’smultiple range tests were applied for statisticalanalysis. The means were considered statisticallydifferent if p < 0.01.

Results

Cloning and characterization of the goldfishSF-1 cDNA

We designed a set of nucleotide primers based onthe nucleotide sequence, which is highly con-served regions of known various species, as de-scribed in Materials and methods. One majorPCR fragment (1169 base pair) was amplifiedfrom the goldfish pituitary and was separated byelectrophoresis. Goldfish SF-1 cDNA generatedby the 3¢ and 5¢ RACE procedures were subse-quently combined to generate a full-lengthcDNA sequence. The 1,948 bp cDNA had anopen reading frame (ORF) of 1,509 bp that be-gan with the first ATG codon at position 22 bpand ended with a TGA stop codon at position1,531 bp (accession number AF526537). A puta-tive polyadenylation signal AATAAA (Proudfoot& Brownlee 1976) occurred at position 1,908 bp(accession number AF526537).

Homologies of goldfish SF-1 with other speciesof nuclear hormone receptor subfamily

Homological analyses using the GenBank andthe EMBL general database searches indicatedthat the amino acid sequence of the goldfish

SF-1 cDNA has a related high homology withthe other species as follow: medaka (O. latipes)FTZ-F1 (66.9% identity and 82.3% similarity)(Watanabe et al. 1999), frog (R. rugosa) FTZ-F1b (63.2% identity and 81.4% similarity) (Nak-ajima et al. 2000), chicken (G. gallus) FTZ-F1(62.9% identity and 82.1% similarity) (Kudo &Sutou 1997) and zebrafish (D. rerio) FTZ-F1b(61.6% identity and 78.4% similarity) (Chai &Chan 2000) are shown in Figure 2.

The zinc fingers which are in region I (DBDdomain); regions II and III are located in the LBDdomain. When amino acid sequences of these do-mains are aligned with those of FTZ-F1 homo-logues cDNA of other species, regions I and II areappeared to high similarity. Goldfish SF-1 con-tains a high degree of conservation in the region I(90.9–97.0% identity), region II (92.9–100% iden-tity) and region III (56.5–73.9% identity) whencompared to medaka FTZ-F1, zebrafish FTZ-F1,zebrafish FTZ-F1b, chicken FTZ-F1 and frogFTZ-F1b (Figure 2). Apart from the regions I–III,this receptor subclass has a characteristic con-served amino acid sequence, called the FTZ-F1box (Figure 2). It lies adjacent to the zinc fingersare required for high affinity and sequence-specificbinding. The FTZ-F1 box is 30 amino acids and is100% identical among medaka FTZ-F1, zebrafishFTZ-F1, chicken FTZ-F1 and frog FTZ-F1b,except for zebrafish FTZ-F1b (Figure 2)

Tissue distribution of the goldfish SF-1

Using RT-PCR, we investigated the expressionof SF-1 mRNA in various tissues in female gold-fish at sexually immature and mature stage. Theexpression of b-actin was monitored in all tissuesand used as control to normalize for loading.Lower expression of SF-1 was observed in theliver, but extremely low expression was detectedin brain and ovary of the immature female gold-fish. SF-1 mRNA was very highly expressed inmature female goldfish pituitary and brain, fol-lowed by liver which, was lower expressed inovary (Figure 3).

Discussion

Mammalian reproductive function is regulated bythe hypothalamic-pituitary-gonadal axis. SF-1 is

1286

Fig. 2. Deduced amino acid comparisons of the goldfish (Carassius auratus) SF-1. Goldfish amino acid sequences are comparedwith the deduced amino acid sequences of the medaka (Oryzias latipes) FTZ-F1, zebrafish (Danio rerio) FTZ-F1, zebrafish (D. re-rio) FTZ-F1b, chicken (Gallus gallus) FTZ-F1 and frog (Rana rugosa) FTZ-F1b. The sequences were taken from the GenBank/EMBL/DDBJ sequence databases. SF-1 homologues sequences used for alignment are goldfish SF-1 (gfSF-1, AF526537), medakaFTZ-F1 (mFTZF1, AB026834), zebrafish FTZ-F1 (zFTZF1, AF014926), zebrafish FTZ-F1b (zFTZFb, AF198086), chicken FTZ-F1 (cFTZF1, AB002404) and frog FTZ-F1b (fFTZF1, AB035499). To aid comparisons, functional motifs as described by Wonget al. (1996) are indicated; regions I-III, P box, D box, FTZ-F1 box and the activation function (AF)-2 motif. Gaps in the se-quences are indicated as dashes. Dots indicate residues identical to those of gfSF-1.

1287

essential for gonadal development and function,implicates SF-1 as important regulatory role inthe reproductive system (Ingraham et al. 1994).

This study provides the complete coding se-quence of the SF-1 cDNA in the goldfishpituitary. The goldfish SF-1 cDNA was found tocontain an open reading frame of 1509 nucleo-tides encoding a protein of 503 amino acids.SF-1 proteins have three conserved regions. Theregion I functions as the DBD contains the zincfinger motifs and has the activity of binding tohormone response elements, and the regions IIand III as the LBD (Honda et al. 1993). The zincfinger motifs of the steroid hormone receptorsuperfamily have two functional domains; the so-called P and D boxes. Region I recognizes andbinds the specific sequences of target genes, theexpression of which is regulated by the nuclearreceptor. The P box in the first zinc finger distin-guishes between the sequences of hormone re-sponse elements. Goldfish SF-1 was found to be

a P box, which is positioned at the base of thezinc finger, and is ESCKG in all FTZ-F1 box-containing receptors. The D box in the secondzinc finger recognizes the spacing of those ele-ments. Furthermore, the two sequences in theputative LBD, called region II and III, have beenused to classify nuclear hormone receptor (Wanget al. 1989). Region II is highly conserved andregion III is less conserved (Figure 2).

Some studies revealed that transcriptionallyactive FTZ-F1 homolog in vertebrate showedconserved activation function (AF)-2 motif se-quence, LLIEML, is located in LBD at aminoacid 491–496 (Figure 2), indicating that transcrip-tional activation is dependent on ligand bindingas reported previously (Tora et al. 1989). Galar-neau et al. (1996) reported that complete removalof the AF-2 motif from the LBD of rat LRH-1also caused the loss of its trans-activationfunction. This result suggests that the transcrip-tional activities of the members of the FTZ-F1

Fig. 3. One microgram of total RNA prepared from brain (B), pituitary (P), ovary (O), liver (L) and control as a no tissue (N)were reverse transcribed and amplified in immature and mature male goldfish using SF-1 specific primer. Tissue distribution ofgoldfish SF-1 was analyzed by RT-PCR. The expression of b-actin mRNA was evaluated in each RT reaction product as a loadingcontrol. The expression level of each tissue was normalized with respect to the b-actin signal, and expressed as relative expressionlevel. An asterisk indicates a significant difference compared between immature and mature goldfish (p < 0.01). Values aremean±the standard deviations of these four experiments, each using separate female goldfish.

1288

subfamily are mainly due to their C-terminalregions.

SF-1 is apparently able to stimulate the syn-thesis of not only corticosteroids, such as gluco-corticoids, in the adrenal cortex but also theproduction of estrogens and androgens in theovary and testis by activating the expression ofthe enzymes that synthesize precursors to thesehormones and by increasing the expression ofaromatase (Lynch et al. 1993), which conductsbioconversion of androgens to estrogens. Addi-tionally, it is also able to indirectly control thesynthesis of primary steroid sex hormones byacting at the pituitary level to influence theexpression of the gonadotropin hormone genes(Barnhart & Mellon 1994).

In this study, we also compared the expres-sion pattern of SF-1 in a variety of tissues in fe-male goldfish at immature and mature stage byRT-PCR. Lower expression of SF-1 was ob-served in the liver, but not detected in brain andovary of the immature female goldfish. Presenceof SF-1 was the predominant expression in thepituitary and brain of mature female goldfish.Also, in the mature female goldfish, weak tran-script was detected in brain and ovary. Interest-ingly, RT-PCR analysis revealed that theexpression of goldfish SF-1 became higher in thebrain and weaker in the liver during maturing fe-male, respectively (Figure 3). This agrees with aprevious report (Kawano et al. 1998) that highexpression of mature frog FTZ-F1 mRNA wasdetected in the brain, but not ovary. Interest-ingly, mature female goldfish liver SF-1 mRNAlevels were decreased significantly compared toimmature female goldfish. Taken together, thesedata indicate a global role for FTZ-F1 at eachstages of liver and gonadal function.

On the other hand, the FTZ-F1 gene familyconstitutes a subgroup of orphan nuclear recep-tors, which can be divided into two groups,LRH/FTF (liver receptor homologue protein/a-fetoprotein transcription factor) and SF-1, basedon function, tissue distribution (Galarneau et al.1996, Hofsten et al. 2001) and phylogenetic treeanalysis (Nakajima et al. 2000). In zebrafish,however, Hofsten et al. (2001) demonstrated thatwhile the expression in liver is indicative of aLRH/FTF function, the expression in other tis-sues, such as the pituitary, brain and gonad isindicative of SF-1 function. Moreover, amino

acid sequences in the regions I and II of goldfishSF-1 showed high similarity to not only LRHbut also SF-1, suggesting the importance of eachregion for the function of these proteins. In theseobservations, coupled with obtained expressionpatterns, indicate that goldfish FTZ-F1 homo-logues exhibit characteristics that are indicativeof both LRH/FTF- and SF-1-like genes, mayhave other developmental roles in goldfish. SF-1mRNA was strongly expressed in immature andmature goldfish pituitary. SF-1 regulates theactivity of steroidogenic enzyme in bovine adre-nal cells and luteal cells (Liu & Simpson 1997).Therefore, SF-1 may also control steroidogenesisin goldfish pituitary.

Interestingly, RT-PCR analysis revealed thatthe expression of goldfish SF-1 was found in thegoldfish liver but it became weaker in the liver ofmature female goldfish. The findings suggest thatSF-1 homologues may plays an important rolesin the regulation of liver function since it is par-ticularly less expression in developing femalegoldfish liver. Moreover, at present, little infor-mation is available on the relative importance ofSF-1 homologues in the regulation of vitellogene-sis and reproduction in goldfish and other teleo-sts species. The present results are in accord withfindings in mammals and other vertebrates con-cerning tissue specific expression of SF-1 homo-logues. Information obtained in this study ontissue distribution of SF-1 homologues providesa framework for better understanding of physio-logical significance of these groups of the orphannuclear receptor in goldfish. The significance ofthese findings remains to be investigated as moreinformation on regulation of brain and liver, andseasonal variation of SF-1 homologues becauseavailable in goldfish and other vertebrates.

Acknowledgement

This work was supported by a grant from theNatural Sciences and Engineering ResearchCouncil of Canada.

References

Barnhart KM, Mellon PL (1994) The orphan nuclear receptor,Steroidogenic factor-1, regulates the glycoprotein hormone

1289

a-subunit gene in pituitary gonadotropes.Mol. Endocrinol. 9:878–885.

Chai C, Chan W-K (2000) Developmental expression of a novelFTZ-F1 homologue, ff1b (NR5A4), in the zebrafish (Daniorerio).Mech. Develop. 91: 421–426.

Choi CY, Habibi HR (2003) Molecular cloning of estrogenreceptor a and expression pattern of estrogen receptorsubtypes in male and female goldfish. Mol. Cell Endocrinol.204: 169–177.

Galarneau L, Pare J-F, Allard D, Hamei D, Levesque L,Tugwood JD, Green S, Belanger L (1996) The a-1 fetopro-tein locus is activated by a nuclear receptor of the DrosophilaFTZ-F1 family. Mol. Cell Biol. 16: 3853–3865.

Galarneau L, Drouin R, Belanger L (1998) Assignment of thefetoprotein transcription factor gene (FTF) to humanchromosome band 1q32.11 by in situ hybridization. Cytoge-net. Cell Genet. 82: 269–270.

Hofsten JV, Jones I, Karlsson J, Olsson P-E (2001) Develop-mental expression patterns of FTZ-F1 homologues in zebra-fish (Danio rerio). Gen. Comp. Endocrinol. 121: 146–155.

Honda S, Morohashi K, Nomura M, Takeya M, Kitajima M,Omura T (1993) Ad4BP regulating steroidogenic P-450 geneis a number of steroid hormone receptor superfamily. J. Biol.Chem. 268: 7498–7502.

Ikeda Y, Shen WH, Ingraham HA, Parker KL (1994) Devel-opmental expression of mouse steroidogenic factor-1, anessential regulator of the steroid hydroxylases. Mol. Endo-crinol. 8: 654–662.

Ingraham HA, Lala DS, Ikeda Y, Luo X, Shen WH, NachtigalMW, Abbud R, Nilson JH, Parker KL (1994) The nuclearreceptor SF-1 acts multiple levels of the reproductive axis.Genes Dev. 8: 2302–2312.

Kawano K, Miura I, Morohashi K, Takase M, Nakamura M(1998) Molecular cloning and expression of the SF-1/Ad4BPgene in the frog. Rana Rugosa. Gene 222: 169–176.

Kudo T, Sutou S (1997) Molecular cloning of chicken FTZ-F1-related orphan receptors. Gene 197: 261–268.

Liu D, Le Drean Y, Ekker M, Xiong F, Hew CL (1997) TeleostFTZ-F1 homolog and its splicing variant determine theexpression of the salmon gonadotropin II beta subunit gene.Mol. Endocrinol. 11: 877–890.

Liu Z, Simpson ER (1997) Steroidogenic factor 1 (SF-1) andsp1 are required for regulation of bovine CYP11A geneexpression in bovine luteal cells and adrenal Y1 cells. Mol.Endocrinol. 11: 127–137.

Lynch JP, Lala DS, Peluso JJ, Luo W, Parker KL, White BA(1993) Steroidogenic factor 1, an orphan nuclear receptor,regulates the expression of the rat aromatase gene in gonadaltissues. Mol. Endocrinol. 7: 776–786.

Mangelsdorf DJ, Thummel C, Beato M, Herrlich P, Schutz G,Umesono K, Blumberg B, Kastner P, Mark M, Chambon P,Evans RM (1995) The nuclear receptor superfamily: thesecond decade. Cell 83: 835–839.

Nakajima T, Takase M, Nakamura M (2000) Two isoforms ofFTZ-F1 messenger RNA: molecular cloning and theirexpression in the frog testis. Gene 248: 203–212.

Proudfoot NJ, Brownlee GG (1976) 3¢ Non-coding regionsequences in eukaryotic messenger RNA. Nature 263:211–214.

Smith C, Smith MJ, Sinclair AH (1999) Expression of chickensteroidogenic factor-1 during gonadal sex differentiation.Gen. Comp. Endocrinol. 113: 187–196.

Tora L, White J, Brou C, Tasset D, Webster N, Scheer E,Chambon P (1989) The human estrogen receptor has twoindependent nonacidic transcriptional activation functions.Cell 59: 477–487.

Wang LH, Tsai SY, Cook RG, Beattie WG, Tsai MJ,O¢Malley BW (1989) COUP transcription factor is amember of the steroid receptor superfamily. Nature (Lon-don) 340: 163–166.

Watanabe M, Tanaka M, Kobayashi D, Yoshiura Y, Oba Y,Nagahama Y (1999) Medaka (Oryzias latipes) FTZ-F1potentially regulates the transcription of P-450 aromatasein ovarian follicles. cDNA cloning and functional charac-terization. Mol. Cell Endocrinol. 149: 221–228.

Wong M, Ramayya MS, Chrousos GP, Driggers PH, ParkerKL (1996) Cloning and sequence analysis of the human geneencoding steroidogenic factor-1. J. Mol. Endocrinol. 17:139–147.

Yu Y, Li W, Su K, Yussa M, Han W, Perrimon N, Pick L(1997) The nuclear receptor Ftz-F1 is a cofactor for theDrosophila homeodomain protein Ftz. Nature 385: 552–555.

1290

Enhancement of isoflavone synthase activity by co-expression of P450reductase from rice

Dae Hwan Kim1, Bong Gyu Kim1, Hyo Jung Lee1, Yoongho Lim1, Hor Gil Hur2 &Joong-Hoon Ahn1,*1Bio/Molecular Informatics Center, Department of Molecular Biotechnology, Konkuk University, 143-701,Seoul, Korea2Department of Environmental Science and Engineering and International Environmental Research Center,Kwangju Institute of Science and Technology, Gwangju, Korea*Author for correspondence (Fax: +82-2-3437-6106; E-mail: jhahn@konkuk.ac.kr)

Received after revisions 20 June 2005; Accepted 20 June 2005

Key words: cytochrome P450, cytochrome P450 reductase, Saccharomyces cerevisiae, secondarymetabolism

Abstract

Plant cytochrome P450s interact with a flavoprotein, NADPH-cytochrome P450 reductase (CPR), totransfer electrons from NADPH. The gene for rice P450 reductase (RCPR) was cloned and expressed inSaccaromyces cerevisiae, where the specific activity of the expressed RPCR was 0.91 U/mg protein. Whenisoflavone synthase gene (IFS) from red clover, used as a model system of plant cytochrome P450, wasco-expressed with RCPR in yeast, the production of genistein from naringein increased about 4.3-fold,indicating that the RCPR efficiently interacts with cytochrome P450 to transfer electrons from NADPH.

Introduction

Cytochrome P450 monooxygenases (P450s) play acritical role in many biosynthetic pathways,including the detoxification of exogenous com-pounds. For example, in animals, P450s areinvolved in drug, steroid and fatty acid metabo-lism. In plants, their activity mediates the synthe-sis of lignins, UV protectants, pigments, defensecompounds, fatty acids, hormones, and signalingmolecules, as well as the catabolism of herbicides,insecticides, and pollutants (Schuler & Werck-Reichhart 2003). The biological functions of P450srely on an electron donor, NADPH-cytrochromeP450 reductase (CPR). CPR transfers two elec-trons from NADPH to P450s (Porter et al. 1987).

From the genome sequences, Arabidopsis has273 P450s and rice more than 300, though thefunction of most of these is unknown. Thedevelopment of an in vitro assay would be a critical

step an in analyzing the function of individualP450s. Our group has been interested in thefunctional characterization of P450s from riceand we have now developed an in vitro assay sys-tem in which CPR from rice was cloned andexpressed in Saccharomyces cerevisiae. Then,isoflavone synthase, a plant P450, was expressedtogether with CPR in the yeast to enhance theP450 activity.

Materials and methods

Cloning of rice CPR gene and CPR assay

A reverse-transcriptase polymerase chain reaction(RT-PCR) was carried out to clone the CPRgene from rice with primers of the followingsequences: CAACCAAACCCTCGCTTC as for-ward primer and GCTAGAGCGAGCTATTTC

Biotechnology Letters (2005) 27: 1291–1294 � Springer 2005DOI 10.1007/s10529-005-0221-7

TGAAS as reverse primer. The resulting PCRproduct was subcloned into pGMET vector (Pro-mega, Madison, WI, USA) and sequenced. Forthe in vitro rice CPR enzyme assay, a cytochromec reductase [NADPH] assay kit (Sigma) wasused. The rate of reduction was calculated bydifferential absorption at 550 nm (Ro et al.2002). One unit of activity was defined as theamount of CPR produced for the reduction of1 lmol cytochrome c per min.

Functional expression of rice CPRin Saccharomyces cerevisiae

To subclone the rice CPR gene into the pESC-His vector (Stratagene, La Jolla, CA, USA), twonew primers were synthesized: a forward primer,containing the initiation codon ATGGCGCTGGCGCTGGA, a reverse primer, contained arestriction enzyme site SpeI (ACTAGT), and astop codon ATACTAGTTCACCATACGTCACGGAGCCTGC. PCR was carried out with PfuTaq polymerase (Stratagene, La Jolla, CA, USA)and the resulting PCR product was digested withSpeI. The pESC-His vector was cut with EcoRI,blunted with Klenow enzyme, digested with SpeI,and then used for the ligation. Expression ofRCPR was followed according to manufacture’sinstructions. Microsomes from the transformantwere isolated by the method of Stansfield & Kelly(1996).

Microsomal protein (1 mg), containing eitherisoflavone synthase (IFS) or IFS plus RCPRmixed with 1 mM NADPH and 100 lM naringe-nin, was incubated at 30 �C for 2 h. The reactionmixture was extracted twice with ethyl acetate;the solvent was then evaporated in vacuo and theresidue was dissolved in methanol. Analysis offlavonoid was by HPLC as described in Kimet al. (2002). Concentrations of naringenin andgeninstein were determined from a standard curvecalculated with various known concentrations ofboth compounds against the peak area detectedon HPLC. For both compounds, the standardHPLC detection curve was linear up to 1 lmol.

Results and discussion

The rice genome database was searchedwith Arabidopsis CPR gene (GenBank accession

number X66016) to find rice CPR. Several ricegenes, which showed homology with ArabidopsisCPR, were found and some of them were anno-tated as cytochrome P450 reductase. Amongthem, one gene (XP_474161, RCPR) showing thehighest homology was cloned by RT-PCR. TheRCPR consists of a 2088-bp open reading frame.The predicted protein has domains that are com-monly found in other CPR; FMN, FAD, andNADPH binding domains. It showed 82% iden-tity with NADPH-cytochrome P450 reductase(AAG17471) from Triticum aestivum and 72%with that from Populus balsamifera subsp.trichocarpa � Populus deltoids (AAK15259).

To express the RCPR in yeast, it was clonedinto the pESC-His vector under the control of agalactose-inducible promoter and the resultingconstruct was transformed into S. cerevisiae IN-VSc1 (his-, leu-, trp-, ura-). Transformants thatgrew in the absence of histidine were selected. Asa control, the pESEC-His vector was trans-formed into the same yeast strain. From fourindependent yeast transformants, containingeither RCPR or the expression vector, micro-somes were isolated and CPR activity measuredby reduction of cytochrome c. Specific activity ofthe crude RCPR from the four dependent trans-formants ranged from 1.12 to 1.32 lmol min)1

mg)1 with an average of 1.21 lmol min)1 mg)1,while yeast, containing the vector only, showed aspecific activity of 0.28–0.35 lmol min)1 mg)1,with an average of 0.31. Thus, specific activity ofthe expressed RCPR was 0.9 lmol min)1 mg)1.The reduction of cytochrome c by recombinantrice CPR protein was dependent on NADPH,but not on NADH, as shown in other CPRs(Koopmann & Hahlbrock 1997, Mizutani &Ohta 1998). Therefore, based on the in vitroenzyme assays, we can conclude that the RCPRcDNA encodes functional rice CPR enzyme.

As endogenous CPR activity is a limiting fac-tor in yeast when foreign P450s are over-expressed (Urban et al. 1994), we determinedwhether the expressed rice CPR activity en-hanced cytochrome P450 activity in yeast. IFSfrom red clover was used as a model system ofplant cytochrome P450. In a previous study(Kim et al. 2003), it was shown that IFSexpressed in yeast could convert naringenin intogenistein. The yeast strain containing RCPR wasretransformed with an IFS construct (Figure 1).

1292

The resulting transformants were selected onuracil- and histidine-deleted media, since the IFSconstruct contains the ura3 gene as a selectionmarker. Several transformants were analyzed forthe presence of the constructs, RCPR and IFS.As a control, a yeast transformant, containingonly IFS, was generated. Microsomes from threeindependent transformants, with either IFS orIFS and RCPR, were isolated, respectively andIFS activity was measured.

The reaction product was analyzed with HPLCto determined whether a reaction had occurred.Naringenin was eluted at 12.5 min and genistein at15.5 min (Figure 2 a, b). Reaction products withIFS showed a smaller genistein peak than thosewith IFS plus RCPR (Figure 2 c, d). To measurethe amount of genistein generated by IFS, severalconcentrations of narigenin and genistein wereanalyzed by HPLC and the HPLC absorbancewas used as the standard for calculating the con-centration of the reaction product. HPLC absor-bance vs. concentration of each flavonoids showeda linear relationship from 0 to 200 lM. IFS itselfproduced an average 18 lM genistein from 100 lM

naringenin, while IFS co-expressed with RCPRproduced 77 lM, equal to about 4.3-fold increasein IFS activity. This indicates that RCPR en-hanced the IFS activity in yeast cells. Even thoughIFS was cloned from red clover, RCPR couldeffectively transfer electrons to IFS. Therefore, thecurrent CYP expression system could be applica-ble for the functional studies of CYPs, not onlyfrom rice but also from other plants.

S.cerevisiae

IFS (P450)

URA 3

pGA

L1

pYES2

IFS (P450)

URA 3

pGA

L1

pGA

L1

pYES2

RCPR

pGA

L10

pGA

L10

HIS 3

pESC-HIS

(His3-, Leu2-, Trp1-, Ura3-)

Fig. 1. Co-expression strategy of isoflavone synthase (IFS) andrice cytochrome P450 reductase (RCPR) in Saccharomycescerevisiae. IFS converts naringenin into genistein. RCPR trans-fers two electrons from NADPH to IFS. RCPR, with his3 geneas a selection marker, was transformed into S. cerevisiae andthen IFS, a plant P450, with ura3 gene as a selection marker,was transformed again into the yeast strain that was originallytransformed with RCPR.

0 5 10 15 200

2040

60

80

100

120mAU

0 5 10 15 200

2040

60

80

100

120mAU

Naringenin

Genistein

mAU

0 5 10 15 200

2040

60

80

100

120

0 5 10 15 200

2040

60

80

100

120mAU

O

O

HO

OHOH

O

O

HO

OH

OH

(a)

(b)

(c)

(d)

Time(min)

Fig. 2. HPLC elution profile of reaction product by red clo-ver isoflavone synthase co-expressed with rice P450 reductase.Microsomes from either vector pES-His (a, b), isoflavone syn-thase gene (c), or isoflavone synthase co-expressed with riceP450 reductase (d) were incubated with 1 mM NADPH and100 lM naringenin (genistein was used instead of naringeninin b). The reaction products were analyzed by HPLC.

1293

Acknowledgements

This work was supported by a grant from theBiogreen 21 Program, Rural DevelopmentAdministration, Republic of Korea and partiallyby KRF2004-F00019 (KRF).

References

Kim BG, Kim SY, Song HS, Lee C, Hur HG, Kim SI, Ahn J-H(2003) Cloning and expression of the isoflavone synthase gene(IFS-Tp) from Trifolium pretense.Mol. Cells 15: 301–306.

Koopmann E, Hahlbrock K (1997) Differentially regulatedNADPH:cytochrome P450 oxidoreductases in parsley. Proc.Natl. Acad. Sci. U S A. 94: 14954–14559.

Mizutani M, Ohta D (1998) Two isoforms of NADPH:Cyto-chrome P450 reductase in Arabidopsis thaliana. PlantPhysiol. 116: 357–367.

Porter TD, Wilson TE, Kasper CB (1987) Expression of afunctional 78,000 dalton mammalian flavoprotein,

NADPH–cytochrome P-450 oxidoreductase, in Escherichiacoli. Arch. Biochem. Biophys. 254: 353–367.

Ro D-K, Ehlting J, Douglas CJ (2002) Cloning, functionalexpression, and subcellular localization of multipleNADPH-cytochrome P450 reductase from hybrid poplar.Plant Physiol. 130: 1837–1851.

Schuler MA, Werck-Reichhart D (2003) Functional genomicsof P450s. Annu. Rev. Plant. Biol. 54: 629–667.

Stansfield I, Kelly SL (1996) Purification and quantification ofSaccharomyces cerevisiae cytochrome P450. In: Evans IV,ed. Yeast Protocols: Methods in Cell and Molecular Biology,New Jersey: Humana Press, pp. 355–366.

Urban P, Werck-Reichhart D, Teutsch HG, Durst F, RegnierS, Kazmaier M, Pompon D (1994) Characterization ofrecombinant plant cinnamate 4-hydroxylase produced inyeast. Kinetic and spectral properties of the major plantP450 of the phenylpropanoid pathway. Eur. J. Biochem. 222:843–850.

Urban P, Mignotte C, Kazmaier M, Delorme F, Pompon D(1997) Cloning, yeast expression, and characterization of thecoupling of two distantly related Arabidopsis thalianaNADPH-cytochrome P450 reductase with P450 CYP73A5.J. Biol. Chem. 272: 19176–19186.

1294

Structural characterization of b-glucans of Agaricus brasiliensis in differentstages of fruiting body maturity and their use in nutraceutical products

Carla Maısa Camelini1, Marcelo Maraschin2, Margarida Matos de Mendonca1,*,Cezar Zucco3, Antonio Gilberto Ferreira4 & Leila Aley Tavares41Departamento de Microbiologia e Parasitologia, Centro de Ciencias Biologicas, 88040-900, Santa Catarina,Florianopolis, Brasil2Departamento de Fitotecnia, Centro de Ciencias Agrarias, 88040-900, Santa Catarina, Florianopolis, Brasil3Departamento de Quımica, Universidade Federal de Santa Catarina, 88040-900, Santa Catarina,Florianopolis, Brasil4Departamento de Quımica, Universidade Federal de Sao Carlos, 13565-905, Sao Paulo, Sao Carlos, Brasil*Author for correspondence (Fax: +55-48-331-9258; E-mail: margarid1@terra.com.br)

Received 17 February 2005; Revisions requested 22 February 2005; Revisions received 20 June 2005; Accepted 21 June 2005

Key words: Agaricus blazei, Agaricus brasiliensis, b-glucan, nutraceutical, polysaccharide

Abstract

b-Glucans of Agaricus brasiliensis fruiting bodies in different stages of maturity were isolated and char-acterized by FTIR and NMR. These fractions had greater amount of (1 fi 6)-b-glucan and the (1 fi 3)-b-glucan increased with fruiting bodies maturation. Yields of b-glucans increased from 42 mg b-glucans g)1

fruiting bodies (dry wt) in immature stage to 43 mg g)1 in mature stage with immature spores, and de-creased to 40 mg g)1 in mature stage with spore maturation. Mature fruiting bodies, which included theseglucans, have potential therapeutical benefits for use in nutraceutical products.

Introduction

Agaricus brasiliensis Wasser & Didukh (Wasseret al. 2002) (=Agaricus blazei ss. Heinem.), knownin Brazil as Cogumelo medicinel, has been widelycultivated in the country because of its medicinalproperties such as immunomodulatory and anti-tumor activities (Kawagishi et al. 1989, Mizunoet al. 1990, Ohno et al. 2001, Dong et al. 2002).The mushroom is commercialized in severalcountries as a nutraceutical product which is anovel class of dietary supplements including par-tially refined extract or dried biomass from themushroom made into a capsule or tablet (Chang& Buswell 1996). Agaricus brasiliensis is har-vested in Brazil mostly in the immature stagewhen the cap is still closed to meet exportationstandards. Although immature fruiting bodieshave not yet achieved their highest biomass, it is

at this stage that they reach the highest marketvalue for exportation. Farmers usually discardthe mature fruiting bodies. The quality of a nu-traceutical is dependent on the chemical compo-sition of the fruiting body, particularly inrelation to the content of b-glucans. However, nostudies have been developed to characterize theb-glucans at different stages of fruiting bodymaturity and on potential use on the preparationof nutraceutical. Kawagishi et al. (1989) charac-terized the b-glucans with antitumoral propertiesof A. blazei and detected an alkali soluble(1 fi 6)-b-D-glucan with no (1 fi 3)-b-linkages.Mizuno et al. (1990), using a water extractionmethod, identified (1 fi 6)–(1 fi 3)-b-D-glucansin the same species. Recently, Ohno et al. (2001)working on an alkali-soluble fraction detected(1 fi 6)-b-D-glucans and a small but significantamount of (1 fi 3)-b-D-glucans in A. blazei.

Biotechnology Letters (2005) 27: 1295–1299 � Springer 2005DOI 10.1007/s10529-005-0222-6

None of these authors indicated the stage ofdevelopment of the fruiting bodies selected intheir studies. In this study, we examined thestructural evolution of water-soluble polysaccha-rides such as b-glucans of A. brasiliensis in threedifferent stages of fruiting body maturity.

Materials and methods

Fruiting body selection

Commercially cultivated fruiting bodies of Agari-cus brasiliensis (strain UFSC-51) were obtainedin Biguacu, Santa Catarina, Southern Brazil, in2003. The fruiting bodies were harvested anddried in different stages of maturity: immature(cap closed) and mature (cap opened). The ma-ture stage was further characterized into imma-ture spores and mature spores as seen onFigure 1.

Extraction of the polysaccharide fraction

Cell wall polysaccharides were extracted andpurified according to Mizuno et al. (1990). Sam-ples of dried fruiting bodies of A. brasiliensis(20 g) were grounded, washed with 120 ml 85%(v/v) ethanol and filtered. The residue was wa-shed with 350 ml 85% (v/v) ethanol holding at80 �C for 3 h (3 times). The polysaccharides weresequentially extracted with 350 ml water holdingsat 100 �C for 3 h (3 times). These aqueous frac-tions were collected by filtration, followed by theaddition of 4 vol. 95% (v/v) ethanol. The mix-ture was then held overnight to obtain the poly-saccharide fraction, which was concentrated anddialyzed against distilled water. This fraction wasthen freeze-dried, weighed and analyzed.

Polysaccharide purification

Each sample (1 g) was dissolved in distilled waterand passed through a DEAE-cellulose columnchromatography (2 cm width� 35 cm length).The neutral fraction eluted with water (105 ml)was discarded and selected fractions of 0.25, 0.5and 0.75 M NaCl (105 ml each) were concentratedand then dialyzed extensively. The polysaccharidewas fractionated according to the highest molecu-lar weight on a Toyopear HW-65F (Tosoh) col-umn (2 cm width� 35 cm length) selecting thefirst 35 ml eluted. The b-glucans (unadsorbed)was separated using Con A-Sepharose 4B (FlukaBiochemika) column (1.5 cm width� 10 cmlength), freeze-dried, weighted and analyzed.

Polysaccharide analyses

The polysaccharide fractions (before and afterpurification of b-glucans) of A. brasiliensis fromeach stage of maturity were analyzed. Proteincontent was determined using the Bradfordmethod with BSA as reference. The FTIR spec-tra were determined with an ABD Bomem Inc.FTLA 2000 spectrometer and KBr discs. 1H and13C NMR spectra of 30 mg and 120 mg polysac-charides fractions, respectively in D2O (600 ll)were recorded at 298 K using a Bruker DRX400spectrometer operating at 9.4 Tesla and TSPA-d4like external reference.

Results and discussion

Polysaccharide analyses

The FTIR spectra (Figure 2) suggested the pres-ence of a small amount of protein (band at1540 cm)l) which was confirmed by the Bradfordmethod (Table 1) and the absence of uronic acids(no carbonyl bands over 1700 cm)l). The charac-teristic bands of b-glucans occurred in the 1000–1100 cm)1 region due to O-substituted glucoseresidues. The band at 1400 cm)l evidenced thepresence of a b-glucan. These spectra showed aweak band at 890 cm)l revealing a b configura-tion on the main glucan. The weak band at910 cm)l and 850 cm)l indicated the presence ofa a-glucan (Gutierrez et al. 1996).

Fig. 1. Agaricus brasiliensis fruiting bodies in different stagesof maturity: (SI) immature (cap closed), (SII) mature (capopened) with immature spores, and (SIII) mature with maturespores.

1296

13C NMR spectra from polysaccharide frac-tions (before and after purification of b-glucans)from A. brasiliensis, in each stage of maturity,are shown in Figure 3. Assignment of each spec-trum was made by comparison with previouslypublished spectra (Dong et al. 2002, Mizunoet al. 1990, Ohno et al. 1985, 2001, Saito et al.1976, York 1995).

Agaricus brasiliensis fruiting bodies in differ-ent stages of maturity contained b-glucans andalso a-glucans. All spectra showed six major

signals on the spectra assigned to (1 fi 6)-b-glu-cosidic linkages (Figure 3). The anomeric C-1signal around 105.1 ppm with closely located sig-nals at 104.7 ppm and 104.4 ppm were attributedas b configurations. The substituted C-6 signalcould be identified at 71.0 ppm from the reversepeak in the DEPT spectrum and non-substitutedC-6 signal at 62.9 ppm, suggesting a higheramount of (1 fi 6)-b-glucan than (1 fi 3)-b-glu-can. Fractions from all stages before purificationof b-glucans showed the signal at 86.5 ppm,attributed to substituted C-3, which was weakerthan the non-substituted C-3 signal at 77.8 ppm(Figure 3). This suggested that most C-3 was notsubstituted, and a small amount of (1 fi 3)-b-glucan was present. The signal at 86.5 ppm wasnot detected in the fraction after purification ofb-glucans from immature fruiting bodies (SI-b),suggesting a smaller amount of the (1 fi 3)-b-glucans than other fractions. All spectra showedsignals at 83.0 ppm, suggesting a (1 fi 2)-b-glu-cosidic linkage. The signals at 102.0 ppm and100.2 ppm indicated a-configurations as well as(1 fi 4)-a- and (1 fi 6)-a-linkages at 79.0 ppmand 68.0 ppm, respectively. Fruiting bodies inmature stages showed an increased signal on C-1(102 ppm) of a (1 fi 4)-a-glucosidic linkage, sug-gesting an increased amount of this type of glu-can with spore matured. The 1H NMR spectra ofboth fractions showed the anomeric signalsaround 4.6 and 3.2 ppm and confirmed theb-glucans. Table 1 shows characteristics of glu-cans in fruiting bodies in different stages of

Fig. 2. FTIR spectra of polysaccharide fractions obtainedfrom Agaricus brasiliensis fruiting bodies in different stages ofmaturity. (SI-a) aqueous fraction from immature stage iso-lated at 100 �C (before purification of b-glucans) and (SI-b)same fraction after purification of b-glucans; (SII-a) aqueousfraction from mature stage with immature spores isolated at100 �C and (SII-b) same fraction after purification of b-glu-cans; (SIII-a) aqueous fraction from mature stage with ma-ture spores isolated at 100 �C and (SIII-b) same fraction afterpurification of b-glucans.

Table 1. Structural characterization, the yield of water soluble glucans at 100 �C from A. brasiliensis fruiting bodies in differentstages of maturity and the amount of protein detected in the fractions before and after purification of b-glucans.

Stages of

maturityaGlucans Protein

(mg g)1 dry wt)

(1 fi 6)-b- (1 fi 2)-b- (1 fi 3)-b- (1 fi 4)-a- (1 fi 6)-a- Yield (mg g)1 dry wt)

SI- a +++b +d + ) + 102 6.0

b +++ + )e ) ) 42 5.8

SII- a +++ + ++c ++ + 106 7.2

b +++ + ++ ) ) 43 6.0

SIII- a +++ + ++ +++ + 111 9.9

b +++ + ++ ) ) 40 6.7

aMaturity stages of fruiting bodies from which glucans were isolated: (SI-a) aqueous fraction from immature stage isolated at 100 �C(before purification of b-glucans) and (SI-b) same fraction after purication of b-glucans; (SII-a) aqueous fraction from mature stagewith immature spores isolated at 100 �C and (SII-b) same fraction after purification of b-glucans; (SIII-a) aqueous fraction frommature stage with mature spores isolated at 100 �C and (SIII-b) same fraction after purification of b-glucans.b+++ Highest amount, c++ intermediate amount, d+ smallest amount, e- no/low level of glucans detected in the fractions.

1297

maturity and the yield of each polysaccharidefraction before and after purification of b-glu-cans. Additionally, the small amount of proteindetected is shown.

The yield and structural diversity of glucansincreased as the fruiting bodies matured. Inmature stages the amount of (1 fi 3)-b-glucans(SII-b and SIII-b) was higher than in the imma-ture stage. These glucans are possibly side bran-ches of a (1 fi 6)-b-backbone as indicated byDong et al. (2002) and Ohno et al. (2001) who de-scribed that b-glucans of A. blazei fruiting bodieshad a (1 fi 6)-b-backbone structure with (1 fi 3)-b-side branches. Mol & Wessels (1990) alsoshowed a greater proportion of (1 fi 6)-b-side

branches on the (1 fi 3)-b-backbone during mat-uration of fruiting bodies of A. bisporus. It isimportant to emphasize that linear (1 fi 6)-b-glu-can extracted from Penicillium islandicum did notpresent bioactivity (Ohno et al. 1986). However,(1 fi 3)-b-side branches are structurally impor-tant and enhance the immunomodulatory activityat polysaccharides (Dong et al. 2002). As Mizunoet al. (1990) evidenced, important anti-tumoractivity is linked to a water-soluble (1 fi 6)–(1 fi 3)-b-D-glucan. As a consequence, maturefruiting bodies of A. brasiliensis should be usedfor nutraceutical products because they containthese important glucans. Furthermore, a signifi-cant increase on another water soluble (1 fi 4)-a-glucan with anti-tumor activity also occurred dur-ing maturation from SII to SIII (Mizuno et al.1990). Cap-opened, more fragile mature fruitingbodies of A. brasiliensis should be selected overimmature ones for the production of nutraceuti-cals. Additionally, this strategy will provide theconsumer with a higher diversity of glucans, opti-mizing bioactivities such as the antitumoral one,additionally allowing farmers an efficient andprofitable use of the mushroom biomass.

Acknowledgements

This work was partially supported by a grantfrom the Conselho Nacional de DesenvolvimentoCientıfico e Tecnologico (CNPq). The senior au-thor would like to thank CNPq for the Biotech-nology fellowship and the Mushroom Farmers ofSanta Catarina. We thank Dr. Admir Giachinifor editorial comments on this manuscript.

References

Chang ST, Buswell JA (1996) Mushroom nutriceuticals. WorldJ. Microb. Biotech. 12: 473–476.

Dong Q, Yao J, Yang X, Fang J (2002) Structural character-ization of water-soluble b-D-glucan from fruiting bodies ofAgaricus blazei Murr. Carbohyd. Res. 337: 1417–1421.

Gutierrez A, Prieto A, Martlnez AT (1996) Structural charac-terization of extracellular polysaccharides produced by fungifrom the genus Pleurotus. Carbohyd. Res. 281: 143–154.

Kawagishi H, Inagaki R, Kanao T, Mizuno T (1989) Fractionand antitumor activity of the water-insoluble residue ofAgaricus blazei fruiting bodies. Carbohydr. Res. 186:267–273.

Fig. 3. 13C NMR spectra of polysaccharide fractions obtainedfrom Agaricus brasiliensis fruiting bodies, in different stages ofmaturity, in D2O at 298 K and a number of scans between30,000 and 35,000. (SI-a) aqueous fraction from immaturestage isolated at 100 �C (before purification of b-glucans) and(SI-b) same fraction after purification of b-glucans; (SII-a)aqueous fraction from mature stage with immature sporesisolated at 100 �C and (SII-b) same fraction after purificationof b-glucans; (SIII-a) aqueous fraction from mature stagewith mature spores isolated at 100 �C and (SIII-b) same frac-tion after purification of b-glucans.

1298

Mizuno T, Hagiwara T, Nakamura T, Ito H, Shimura K,Sumiya T, Asakura A (1990) Antitumor activity and someproperties of water-soluble polysaccharides from ‘‘Himema-tsutake’’, the fruiting body of Agaricus blazei Murill. Agric.Biol. Chem. 54: 2889–2896.

Mol PC, Wessels JGH (1990) Differences in wall structurebetween substrate hyphae and hyphae of fruit-body stipes inAgaricus bisporus. Mycol. Res. 94: 472–479.

Ohno N, Furukawa M, Miura NN, Adachi Y, Motoi M,Yadomae T (2001) Antitumor b-glucan from the culturedfruit body of A. blazei. Biol. Pharm. Bull. 24: 820–828.

Ohno N, Iino K, Takeyama T, Suzuki I, Sato K, Oikawa S,Miyazaki T, Yadomae T (1985) Structural characterizationand antitumor activity of the extracts from matted myceliumof cultured Grifola frondosa. Chem. Pharm. Bull. 33:3395–3401.

Ohno N, Hayashi M, Iino K, Suzuki I, Oikawa S, Sato K,Yadomae T (1986) Effect of glucans on the antitumoractivity of Grifolan. Chem. Pharm. Bull. 34: 2149–2154.

Saito H, Ohki T, Yoshioka Y, Fukuoka F (1976) A 13C nuclearmagnetic resonance study of a gel-forming branched(1 fi 3)-b-D-glucan from Pleurotus ostreatus (fr.): determi-nation of side-chains and conformation of the polymer-chainin relation to gel-structure. FEBS Lett. 68: 15–18.

Wasser SP, Didukh MY, Amazonas MALA, Nevo E, StametsP, Eira AF (2002) Is a widely cultivated culinary-medicinalroyal sun Agaricus (the Himematsutake mushroom) indeedAgaricus blazei Murrill?. Int. J. Med. Mush. 4: 267–290.

York WS (1995) A conformational model for cyclic b-(1 fi 2)-linked glucans based on NMR analysis of the glucansproduced by Xanthomonas campestris. Carbohyd. Res. 278:205–225.

1299

Taxane production in suspension culture of Taxus 3 media var. Hicksiicarried out in flasks and bioreactor

Katarzyna Syklowska-Baranek* & Miroslawa FurmanowaDepartment of Biology and Pharmaceutical Botany, Medical University of Warsaw, ul. Banacha 1, 02--097,Warsaw, Poland*Author for correspondence (E-mail: kasiasb@farm.amwaw.edu.pl)

Received: 6 April 2005; Revisions requested 14 April 2005; Revisions received 20 June 2005; Accepted 21 June 2005

Key words: bioreactor, 10-deacetylbaccatin III, paclitaxel, L-phenylalanine, Taxus · media var. Hicksii

Abstract

Paclitaxel and 10-deacetylbaccatin III (10-DAB III) were produced in suspension cultures of Taxus · mediavar.Hicksii grown in shake-flasks and in a 7-l bioreactor reaching, in the bioreactor, 4.4 mg l)1 (onday 14) and37.5 mg l)1 (on day 11). In shake-flasks the highest total content of paclitaxel and 10-DAB III was 7.3 mg l)1

(on day 4) and 8.8 mg l)1 (on day 18). Phenylalanine, at 0.05 mM, increased paclitaxel accumulation in cellscultivated in bioreactor and flasks 30-fold and 9-fold (from 0.02 mg l)1 to 0.6 mg l)1 and to 0.2 mg l)1,respectively). The 10-DAB III content in cells from flasks was increased from 0.4 mg l)1 to 1.6 mg l)1.

Introduction

The structurally complex taxane diterpenoid,paclitaxel, first isolated from the bark of theTaxus brevifolia, is a highly effective anti-can-cer drug used widely in the treatment of vari-ous carcinomas, melanomas, and sarcomas.The 10-deacetylbaccatin III (10-DAB III) is anintermediate in paclitaxel biosynthesis (Walker& Croteau 2001) and currently paclitaxel andits analogue, docetaxel, are produced semi-syn-thetically through acylation of 10-DAB III iso-lated from needles of various Taxus species.

Taxus-derived cell cultures may be a useful analternative source of paclitaxel and its derivativesas the complete chemical synthesis is uneconomicand, moreover, an increasing demand for newtaxoids with improved biological activity andpossible application against other diseases hasbeen observed. Currently paclitaxel, manufac-tured by plant cell culture technology, is the ac-tive compound of Genexol produced bySamyang Genex (www.genexol.com). Amongmany applied strategies to enhance paclitaxel

accumulation in cell culture the medium supple-mentation with L-phenylalanine, the precursor ofpaclitaxel’s side chain, resulted in considerableincrease in paclitaxel production in suspensioncultures of Taxus species as it was reviewed byZhong (2002). The aim of this work was to con-duct a comparative study of biomass growth andexamined the influence of L-phenylalanine onproduction of paclitaxel and 10-DAB III in sus-pension culture of Taxus · media var. Hicksiicarried out in flasks and 7-l bioreactor. This isthe first report describing the 10-DAB III andpaclitaxel accumulation in suspension cultureperformed in shake flasks and bioreactor.

Materials and methods

Flask cultures

Suspension culture of Taxus · media var. Hicksiiwas initiated from callus of seedling origindifferentiating into roots. Shake-flask cultureswere carried out in 250-ml Erlenmeyer flasks

Biotechnology Letters (2005) 27: 1301--1304 � Springer 2005DOI 10.1007/s10529-005-0223-5

containing 30 ml modified DCR medium (Gupta& Durzan 1985). The medium was supplementedwith 1 mg NAA l)1, 4.8 mg picloram l)1, 500 mgcasein hydrolysate l)1 and 30 g sucrose l)1. Cellswere transferred to the fresh medium every fourweeks.

To perform the comparative studies the cul-ture was continued in the same medium butwith addition of 8.25 mg L-phenylalanine l)1

(0.05 mM). To examine the time growth parame-ters about 2.5 g fresh weight of 28-day old cellswas placed onto 25 ml fresh medium in 250-mlErlenmeyer flasks. Every 3--4 days, samples from2 flasks were collected. The fresh and dry weightof the cells was recorded. The medium was sub-mitted for detection of the pH, content ofsucrose and its conductivity, as well for chemicalanalysis.

Bioreactor cultures

The culture was performed in 7-l mixed-type air-lift reactor implemented with Rushton-type stir-red tank (Biostad Ed, Braun). The workingvolume was 5 l. The inoculum was 2% (w/v)DCR medium as described above. From day 1--4of culture stirring was set at 400 rpm, and next tillthe end of experiment at 200 rpm. The pH valuewas set at 5.42. To determine growth parametersthe procedure mentioned above was used.

The cultures were maintained at 25±2 �Cunder fluorescent light (40 lmol m)2 s)1) and12 h light/dark cycle.

Chemical analysis

The content of paclitaxel (Sigma) and 10-DABIII (donated by Prof. Jaziri from Free Universityof Brussels) in cells and samples of medium wasdeterminated using method presented earlier(Furmanowa & Syklowska-Baranek 2000).

All these experiments were conducted twice.Statistical analysis was performed using the Stat-Soft STATISTICA PL software.

Results and discussion

Paclitaxel and 10-DAB III production duringbatch culture, carried out in bioreactor and inthe Erlenmeyer flasks, were compared.

The growth of biomass in Erlenmeyer flasksand bioreactor was similar up to day 14. A lagphase was not observed. In the culture carriedout in the bioreactor, starting from day 14, therewas a rapid decline of biomass accumulationwhich could be attributed to cell breakage.Moreover from day 4 until the end of cultivationin the bioreactor sucrose was not taken up fromthe medium and remained on unchanged. On theother hand, the consumption of sucrose in flaskscorresponded well to the cell growth profile, andwas 90% exhausted by day 14 -- at the onset ofstationary growth phase. Navia-Osorio et al.(2002) also reported the lack of the lag phaseafter inoculation in experiments conducted in 20-l bioreactor. Data recorded during cultivatingcells in bioreactor might be contributed to highercell damage by sharing as the medium stirringwas set at 400 rpm. The similar growth patternof suspension culture performed in flasks and bi-oreactors was also reported by Pestchanker et al.(1996).

The supplementation of medium with L-phen-ylalanine increased the production of cell-associ-ated paclitaxel both in the bioreactor and flasks(see Tables 1 and 2). The influence of this aminoacid was the most pronounced in cultures per-formed in bioreactor where a 30-fold rise in pac-litaxel accumulation in cells was obtained(Table 2). Jha et al. (1998) reported a 3-fold en-hance of paclitaxel with addition of 2.5 mg IAA-and phenylalanine l)1 to the medium. Fett-Netoet al. (1994) demonstrated 4-fold increase ofcell-associated paclitaxel content but that theprecursor did not promote the accumulation ofpaclitaxel in the medium relative to the control.L-Phenylalanine added to the medium along withother precursors and elicitors at optimized dosesdoubled the paclitaxel content of the cells (Luo& He 2004). A combination of in situ extractionwith organic solvents, precursor feeding andadditional carbon source introduction gave 5times higher paclitaxel production in comparisonto the control (Yuan et al. 2001). Also a 3-foldrise of paclitaxel accumulation in Taxus · mediavar. Hicksii transgenic root culture was foundafter elicitation with methyl jasmonte (Furma-nowa & Syklowska-Baranek 2000).

In our experiments the highest total paclitaxelproduction (cell-associated and extracellular)was higher in flasks (Tables 1 and 2). However,

1302

paclitaxel in cells from flasks was produced onlyup to day 18 and was not detected after day 7 inthe medium (Table 1).

Higher amounts of 10-DAB III were accumu-lated by cells cultivated in flasks than in bioreac-tor. But substantially higher quantities of10-DAB III were secreted to the medium in bio-reactor than in flasks (Tables 1 and 2).

In our experiments the highest total concen-trations of paclitaxel detected in bioreactor coin-cided with the lowest 10-DAB III amounts(Table 2).

In the bioreactor paclitaxel and 10-DAB IIIaccumulation in cells and their secretion to themedium was continuous, moreover at day 11 fourpeaks of unidentified paclitaxel analogues but inconcentration several times higher than paclitaxel,were observed both in cells and medium.

During culturing, the excretion of both tax-anes in cells grown in flasks and in the bioreactorexceeded on average 72% of the total compoundcontent which is consistent with results obtainedby Navia-Osorio et al. (2002) and Pestchankeret al. (1996) although Wickremesinhe & Arteca(1994) reported only a 10% release of paclitaxelin suspension culture of Taxus · media. It wasearlier demonstrated that at low inoculum sizesof 1.5 and 2.0 g dry wt l)1 the extracellular pac-litaxel concentration was relatively higher (Wanget al. 1997) which is in accordance with ourresults.

Precursor feeding strategy employed in ourexperiments seems to be promising for futureimproving and scale-up of taxane production insuspension culture of Taxus species, when 10-DAB III production amounted to 37.5 mg l)1

Table 1. Paclitaxel and 10-DAB III content (mg l)1) in cells and medium containing L-phenylalanine from suspension culture ofTaxus · media var. Hicksii carried out in flasks (values are means of four samples ±SD).

Growth (days) Cells Medium Total

Dry weight (g l)1) Paclitaxel 10-DAB III Paclitaxel 10-DAB III Paclitaxel 10-DAB III

4 14.2±0.890 0.03±0.002 1.0±0.008 7.3±0.052 6.6±0.027 7.3±0.171 7.6±0.810

7 16.1±1.002 0.2±0.003 0.7±0.007 1.1±0.002 0.3±0.002 1.3±0.024 1.0±0.254

11 18.6±2.001 0.04±0.001 1.0±0.008 0 2.6±0.003 0.04±0.002 3.6±0.436

14 22.2±2.124 0.2±0.003 0.8±0.007 0 4.1±0.012 0.2±0.021 4.9±0.564

18 21.5±1.954 0.1±0.001 0.5±0.004 0 8.2±0.044 0.1±0.002 8.8±0.641

21 20.2±1.025 0 0.8±0.005 0 3.7±0.039 0 4.5±0.123

25 18.2±1.047 0 1.6±0.014 0 4.3±0.081 0 5.9±0.921

28 21.8±0.712 0 0.2±0.002 0 4.8±0.056 0 5.0±1.002

Taxane content determined in cells from suspension culture carried out in medium without L-phenylalanine in shake flasks (control):paclitaxel 0.02±0.003 mg l)1, 10-DAB III 0.4±0.004 mg l)1; 0 -- taxanes not detected.

Table 2. Paclitaxel and 10-DAB III content (mg l)1) in cells and medium containing L-phenylalanine from suspension culture ofTaxus · media var. Hicksii carried out in bioreactor (values are means of four samples ± SD).

Growth (days) Cells Medium Total

Dry weight (g l)1) Paclitaxel 10-DAB III Paclitaxel 10-DAB III Paclitaxel 10-DAB III

1 3.5±0.451 0 0 1.2±0.081 4.2±0.065 1.2±0.054 4.2±0.029

4 5.4±0.612 0.04±0.003 0.05±0.007 2.2±0.064 11.5±1.018 2.3±0.036 11.6±0.764

7 5.0±0.398 0.3±0.008 0.07±0.006 1.1±0.032 12.9±1.425 1.4±0.014 13.0±0.0214

11 7.9±0.887 0.6±0.007 0.08±0.005 0 37.4±2.523 0.6±0.052 37.5±2.395

14 18.6±1.014 0.4±0.010 0.04±0.002 4.0±0.076 9.3±0.874 4.4±0.321 9.9±0.845

18 15.2±2.007 0 0 0 0 0 0

21 11.9±1.985 0 0 0 0 0 0

Taxane content determined in cells from suspension culture carried out in medium without L-phenylalanine in shake flasks (control):paclitaxel 0.02±0.003 mg l)1, 10-DAB III 0.4±0.004 mg l)1; 0 -- taxanes not detected.

1303

within 11 days of culture and paclitaxel yielded4.4 mg l)1 within 14 days of culture.

Acknowledgements

This research work was supported by the grantfrom the State Committee for Scientific ResearchNo. 4 P05F 028 12. We are grateful to ProfessorMondher Jaziri from the Laboratory of Biotech-nology and Plant Morphology, Free University ofBrussels for the 10-deacetylbaccatin III standard.

References

Fett-Neto AG, Melanson SJ, Nicholson SA, Pennington JJ,DiCosmo F (1994) Improved taxol yield by aromaticcarboxylic acid and amino acid feeding to cell cultures ofTaxus cuspidata. Biotechnol. Bioeng. 44: 967--971.

Furmanowa M, Syklowska-Baranek K (2000) Hairy rootcultures of Taxus · media var. Hicksii Rehd. as a newsource of paclitaxel and 10-deacetylbaccatin III. Biotechnol.Lett. 22: 683--686.

Gupta PK, Durzan DJ (1985) Shoot multiplication frommature trees of Douglas-fir (Pseudotsuga menziesii) andSugar pine (Pinus lambertiana). Plant Cell Rep. 4: 177--179.

Jha S, Sanyal D, Ghosh B, Jha TB (1998) Improved taxol yieldin cell suspension culture of Taxus wallichiana (HimalayanYew). Planta Med. 64: 270--272.

Luo J, He GY (2004) Optimization of elicitors and precursorsfor paclitaxel production in cell suspension culture of Taxuschinensis in the presence of nutrient feeding. ProcessBiochem. 39: 1073--1079.

Navia-Osorio A, Garden H, Cusido RM, Palazon J, AlfermanAW, Pinol TM (2002) Production of paclitaxel and baccatinIII in a 20-L airlift bioreactor by a cell suspension of Taxuswallichiana. Planta Med. 68: 336--340.

Pestchanker LJ, Roberts SC, Shuler ML (1996) Kinetics oftaxol production and nutrient use in suspension cultures ofTaxus cuspidata in shake flasks and a Wilson-type bioreac-tor. Enzyme Microb. Technol. 19: 256--260.

Walker K, Croteau R (2001) Taxol biosynthesis genes. Phyto-chemistry 58: 1--7.

Wang HQ, Zhong JJ, Yu JT (1997) Enhanced production oftaxol in suspension cultures of Taxus chinensis by controllinginoculum size. Biotechnol. Lett. 19: 353--355.

Wickremesinhe ERM, Arteca RN (1994) Taxus cell suspensioncultures: optimizing growth and production of taxol. J. PlantPhysiol. 144: 183--188.

Yuan YJ, Wei ZJ, Wu ZL, Wu JC (2001) Improved taxolproduction in suspension cultures of Taxus chinensis var.mairei by in situ extraction combined with precursor feedingand additional carbon source introduction in airlift loopreactor. Biotechnol. Lett. 23: 1659--1662.

Zhong JJ (2002) Plant cell culture form production of paclitaxeland other taxanes. J. Biosci. Bioeng. 94: 591--599.

1304

Regio- and stereo-selective hydroxylation of abietic acid derivativesby Mucor circinelloides and Mortierella isabellina

Koichi Mitsukura, Takeshi Imoto, Hirokazu Nagaoka, Toyokazu Yoshida &Toru Nagasawa*Department of Biomolecular Science, Gifu University, 1-1 Yanagido, Gifu, 501-1193, Japan*Author for correspondence (Fax: +81-58-293-2647; E-mail: tonagasa@biomol.gifu-u.ac.jp)

Received 27 April 2005; Revisions requested 24 May 2005; Revisions received 20 June 2005; Accepted 21 June 2005

Key words: abietic acid derivatives, microbial hydroxylation, regio-selective, stereo-selective

Abstract

Mucor circinelloides and Mortierella isabellina hydroxylated dehydroabietic acid (DehA). DehA was con-verted regio- and stereo-selectively by whole cells of Mr. circinelloides to give 2a-hydroxydehydroabieticacid in a 75% molar conversion yield (11 mM from 14.7 mM DehA) after 72 h in the cultivation mediumcontaining 3% (v/v) Tween 80. With cells of Ma. isabellina, under the same conditions, 20.5 mM (6.5 g l)1)2–hydroxydehydroabietic acid (a/b=81/19) was formed from 26.4 mM DehA.

Introduction

Rosin, commonly known as resin acid, is anabundant renewable resource obtained frompitch pine. It is comprised of several diterpe-noic acid derivatives, such as abietic acid,dehydroabietic acid, dihydroabietic acid, pima-ric acid and tetrahydroabietic acid. They arewidely used to synthesize sizing agents forpaper, emulsifying agents for synthetic rubber,resin for printing inks, resin adhesives, etc.(Sadhra et al. 1994). On the other hand, theleakage of the resin acid into natural worldthrough the production process of paper andpulp has often caused severe environmentalpollution problems (Owens 1991). Therefore,the microbial degradation or detoxification ofresin acid has been studied (Liss et al. 1997,Martin et al. 1999).

Further broad applications of resin acids canbe expected in the field of synthetic chemicalsdue to their unique chemical and physicalproperties. Several biological activities such asanti-ulcer (Wada et al. 1985), anti-microbial(Savluchinske Feio et al. 1999) and anti-inflammatory

(Fernandez et al. 2001) effects of resin derivativeshave also been reported. In the present study, wehave surveyed microorganisms to catalyze thehydroxylation of abietic acid derivatives andfound a strain of Mucor circinelloides IT25 thatcatalyzes the regio-selective and stereo-selectivehydroxylation of both abietic acid and dehydroa-bietic acid.

Materials and methods

Chemicals

Abietic acid (AbA), dehydroabietic acid (DehA),dihydroabietic acid (DihA) and rosin contain-ing 2% (w/w) AbA, 58% (w/w) DehA and 36%(w/w) DihA were kindly provided by ArakawaChemical Industries (Japan). Products from the fol-lowing suppliers were used: polypeptone (NipponSeiyaku, Japan), meat extract (Mikunikagaku,Japan) and yeast extract (Oriental Yeast, Japan).All other chemicals were of guaranteed reagentgrade.

Biotechnology Letters (2005) 27: 1305–1310 � Springer 2005DOI 10.1007/s10529-005-3224-5

Isolation of microorganisms capable of convertingabietic acid derivatives

Microorganisms capable of converting AbA,DehA or DihA were isolated in two ways usingtwo types of basal media. To collect fungi fromsoil, a mixture of soil samples (approximately2 g) suspended in 5 ml 0.85% (w/v) NaClwas plated onto medium A containing 30 gsucrose l)1, 2 g NaNO l)13, 1 g K2HPO4 l

)1,0.5 g MgSO4Æ7H2O l)1, 0.5 g KCl l)1, 0.01 gFeSO4Æ7H2O l)1 and 20 g agar l)1and incubatedat 28 �C. A single colony on each plate waspicked, and its AbA or DehA-converting activitywas evaluated. Alternatively, a conventionalenrichment culture was carried out aerobicallyat 28 �C for 14 days in a test-tube containing5 ml medium B (pH 7.3) which consistedof 4 g rosin l)1, 10 g sucrose l)1, 1 g yeastextract l)1, 2 g NH4Cl l)1, 1 g K2HPO4 l

)1,0.5 g MgSO4Æ7H2O l)1, 0.5 g KCl l)1and 0.01 gFeSO4Æ7H2O l)1. Each culture was spread on theplate containing the medium A and 2% (w/v)agar and incubated at 28 �C. The conversion ofrosin was checked by thin-layer chromatography(TLC) analysis.

Identification of microorganisms

Among 238 isolates, three bacteria (HR1, HR6,and HR34) and two molds (IT 25 and HR32)exhibited abietic acid derivative-converting activ-ity. Based on morphological, physiological andbiological aspects, these five strains were identi-fied by the National Collections of Industrial,Food and Marine Bacteria, Japan.

Culture conditions and conversion of AbA,DehA and DihA

The pre-culture was carried out at 28 �C for2 days with reciprocal shaking in a test-tube con-taining 5 ml medium (pH 5.5) comprised of5 g polypepton l)1, 2 g yeast extract l)1, 1 gK2HPO4 l

)1, 0.5 g MgSO4Æ7H2O l)1 and 0.01 gFeSO4Æ7H2O l)1. The pre-culture was added to a500 ml shaking-flask containing 30 ml of thesame medium using pre-culture. Cultivationwas carried out at 28 �C for 2 days withreciprocal shaking (115 strokes min)1), and then

0.2% (w/v) AbA, DehA or DihA were added tothe medium. Biotransformation of each substratewas carried out at 28 �C for 7 days.

Identification of reaction products

The reaction products (see Figure 1) convertedfrom AbA, DehA or DihA were purified usinga silica gel column (n-hexane:ethyl acetate=8:1,4/1, 2:1, 1:1, 0:1, and methanol, v/v) and pre-parative TLC (n-hexane:ethyl acetate=2:1, v/v).The analyses of products were carried out usinga Varian INOVA 400 for NMR, a Jeol Auto-mass SUN300 for GC-MS, and a Horiba SEPA-300 for optical rotation.

Fig. 1. Microbial conversion of various abietic acid deriva-tives.

1306

High performance liquid chromatography(HPLC) analysis

Each compound was analyzed by reverse-phaseHPLC using an ODS column (Waters Spherisorb,4.6�150 mm) and methanol/water (8:2, v/v) forelution at 1 ml min)1 and monitoring at 230 nm.

Optimization of culture medium

For the studies on the optimization of culturemedium, a mixture comprised of 0.1% (w/v)K2HPO4, 0.05% (w/v) MgSO4Æ7H2O, and 0.001%(w/v) FeSO4Æ7H2O was used as the basal medium.The pre-culture using medium A was carried outat 28 �C for 2 days with reciprocal shaking. Culti-vation was carried out at 28 �C for 3 days withreciprocal shaking.

Results

Microbial conversion of AbA, DehA and DihA

Among fungi isolates, Mortierella isabellinaHR32 exhibited a powerful hydroxylationactivity for DehA. In addition, we found thatMr. circinelloides IT 25 exhibited a novel DehA-converting activity. Mr. circinelloides IT25 cata-lyzed the regio- and stereo-selective hydroxylationof AbA and DehA to 2a-hydroxyabietic acid and2a-hydroxydehydroabietic acid, respectively.With AbA as a substrate, the aromatizationof AbA to DehA proceeded in the course ofcultivation medium due to its AbA instability.Although Mr. javanicus IAM 6087 also exhibitedDehA-hydroxylation activity, it was lower thanthat of Mr. circinelloides IT25.

Ma. isabellina HR32 also showed the hydrox-ylation activity for DehA or DihA. 2-hydroxyde-hydroabietic acid formed from DehA was amixture of two isomers, i.e., 81% a-isomer and19% b-isomer, as revealed in an analysis ofNMR after esterification with trimethylsilyldia-zomethane in a methanol solution. When usingAbA as a substrate, however, 2a-hydroxyabieticacid was not formed at all. If the other isolateswere examined, Moraxella sp. HR6 convertedDehA and DihA to 3,7-dioxodehydroabietin and3,7-dioxodihydroabietin, respectively, probably

through the oxidation at C3 and C7 followed bydecarboxylation at C4 with a low molar conver-sion yield. Pseudomonas sp. HR34 convertedDihA to 7-oxodihydroabietic acid. On the otherhand, Sphingomonas sp. HR1 catalyzed the rapiddegradation of DehA or DihA, so that theirhydroxylated products were not detected. Thesemicrobial conversions of abietic acid derivativeswe found are summarized in Figure 1.

The product obtained from AbA, DehA orDihA by microbial conversion was purified by sil-ica gel column and identified by NMR and MS.

2a-Hydroxydehydroabietic acid: 1H-NMR(CDCl3, 400 MHz) d 7.18 (1H, d, J=8.0 Hz), 7.02(1H, dd, J=8.0, 1.6 Hz), 6.91 (1H, d, J=1.6 Hz),4.11 (1H, tt, J=11.4, 4.1 Hz), 2.88–2.96 (2H, m),2.82 (1H, septet, J=7.0 Hz), 2.67 (1H, m), 2,25(1H, dd, J=12.6, 2.0 Hz), 2.12(1H, s), 2.07 (1H,m), 1.78–1.90 (1H, m), 1.77 (1H, t, J=11.7 Hz),1.57–1.65 (1H, m), 1.48 (1H, t, J=11.7 Hz), 1.31(3H, s), 1.28 (3H, s), 1.22 (6H, d, J=7.0 Hz); 13C-NMR (CDCl3, 100 MHz) d182.8, 146.1, 145.8,134.2, 127.0, 124.1, 123.9, 65.1, 48.2, 47.1, 45.2,44.3, 38.6, 33.5, 29.8, 26.1, 24.0, 20.9, 17.3; GC-MS (methyl ester) (m/z) 330 (M+), 312, 283, 255,237(100%), 195; [a]24.5D +60.4 (c. 0.16, CHCl3).In the case of Ma. isabellina HR32, the opticalrotation is [a]24.5D +44 (c. 0.185, CHCl3).

2a-Hydroxyabietic acid: 1H-NMR (CDCl3,400 MHz) d 5.78 (1H, s), 5.37 (1H, d,J=4.4 Hz), 3.94 (1H, tt, J=11.6, 4.0 Hz),2.21 (1H, dd, J=13.6, 6.2 Hz), 2.10 (1H, brs),1.92–2.08 (3H, m), 1.83 (1H, d, J=11.7 Hz), 1.78(1H, t, J=11.7 Hz), 1.27 (3H, s), 1.21–1.26 (5H,m), 1.13 (1H, t, J=12.0 Hz), 1.10 (3H, d,J=7.0 Hz), 1.00 (1H, t, J=6.6 Hz), 0.86 (3H s,);13C-NMR (CDCl3, 100 MHz) d182.9, 145.4,135.1, 122.1, 120.3, 64.6, 50.9, 47.5, 45.3, 44.5,36.3, 34.8, 27.3, 25.1, 22.5, 21.4, 20.8, 17.7, 14.9;GC-MS (methyl ester) (m/z) 332 (M+), 314,255(100%), 239.

2-Hydroxydihydroabietic acid: 1H-NMR(CDCl3, 400 MHz) d 3.97 (1H, tt, J=11.4,4.2 Hz), 2.11–2.17 (1H, m), 2.04–2.11 (1H, m),2.03 (1H, dd, J=13.2, 2.6 Hz), 1.69–1.83(2H, m), 1.53–1.63 (2H, m), 1.19 (2H, t,J=8.4 Hz), 1.14 (2H, t, J=11.7 Hz), 1.02 (3H,s), 0.95–1.01 (1H, m), 0.88 (3H, d, J=7.0 Hz),0.87 (3H, d, J=7.0 Hz); 13C-NMR (CDCl3,100 MHz) d182.6, 136.8, 126.3, 65.1, 48.3, 45.9,45.3, 44.8, 40.4, 38.6, 34.3, 32.5, 31.8, 27.1, 24.7,

1307

20.8, 20.4, 19.9, 19.5, 17.3; GC-MS (methyl ester)(m/z) 334 (M+), 316, 301, 273, 257, 241(100%).

7-Oxodihydroabietic acid: 13C-NMR (CDCl3,100 MHz) d200.0, 182.3, 166.5, 130.6, 46.3, 44.6,39.5, 39.2, 36.7, 36.2, 34.4, 32.3, 26.7, 26.2, 26.0,19.8, 19.5, 17.8, 17.7, 16.2; GC-MS (methyl ester)(m/z) 332 (M+), 317, 289, 273, 257(100%), 229.

3,7-Dioxodehydroabietin: 13C-NMR (CDCl3,100 MHz) d210.7, 197.4, 149.8, 147.6, 132.7,130.8, 125.4, 124.4, 47.8, 44.9, 39.4, 37.7, 37.2,36.7, 33.6, 23.8, 23.7, 21.2, 11.0; GC-MS (m/z)284 (M+), 269(100%), 227, 199.

3,7-Dioxodihydroabietin: 13C-NMR (CDCl3,100 MHz) d210.7, 197.8, 163.5, 131.9, 48.0, 44.7,39.4, 38.7, 38.3, 37.5, 35.2, 32.3, 27.4, 26.8, 25.9,19.7, 19.5, 16.0, 10.9.

Optimal conditions for accumulationof 2a-hydroxydehydroabietic acid

To enhance the hydroxylation activity of Mr. cir-cinelloides IT25 and Ma. isabellina HR32, theculture conditions were optimized. The activ-ity was evaluated by measuring the amount of2a-hydroxydehydroabietic acid formed in themedium at 28�C after 72 h cultivation. The opti-mal medium for Mr. circinelloides IT25 consistedof 20 g sodium L-glutamate l)1, 10 g maltextract l)1, 1 g K2HPO4 l

)1, 0.5 g MgSO4Æ7-H2O l)1, and 0.01 g FeSO4Æ7H2O l)1 at pH 7.5.Under these conditions, it produced 5.4 mM 2a-hydroxydehydroabietic acid from 14.7 mM DehAwith a 37% molar conversion yield.

Culture conditions of Ma. isabellina HR32were also examined. The optimized medium wascomprised of 20 g polypeptone l)1, 20 g yeastextract l)1, 1 g K2HPO4 l

)1, 0.5 g MgSO4Æ7-H2O l)1 and 0.01 g FeSO4Æ7H2O l)1 at pH 5.0.When the cultivaion was carried out using themedium, 5.6 mM 2-hydroxydehydroabietic acidwas formed from 26.4 mM DehA with a 21%molar conversion yield.

For further improvement of the productivityof 2a-hydroxydehydroabietic acid, the effect oforganic solvents and detergents on DehA hydrox-ylation was examined (Table 1). The addition of1% (v/v) methanol or ethanol was effective inslightly extending of the formation of 2a-hydrox-ydehydroabietic acid by Ma. isabellina HR32. Theadditon of 3% (v/v) Tween 80 resulted in a signifi-cant enhancement of 2a-hydroxydehydroabietic

acid production by Mr. circinelloides IT 25 or Ma.isabellina HR32. Mr. circinelloides IT 25 produced11 mM (3.5 g l)1) 2a-hydroxydehydroabietic acidfrom 14.7 mM DehA with a 75% molar conver-sion yield. In the case of Ma. isabellina HR32,20.5 mM (6.5 g l)1) 2-hydroxydehydroabietic acid(a/b=81/19) was produced from 26.4 mM DehAwith a 78% molar conversion yield (Table 1).

Effects of P450 inhibitors on DehA hydroxylation

The effect of common P450 inhibitors on thehydroxylation of DehA by Mr. circinelloides IT25 was examined by adding 0.5 mM P450 inhibi-tors to the cultivation medium (Table 2). Thehydroxylation activity of Mr. circinelloides IT 25was inhibited by all P450 inhibitors tested, with1-aminobenzotriazole causing the most significantinhibition.

Discussion

In previous studies of biodegradation of DehA,some bacteria could grow on DehA as a carbonsource, and the biodegradation pathways of

Table 1. Effects of organic solvents and detergents on pro-duction of 2-hydroxydehydroabietic acid.

Organic

solvent or

detergent

Cone

(% v/v)

Relative

activity (%)

Mucor

circinelloides

Mortierella

isabellina

None 100 100

Methanol 1 92 149

Ethanol 1 92 145

Acetone 1 93 90

Dimethyl sulfoxide 1 55 96

Tween 20 1 98 124

Tween 80 1 128 195

2 166 353

3 204 365

4 204 360

Triton X)100 1 96 0

DehA (150 mg, 0.50 mmol) and (270 mg, 0.90 mmol), respec-tively, was added to the cultivation medium of either Mr. cir-cinelloides orMa. isabellina. Strain IT25 and HR32 cells formed5.4 mM and 5.6 mM 2-hydroxydehydroabietic acid, respectively,without organic solvents and detergents after 72 h cultivation at28 �C. These values were taken as 100%.

1308

DehA have been outlined as shown in Figure 2abased on studies of their metabolic intermediatesor molecular genetics. The microbial degradationor conversion of AbA has scarcely been studieddue to its chemical lability, causing it to changereadily into DehA (Martin & Mohn 2000).

In the present study, we found that Moraxel-la sp. HR6 acted on DehA, and forming a newmetabolite, 3,7-dioxodehydroabietin. This resultsuggested that Moraxella sp. HR6 probablydegraded DehA through a pathway similar topathway A in Figure 2a. A low amount ofaccumulated of 3,7-dioxodehydroabietin sug-gested that the activity of the dioxygenaserequired to catalyze cleavage of the subsequentrings might be high. When DehA was replacedwith DihA, the accumulation of a new metabolite,3,7-dioxodihydroabietin was observed at ahigher concentration than that of 3,7-dioxode-hydroabietin. Thus, the degradation of 3,7-diox-odihydroabietin proceeded at lower rate than3,7-dioxodehydroabietin because the cyclohexanering of DihA is resistant to the meta-cleavagecatalyzed by dioxygenase. Pseudomonas sp.HR34 also catabolized DihA rapidly and a newmetabolite, 7-oxodihydroabietic acid, wasdetected. This suggests that the degradation ofDehA by Pseudomonas sp. HR34 probably

Table 2. Effect of P450 inhibitors at 0.5 mM on DhA hydrox-ylation by Mr. circinelloides IT25.

Inhibitor Relative activity (%)

None 100

Methoxsalen 49

1-Aminobenzotriazole 36

Ketoconazole 52

Miconazole 75

Menadione 76

Hydroxylation of DhA was carried out at 28 �C for 72 h cul-tivation in the optimum medium containing 3% (v/v) Tween 80.The formation of 10 mM 2a-hydroxydehydroabietic acid in theabsence of P450 inhibitors was taken as 100% – see Table 1 forabsolute values.

Fig. 2. (a) Possible degradation pathway of DehA by Flavobacterium resinovorum (Biellmann et al. 1973a) (A), Alcaligenes eutro-phus and Pseudomonas sp. (Biellmann et al. 1973b) (B), Pseudomonas abietaniphila BKME)9 (Martin and Morn 1999) (b) Conver-sion of DehA by Fusarium oxysporum or F. moniliforme (Tapia et al. 1997) (C), Mortierella isabellina (Kutney et al. 1981) (D) andChaetomium cochliodes (Yano et al. 1994) (E).

1309

proceeded through pathway B (see Figure 2).When Sphingomonas sp. HR1 cells were used,DehA was completely degraded.

In contrast to bacteria, no mold capable ofassimilating DehA as a sole carbon source hasbeen reported. As shown in Figure 2b, twomolds convert DehA into the hydroxylated ca-tabolites. In our studies, Ma. isabellina HR32produced a mixture of 2a- and 2b-hydroxydehy-droabietic acid through pathway D. We alsofound that Ma. isabellina HR32 catalyzed thehydroxylation of DihA at C2 and Mr. circinello-ides IT25 catalyzed the hydroxylation of AbA atC2. We isolated these metabolites and identifiedtheir chemical structures.

The addition of nonionic detergent Tween 20or Tween 80 enhanced production of 2a-hydrox-ydehydroabietic acid. We assume that Tween 20and Tween 80 probably enhanced the solubilityof rosin or the permeability of the cell membrane(Baklashova & Koshcheenko 1980).

Based on the effect of P450 inhibitors onDehA hydroxylation, we suggested that Mr. cir-cinelloides IT25 might proceed through a systemincorporating P450.

Conclusion

Until now, the production of catabolic intermedi-ates of DehA, a renewable natural resource, hasnot been undertaken from the viewpoint of theirapplication and development. In the presentstudies, we have produced, for the first time, alarge amount of 2a-hydroxydehydroabietic acidby optimizing the culture conditions and are pro-ceeding to examine its applications as a func-tional material.

Acknowledgements

We are grateful to Life Science Research Centerin Gifu University for NMR and mass measure-ments.

References

Baklashova TG, Koshcheenko KA (1980) Effect of detergentson the hydroxylation of indolyl)3-acetic acid by an Asper-gillus niger culture. Mikrobiologiia 49: 546–550.

Biellmann JF, Branlant G, Gero-Robert M, Poiret M (1973a)Degradation bacterienne de l’acide dehydroabietique par unFlavobacterium resinovorum. Tetrahedron 29: 1227–1236.

Biellmann JF, Branlant G, Gero-Robert M, Poiret M (1973b)Degradation bacterienne de l’acide dehydroabietique parun Pseudomonas et une Alcaligenes. Tetrahedron 29:1237–1241.

Fernandez MA, Tornos MP, Garcia MD, de las Heras B, VillarAM, Saenz MT (2001) Anti-inflammatory activity of abieticacid, a diterpene isolated from Pimenta racemosa var.grissea. J. Pharm. Pharmacol. 53: 867–872.

Kutney JP, Singh M, Hewitt GM, Salisbury PJ, Worth BR,Servizi JA, Martens DW, Gordon RW (1981) Studies relatedto biological detoxification of kraft pulp mill effluent I. Thebiodegradation of dehydroabietic acid with Mortierellaisabellina. Can. J. Chem. 59: 2334–2341.

Liss SN, Bicho PA, Saddler JN (1997) Mini-review: Microbi-ology and biodegradation of resin acids in pulp mill effluents.Can. J. Microbiol 43: 599–611.

Martin VJJ, Mohn WW (1999) A novel aromatic-ring-hydrox-ylating dioxygenase from the diterpenoid-degrading bacte-rium, Pseudomonas abietaniphila BKME)9. J. Bacteriol.181: 2675–2682.

Martin VJJ, Yu Z, Mohn WW (1999) Mini review: Recentadvances in understanding resin acid biodegradation:microbial diversity and metabolism. Arch. Microbiol. 172:131–138.

MartinVJJ,MohnWW(2000)Genetic investigation of the catabolicpathway for degradation of abietic diterpenoids by Pseudomonasabietaniphila BKME)9. J. Bacteriol. 182: 3784–3983.

Owens JW (1991) The hazard assessment of pulp and papereffluents in the aquatic environment: a review. Environ.Toxicol. Chem. 10: 1511–1540.

Sadhra S, Foulds IS, Gray CN, Koh D, Gardiner K (1994)Colophony – uses, health effects, airborne measurements andanalysis. Ann. Occup. Hyg. 38: 385–396.

Savluchinske Feio S, Gigante B, Roseiro JC, Marcelo-CurtoMJ (1999) Antimicrobial activity of diterpene resin acidderivatives. J. Microbiol. Methods 35: 201–206.

Tapia AA, Vallejo MD, Gouiric SC, Feresin GE, RossomandoPC, Bustos DA (1997) Hydroxylation of dehydroabietic acidby Fusarium species. Phytochemistry 46: 131–133.

Wada H, Kodato S, Kawamori M, Morikawa T, Nakai H,Takeda M, Saito S, Onoda Y, Tamaki H (1985) Antiulceractivity of dehydroabietic acid derivatives. Chem. Pharm.Bull. (Tokyo) 33: 1472–1487.

Yano S, Nakamura T, Uehara T, Furuno T, Takahashi A(1994) Biotransformation of terpenoids in conifers bymicroorganisms I. Hydroxylation of dehydroabietic acidby Chaetomium cochliodes. Mokuzai Gakkaishi 40:1226–1232.

1310

Increased conformational and thermal stability propertiesfor phenylalanine dehydrogenase by chemical glycosidation with end-groupactivated dextran

Reynaldo Villalonga1,*, Shinjiro Tachibana2, Yunel Perez1 & Yasuhisa Asano21Enzyme Technology Group, Center for Biotechnological Studies, University of Matanzas, 44740, Matanzas,C.P, Cuba2Biotechnology Research Center, Toyama Prefectural University, 5180 Kurokawa, 939-0398, Kosugi,Toyama, Japan*Author for correspondence (Fax: +53-45-253101; E-mail: reynaldo.villalonga@umcc.cu)

Received 20 May 2005; Revisions requested 26 May 2005; Revisions received 20 June 2005; Accepted 21 June 2005

Key words: dextran, enzyme stability, glycosidation, phenylalanine dehydrogenase

Abstract

A mono-aminated dextran derivative was attached to Bacillus badius phenylalanine dehydrogenase via acarbodiimide-catalyzed reaction. The optimum temperature for the conjugate was 10 �C higher than fornative enzyme, and its thermostability was improved by 8 �C. The activation free energy of thermalinactivation at 45 �C was increased by 16.8 kJ/mol. The improved conformational stability of the modifiedenzyme was confirmed by fluorescence spectroscopy.

Introduction

Phenylalanine dehydrogenase (PheDH, EC1.4.1.20) from Bacillus badius is a NAD+-depen-dent octameric enzyme that catalyzes thereversible oxidation–reduction reactions for L-phenylalanine (Asano 1999). This enzyme hasbeen used in the colorimetric screening of phen-ylketonuria in neonates in Japan (Asano et al.1987), and also constitutes a valuable catalyst forthe enantioselective synthesis of Phe and relatedL-amino acids from their keto analogs (Asanoet al. 1990). However, the biocatalytic and ana-lytical applications of this enzyme is limited byits rapid inactivation at elevated temperatures(Asano et al. 1987).

Cross-linking of enzymes with polyactivatedpolymers has been widely used for increasingfunctional stability (Srivastava 1991, Gomez &Villalonga 2000, Darias & Villalonga 2001,Villalonga et al. 2003). Site-specific modification

with polyethylenglycols has been also reported asstabilizing method for enzymes (Veronese et al.2002). We recently described the use of mono-activated cyclodextrin derivatives as glycosidationagents for preparing thermostable neoglycoen-zymes (Cao et al. 2003, Villalonga et al. 2003,Fernandez et al. 2004). However, chemically acti-vated cyclodextrins are expensive materials, andthis fact limits their wide use in the synthesis ofneo-glycoenzymes for industrial application.

Dextrans are less expensive and non-toxicpolysaccharides produced by bacteria from su-crose, and consisting of linear a-1,6-linked D-glu-copyranose units with some degree of branchingvia 1,3-linkages (Mehvar 2000). High stable en-zyme derivatives have been prepared by cross-linking the protein surfaces with polyactivateddextrans (Srivastava 1991), but this approachyields to enzyme preparations with low catalyticactivity. This problem could be solved by usingmono-activated polymer derivatives, but the

Biotechnology Letters (2005) 27: 1311–1317 � Springer 2005DOI 10.1007/s10529-005-3225-4

preparation of such kind of mono-activated mac-romolecules requires well specific reaction condi-tions as well as the adequate selection of thepolymer size (Bruneel & Schacht 1995).

The present paper reports the preparation ofan end-group mono-aminated dextran derivative,and its use as glycosidation agent for B. badiusPheDH. The influence of this modification on thecatalytic and stability properties of this oxidore-ductase is evaluated.

Materials and methods

Materials

Phenylalanine dehydrogenase from Bacillus badi-us (18.4 U/mg), recombinantly expressed inE. coli, was prepared as previously described(Asano et al. 1987). L-Phenylalanine, NAD+ and1-ethyl-3-(3-dimethylaminopropyl) carbodiimidehydrochloride (EDAC) were purchased fromWako Pure Chemicals. Dextran 5000 was fromServa. All other chemicals were analytical grade.

Synthesis of end-group aminated dextran

Dextran (2 g), dissolved in 10 ml distilled waterwas treated with 1 ml 1,6-hexylenediamine andstirred for 2 h. NaBH3CN, 150 mg, was then ad-ded and the reaction mixture was continuouslystirring at room temperature overnight. The solu-tion was further extensively dialyzed vs. distilledwater using a Spectrapor 6 dialysis tubing (Serva,molecular weight cut-off 1000 Da) and finallylyophilized. The aminated dextran derivative wascharacterized by 1H-NMR spectrometry using aBruker AVANTE 400 MHz apparatus.

Preparation of PheDH-dextran conjugate

EDAC, 10 mg, was added to a reaction mixturecontaining 4 mg PheDH dissolved in 3 ml 50 mM

sodium phosphate buffer, pH 6.0, and 100 mgaminated dextran. The solution was stirred for1 h at room temperature, then at 4 �C for 16 hand finally dialyzed at 4 �C against 10 mM potas-sium phosphate buffer, pH 7.0, containing 1 mM

EDTA and 5 mM 2-mercaptoethanol.

Analytical determinations

The enzymatic activity of native and modifiedPheDH was determined at 25 �C in 100 mM

glycine/KCl/KOH buffer, pH 10.4, containing2.5 mM NAD+ and using 10 mM L-Phe as sub-strate (Asano et al. 1987). One unit of L-phenyl-alanine dehydrogenase activity is defined as theamount of enzyme that catalyzes the formationof 1 lmol NADH per min under the describedconditions. Michaelis-Menten parameters werecalculated from Eadie-Hofstee plots. Protein con-centration was estimated from the absorbance at280 nm using the absorption coefficientA1%

1cm ¼ 6:3 (Asano et al. 1987). Total carbohy-drates were determined by the phenol/sulfuricacid method using glucose as standard (Duboiset al. 1956).

The molecular weight of the enzyme formswas determined by analytical GPC on TSKGELG3000SW column (4.5�60 cm), calibrated withprotein standards from Oriental Yeast Co., Ltd.

The fluorescence emission spectra of nativeand modified PheDH before and after thermaldenaturation at 55 �C were measured with0.7 nmol enzyme in 10 mM sodium phosphatebuffer, pH 7.0, containing 1 mM EDTA and5 mM 2-mercaptoethanol, using a spectrofluorim-eter with excitation at 280 nm and the emissionscanned between 300–400 nm.

Light scattering measurements of native anddextran-modified PheDH were monitored at400 nm after excitation at 280 nm.

Results and discussion

The strategy used for end-group functionaliza-tion of dextran involves the treatment with a mo-lar excess of 1,6-hexylenediamine in the presenceof NaBH3CN in order to reduce only the newimine bonds formed at the reducing end of thepolymer. Through this procedure, a high yield ofmono-activated dextran was obtained (about91%), as determined by 1H-NMR spectra (datanot shown).

The amino dextran derivative synthesized wasfurther attached to the free carboxylate groupsfrom aspartic and glutamic acid residues locatedat the protein surface of PheDH, through the

1312

formation of stable amide links by using a watersoluble carbodiimide as coupling agent. Due tothe mono-activated nature of the modifying poly-mer, each mol of polysaccharide was attached toa single amino acid residue. The overall syntheticprocess employed for preparing this neo-glycoen-zyme is illustrated in Figure 1.

The structural and catalytic properties of thisenzyme-polymer conjugate are reported inTable 1. The molecular weight of PheDH, deter-mined by analytical GPC, was increased in about16.4 kDa after glycosidation with dextran. Thisresult represents an average of 3 mol polysaccha-ride attached to each mol octameric protein. Onthe other hand, the specific dehydrogenase activityretained by the conjugated enzyme was estimatedto be about 89%. It was further demonstratedthat this reduction was partially mediated by thepresence of EDAC as catalyst in the reaction med-ia. Steric hindrance to the diffusion of L-Phe tothe active site of the enzyme, caused by the pres-ence of the bulky polysaccharide moieties at thesurface of PheDH, could be also considered as apossible cause of this reduction. Glycosidationwith dextran enhanced the affinity of the enzymefor L-Phe: Km was decreased by 1.3-fold for Phe-

Fig. 1. Preparation of PheDH-dextran conjugate.

Table 1. Structural and catalytic properties of dextran-modified PheDH.

Parameter PheDH PheDH-dextran

Molecular weight (kDa) 325.6 342

Dextran content

(mol/mol protein)

– 3

Specific activity (Umg)1) 18.4 16.3

Km (lM) 215 164

kcat (s)1) 816 714

kcat/Km (lM)1s)1) 3.8 4.4

1313

DH after modification with dextran. On the con-trary, the rate of dehydrogenation of L-Phe wasonly slightly decreased for the transformed en-zyme, as is revealed by the lower value of kcat.

The catalytic behaviour of PheDH at hightemperatures was noticeably improved after gly-cosidation with the polymer. In order to evaluatethe effect of this modification on the enzymethermoresistance, different types of experimentswere performed. As a, native PheDH was mixedwith mono-aminated dextran in order to elimi-nate any contribution caused by the physicalpresence of the polysaccharide.

Figure 2 shows the temperature-activity pro-files of native and dextran-modified PheDHpreparations. PheDH showed maximum catalyticactivity at 40 �C, and this value of optimum tem-perature was not affected by the physical pres-ence of the aminated dextran. On the contrary,the optimum temperature for dehydrogenation ofL-Phe was increased in about 10 �C for the en-zyme after conjugation with the polysaccharide.This fact could be justified by the improved ther-mostabilization showed by the conjugated en-zyme (see below), avoiding protein denaturationuntil higher temperatures and then favouring toreach a high catalytic activity of PheDH atelevated temperatures.

The effect of 10 min of incubation at differenttemperatures on the catalytic activity of bothPheDH forms is shown in Figure 3. Dextran-

modified enzyme was more resistant to heattreatment at temperatures higher than 45 �C, incomparison with the native counterpart.Consequently, the value of T50, defined as thetemperature at which 50% of the initial activ-ity was retained, was increased from 54 to62 �C for PheDH after glycosidation with thepolysaccharide.

Figure 4 shows the time course of inactivationof PheDH preparations at different temperaturesranging from 45 to 60 �C. As can be observed,both enzyme preparations progressively lostactivity with time though a second order inacti-vation mechanism, and then the data obtainedwere analyzed according to a series-type enzymeinactivation model involving two first-order steps(Sadana & Henley 1987), in which k1 and k2 arethe inactivation rate constants. The kinetics con-stants were calculated by using a non-linearregression procedure based on the Marquardt-Levenberg method of iterative convergence in-cluded into the Microcal Origin 7.0 software(Microcal Software, Inc., MA, USA).

This kind of biphasic thermal inactivationprocess is characteristic for oligomeric proteinslike PheDH (Asano et al. 1987). However, itshould be noted that the modified enzyme pos-sessed lower inactivation rate constants and con-sequently higher values of half life times,

30 40 50 60 700

20

40

60

80

100

Rel

ativ

e A

ctiv

ity (

%)

Temperature (°C)

Fig. 2. Temperature-activity profile of native (s) anddextran-modified PheDH (d).The enzyme activity of nativeand modified enzyme preparations (about 10 U/ml, corre-sponding to 100% in the graphic), was measured at differenttemperatures in 100 mM glycine/KCl/KOH buffer, pH 10.4.As a control, a physical mixture of PheDH and amino dex-tran (X) was also evaluated.

30 40 50 60 700

20

40

60

80

100

Res

idua

l Act

ivity

(%

)

Temperature (°C)

Fig. 3. Thermal stability profile of native (s) and dextran-modified PheDH (d). Native and modified enzyme prepara-tions (about 50 U/ml, corresponding to 100% in the graphic),were incubated in 10 mM potassium phosphate buffer, pH 7.0,containing 1 mM EDTA and 5 mM 2-mercaptoethanol at thestated temperatures for 10 min. Samples were removed, chil-led quickly, and assayed for enzymatic activity. As a control,a physical mixture of PheDH and amino dextran (X) was alsoevaluated.

1314

indicating an enhancement in its thermal stability(Table 2). The thermostabilization effect for dex-tran-modified PheDH was especially noticeable

at 45 �C, temperature at which the activationGibbs energy of the first phase of thermal inacti-vation (DGi) (Darias & Villalonga 2001) in-creased in about 16.8 kJ/mol.

The molecular events behind the thermal sta-bilization showed by modified PheDH may beexplained by the combined contribution of sev-eral factors, previously demonstrated to be effec-tive in the maintenance of the activeconformation of neo-glycoenzymes. Amongthese, the most important factors could be theconformational stabilization of dextran-modifiedPheDH molecules due to the formation of newintramolecular hydrogen bonds (Srivastava1991); and the hydrophilization of the non-polarsurface areas of the enzyme, preventing the ther-mal inactivation mechanisms associated with theformation of intermolecular protein aggregatesdue to hydrophobic interactions (Venkatesh &Sundaram 1998).

0 30 60 90 1200

20

40

60

80

100

Res

idua

l Act

ivity

(%

)

Time (min)

0 30 60 90 1200

20

40

60

80

100

Res

idua

l Act

ivity

(%

)

Time (min)

(a)

(b)

Fig. 4. Kinetics of thermal inactivation of native (a) anddextran-modified PheDH (b) at 45 �C (m), 50 �C (X), 55 �C(s) and 60 �C (n). Native and modified enzyme preparations(about 50 U/ml, corresponding to 100% in the graphic), wereincubated at different temperatures in 10 mM potassium phos-phate buffer, pH 7.0, containing 1 mM EDTA and 5 mM 2-mercaptoethanol. Aliquots were removed at scheduled times,chilled quickly, and assayed for enzymatic activity.

Table 2. Half-life times of native and dextran-modifiedPheDH at different temperatures.

Temperature (�C)/t1/2 (h) PheDH PheDH-dextran

45 t1/2 (1) 2.0 231

t1/2 (2) 19.3 n.d.

50 t1/2 (1) 1.0 9.6

t1/2 (2) 4.4 116

55 t1/2 (1) 0.7 4.4

t1/2 (2) 2.6 23

60 t1/2 (1) 0.07 0.4

t1/2 (2) 0.17 3.9

n.d.: Not determined.

300 325 350 375 4000

300

600

900

Rel

ativ

e In

tens

ity

300 325 350 375 4000

300

600

900

Rel

ativ

e In

tens

ity

λ

λ

emission (nm)

emission (nm)

(a)

(b)

Fig. 5. Fluorescence emission spectra of native (a) anddextran-modified PheDH (b) before (——) and after (ÆÆÆÆÆÆÆÆ)1 h incubation at 55�C.

1315

In order to prove these hypotheses, thefluorescence spectra of native and modified pro-tein were recorded at 25 �C and after 1 h incuba-tion at 55 �C (Figure 5). The fluorescence spectraof both native and modified PheDH uponexcitation at 280 nm showed an identical emis-sion maximum at 313 nm, which is characteristicof tryptophan residues buried into the hydropho-bic protein core. However, thermal treatment at55 �C resulted in the unfolding of non-modifiedenzyme, as evidenced by the decrease in the fluo-rescence intensity and the appearance of a sec-ondary band, shifted to the red zone of thespectra. On the contrary, dextran-modified en-zyme showed smaller decrease in the fluorescenceintensity after identical heat treatment, indicatingthat a more compact protein structure is formedafter glycosidation of PheDH with mono-activated dextran.

Figure 6 shows the influence of heat treat-ment at 55 �C on the light scattering intensity ofnative and modified enzymes. Increased lightscattering was observed for both enzyme formsafter incubation at this temperature, indicatingthe occurrence of intermolecular aggregation pro-cesses in the thermal inactivation mechanism ofthis protein. However, glycosidation of PheDHwith dextran resulted in the reduction ofintermolecular associations as evidenced by re-duced light scattering. This result suggests thatthe hydrophilic polysaccharide moieties pre-vented inactivation of PheDH due to aggregationprocesses when is incubated at elevatedtemperatures.

In this work PheDH was chemically glycosi-dated with an end-group aminated dextran. Thismodification resulted in a noticeable improve-ment of the conformational and thermal stabilityproperties of this oxido-reductase. The influenceof protein aggregation processes on the inactiva-tion mechanism of PheDH at elevated tempera-ture was also determined, as well as thereduction of this phenomenon after attachmentof the polysaccharide. Attending to these results,we suggest the covalent glycosidation of PheDHwith end-group aminated dextran as a usefulmethod for improving its resistance to heatinactivation.

Acknowledgements

This research was supported by grants from TheJapan Society for the Promotion of Sciences toR. Villalonga and Y. Asano (Grant S-04257),and from Toyama Medical-Bio Cluster (TheMinistry of Education, Culture, Sports, Scienceand Technology, Japan) to Y. Asano and S.Tachibana. Dextran derivative was synthesized inMatanzas University, Cuba, and the modificationof PheDH and its properties were investigated inToyama Prefectural University, Japan. Financialsupport to R. Villalonga from the InternationalFoundation for Science, Stockholm, Sweden, andthe Organisation for the Prohibition of ChemicalWeapons, The Hague, The Netherlands (GrantF/3004-1) is also acknowledged.

References

Asano Y (1999) Phenylalanine dehydrogenase. In: FlickingerMC & Drew SW eds. Encyclopedia of Bioprocess Technol-ogy: Fermentation, Biocatalysis and Bioseparation, NewYork: John Wiley & Sons, Inc, pp. 1955–1963.

AsanoY,YamadaA,KatoK,YamaguchiK,HibinoY,HiraiK,KondoK (1990) Enantioselective synthesis of (S)-amino acidsby phenylalanine dehydrogenase fromBacillus sphaericus: useof natural and recombinant enzymes. J. Org. Chem. 55: 5567–5571.

AsanoY,YamadaA,KatoY,YamaguchiK,HibinoY,HiraiK,Kondo K (1987) Phenylalanine dehydrogenase of Bacillusbadius. Purification, characterization and gene cloning.Eur. J.Biochem. 168: 153–159.

Bruneel D, Schacht E (1995) End group modification ofpullulan. Polymer 36: 169–172.

0 30 60 90 1200

45

90

135

180

nIcr

ease

d lig

ht s

catte

ring

(%)

Time of incubation at 55°C (min)

Fig. 6. Influence of time of incubation at 55 �C on lightscattering at 400 nm of native (s) and dextran-modifiedPheDH (d).

1316

Cao R, Fragoso A, Almiral E, Villalonga R (2003) Supramo-lecular chemistry of cyclodextrins in Cuba. Supramol. Chem.15: 161–170.

Darias R, Villalonga R (2001) Functional stabilization ofcellulase by covalent modification with chitosan. J. Chem.Technol. Biotechnol. 76: 489–493.

Dubois MK, Gilles A, Hamilton JK, Rebers PA, Smith F(1956) Colorimetric method for determination of sugars andrelated substances. Anal. Chem. 28: 350–356.

Fernandez M, Fragoso A, Cao R, Banos M, Ansorge-Schum-acher M, Hartmeier W, Villalonga R (2004) Functionalproperties and application in peptide synthesis of trypsinmodified with cyclodextrin-containing dicarboxylic acids.J. Mol. Catalysis B. Enzymatic 31: 47–52.

Gomez L, Villalonga R (2000) Functional stabilization ofinvertase by covalent modification with pectin. Biotechnol.Lett. 22: 1191–1195.

Mehvar R (2000) Dextrans for targeted and sustained delivery oftherapeutic and imaging agents. J. Control. Release 69: 1–25.

Sadana A, Henley JP (1987) Single-step unimolecular non-first-order enzyme deactivation. Biotechnol. Bioeng. 30: 717–723.

Srivastava RAK (1991) Studies on stabilization of amylase bycovalent coupling to soluble polysaccharides. EnzymeMicrob. Technol. 13: 164–170.

Venkatesh R, Sundaram PV (1998) Modulation of stabilityproperties of bovine trypsin after in vitro structural changeswith a variety of chemicalmodifiers.Protein Eng. 11: 691–698.

Veronese FM, Caliceti P, Schiavon O, Sergi M (2002) Polyeth-ylene glycol-superoxide dismutase, a conjugate in search ofexploitation. Adv. Drug Deliv. Rev. 54: 587–606.

Villalonga R, Fernandez M, Fragoso A, Cao R, Di Pierro P,Mariniello L, Porta R (2003) Transglutaminase-catalyzedsynthesis of trypsin-cyclodextrin conjugates. Kinetics andstability properties. Biotechnol. Bioeng. 81: 732–737.

1317

Production of fungal biomass immobilized loofa sponge (FBILS)-discsfor the removal of heavy metal ions and chlorinated compoundsfrom aqueous solution

M. Iqbal1, A. Saeed1, R.G.J. Edyvean2, B. O’Sullivan2 & P. Styring21Environment Biotechnology Group, Biotechnology and Food Research Centre, PCSIR LaboratoriesComplex, 54600, Lahore, Pakistan2Department of Chemical and Process Engineering, University of Sheffield, S1 3JD, Sheffield, UK

Received 10 June 2005; Revisions requested 10 June 2005; Revisions received 23 June 2005; Accepted 28 June 2005

Key words: cadmium, 4-chloroanisole, fungal immobilization, loofa sponge, metal biosorptionPhanerochaete chrysosporium

Abstract

A white rot basidiomycete, Phanerochaete chrysosporium, was immobilized on loofa sponge (FBILS) discs.It removed ca. 37 and 71 mg Cd (II) g)1 from 50 and 200 mg l)1 aqueous solutions and up to 89% of4-chloroanisole from a 10 mg l)1 aqueous solution. FBILS are physically strong and chemically re-calcitrant, resisting temperature, mechanical agitation, and variations in pH without alteration to shape,structure or texture.

Introduction

Pollution in water supplies and in waste-waterdischarge is a cause of increasing legislative andpublic concern. In many countries, industry isrequired to meet ever higher quality standardsand has a need for improved, and environmen-tally sound, waste treatment methodologies forthe removal of toxic chemicals from effluents.The use of biomass as biosorbents for pollutantsoffers an environmentally sound and potentiallylow cost alternative to existing technologies. Fun-gal biomasses have high affinities for toxic metals(Kratochvil & Volesky 1998) and organic chemi-cals (Perez et al. 1997, Reddy et al. 1998) inaqueous solution. Commercial application ofsuch biomass has been hindered by problemsassociated mainly with physical manipulation(McHale & McHale 1994). Low mechanicalstrength and fragmentation of the biomass cancause difficulties in the contacting and separationof the effluent and biomass and this limits pro-cess design.

Immobilization technologies have been sug-gested to overcome these problems (Trujillo et al.1995, Aloysius et al. 1999). Immobilization ofmicrobial biomass in polymeric gel matrices isthe most extensively studied method (Leenenet al. 1996, Arica et al. 2001). However, produc-tion of large amounts of gel beads needed forcommercial applications is expensive and requiresspecialist equipment. Furthermore, the use ofsuch polymeric matrices results in closed struc-tures with restrictive diffusion and low mechani-cal strength (Hu & Reeves 1997).

The ideal immobilization matrix is strong andresistant and has an open structure. The plant-derived Loofa sponge is an inexpensive and eas-ily available biological, and therefore renewable,matrix produced in most tropical and subtropicalcountries. The sponge is made up of intercon-necting voids with an open network of fibroussupport giving the potential for rapid contact ofimmobilized cells to the surrounding aqueousmedium. Merits of the loofa biomatrix system in-clude freedom from materials that might be toxic

Biotechnology Letters (2005) 27: 1319–1323 � Springer 2005DOI 10.1007/s10529-005-0477-y

to microbial cells, simple application and opera-tion technique, and high stability during long-term repeated use.

The white rot basidiomycete, Phanerochaetechrysosporium, was chosen for this study as it hasa known affinity for metal ions but this is thefirst report on the immobilization of P. chrysos-porium on a biomatrix for the bioremediation ofboth inorganic and organic pollutants fromaqueous solution. While the mechanisms forinorganic and organic removal are likely to bedifferent, the inclusion of cadmium sorptionexperiments ensures continuity with previouswork on biosorbents and to test the hypothesisthat this form of immobilization does not affectmetal uptake, and therefore surface reactivity ofthe fungal hyphae.

Materials and methods

Microorganism and culture medium

The white-rot basidiomycete, Phanerochaete chry-sosporium ATTC 24725, was grown on (g l)1 dis-tilled water); D-glucose, 10; KH2PO4, 2; MgSO4 Æ7H2O, 0.5; NH4Cl, 0.1; CaCl2 ÆH2O, 0.1; thia-mine, 0.001; at pH 4.5.

Immobilizing materials and productionof FBILS-discs

Loofa sponge for use as an immobilization matrixwas obtained from the ripened dried fruit of Luffacylindrica. The loofa was cut into discs of approx-imately 2.5 cm diam. and 2–3 mm thick, soakedin boiling water for 30 min, thoroughly washedunder tap water and left for 24 h in distilled wa-ter, changed 3–4 times. The discs were then ovendried at 70 �C and stored in a desiccator.

A mycelium suspension of P. chrysosporium,0.5 ml, was inoculated in 100 ml of autoclavedgrowth medium containing four pre-weighed loo-fa sponge discs in 250 ml Erlenmeyer flasks.Flasks, with no loofa sponge discs in the med-ium, were inoculated to provide free fungal bio-mass controls. The inoculated flasks were shakenat 100 rpm at 34 �C. After 8 days, both free andloofa immobilized biomass of P. chrysosporium(hereafter called FBILS – Fungal biomass immo-bilized loofa sponge) were harvested from the

medium, washed twice with distilled water andstored at 4 �C until use. The dry weight of thefungal biomass was determined by weighing ovendried (70 �C overnight) sponge discs before andafter fungal growth.

Results and discussion

Properties of loofa sponge

The successful use of immobilized biosorbentsrequires that the immobilization matrix providesa high surface contact area and is stable to ad-verse chemical and physical treatments. Neitherautoclaving (10 times for 20 min), nor pH(2.0–12 for 24 days) produced any change in theshape and structure of the sponge. Table 1 showsthe lowest and highest values of physical parame-ters of loofa sponge discs. These results indicatethat loofa sponge discs can be repeatedly reusedin adverse conditions.

Production and properties of FBILS biosorbent

Microscopic examination shows hyphal growthin the sponge matrix within 24 h of incubation.Complete coverage of the sponge disc with thehyphae of P. chrysosporium occurs within 5 days(Figure 1a–c) and growth continues until theattainment of stationary phase at day 7. Whilethe immobilized hyphal biomass is packed tightlywithin the sponge there remain large numbers ofmicro-channels for free movement of solute dur-ing the biosorption process (Figure 1c). In con-trast, the free hyphal growth was compact andpelleted. At day 8, immobilized P. chrysosporiumhad a 21% increase in biomass over the freelygrowing control with biomass levels reaching1900 mg l)1. Such an increase is unusual inimmobilized systems.

Table 1. Some physical characteristics of loofa (Luffa cylindri-ca) sponge.

Physical properties

Structural nature Fibrous network

Porosity (%) 85–95

Density (g/cm3) 0.018–0.05

Specific pore volume (cm3/g) 26–34

1320

No change in the shape, size or weight ofFBILS was observed during exposure to variouspH in the range 2–12 and were stable whenexposed to acid (HCl), alkali (NaOH) and salt(NaCl) solutions for 5 days. FBILS retain 99%of the immobilized biomass within the loofasponge when shaken for 7 days at 150 rpm indistilled water. In contrast to the results forFBILS, significant cell leakage has been re-ported to occur during biosorption from systemsimmobilized using polymer gel systems (Hu &Reeves 1997). These results show that, in con-trast to the polymer gel immobilization methodwhich requires more sophisticated equipmentinvolving high costs, the more robust FBILSsystem can be made simply by adding themicrobial cell/hyphal/spore suspensions to a

growth medium containing the inexpensivesponge discs without any prior chemicaltreatment.

Metal removal studies

Figure 2 shows the efficiency of the removal ofCd (II) from solution by FBILS, free fungal bio-mass and loofa sponge control. While other au-thors have reported reductions in metal ion

Fig. 1. Immobilization of Phanerochaete chrysosporium withinloofa sponge discs: (a) loofa sponge disc; (b) loofa spongedisc covered with P. chrysosporium hyphal biomass; (c) scan-ning electron micrographs of immobilized P. chrysosporiumshowing micro-channel and void volume for free solute move-ment. White scale bars at bottom of micrograph = 10 lm.

Fig. 2. Biosorption of Cd (II) from (a) 50 mg l)1 and (b)200 mg l)1 solutions by free or immobilized Phanerochaetechrysosporium. Cd (II) solutions were prepared fromCd(NO3)2 and adjusted to pH 5 using 0.1 M NaOH. Freshdilutions were used for each experiment. Hundred milligramof free or immobilized (FBILS-discs) fungal biomass was con-tacted with 100 ml Cd (II) solution in 250 ml flasks shaken at100 rpm at 20±2 �C. Free fungal biomass was separated bycentrifugation at 3500� g for 5 min, whereas FBILS-discswere separated by simple decantation. Residual concentra-tions of Cd (II) in the supernatant were determined using anatomic absorption spectrophotometer. Metal-free and fungalbiomass-free solutions were used as controls. Statistical analy-sis of the data was carried out according to the Duncan’s newmultiple range test (Steel & Torrie 1996).

1321

uptake when biomass is immobilized, these re-sults show that this form of immobilization doesnot affect metal uptake capacity of the fungal hy-phae. Mahan and Holcombe (1992) reported a40% reduction in the sorption of Pb (II) whenStichococcus bacillaris was immobilized on silicagel and Lopez et al. (2002) report a 60% de-crease in metal sorption by Pseudomonas fluores-cens cells immobilized in agar beads, both incomparison with free cells. The statistically sig-nificant lower uptake of Cd (II) by free hyphalbiomass found in this work may be due to areduction in the surface area available for sorp-tion due to hyphal aggregation and pelletization.Such reductions due to aggregation have beenfound for yeast cells as well as fungal hyphae (deRome & Gadd 1987, Plette et al. 1996, Aloysiuset al. 1999). The findings presented here indicateno diffusional limitations and demonstrate thatFBILS are better suited for biosorption andother reactions than either free hyphal biomassor polymeric gel immobilization. From 50 mg l)1

metal solution uptake reached 36.8±0.7 mg g)1

fungal biomass for FBILS but only 29.7±0.9 mg g)1 free fungal biomass. For 200 mg l)1

metal solution, uptake reached 71.3±1.3 mg g)1

fungal biomass for FBILS but only 59.9±1.5 mg g)1 free fungal biomass.

From isotherm studies the maximum uptakelevels are 75.9±1.7 mg g)1 for FBILS and63.7±1.5 mg g)1 for free fungal biomass fromCd (II) concentrations of 250 mg l)1 and above.

Chlorinated aromatic organic compound removalstudies

Chlorinated aromatic organic compounds pose se-vere environmental and health hazards. In partic-ular, methods for the removal of poly-chlorinateddioxins and dibenzofurans are of widespreadinterest. Because of the hazardous properties ofthese materials, 4-chloroanisole in aqueous solu-tion was chosen as a model compound as it can beregarded as being structurally similar to half adichlorodioxin molecule (Figure 3).

After 7 days, contact at 34 �C, FBILSremoved 84%, 78% and 69% of 4-chloroanisolefrom a 5, 10 and 20 mg l)1 aqueous solutions(Table 2). No removal was detected in the4-chloroanisole control using untreated loofasponge. Hundred percent of 4-chloroanisole was

removed from 250 ml of 10 mg solution in8 days (Figure 4). There was no evidence in thechromatograms for the presence of aromatic deg-

Fig. 3. Structures of 4-chloroanisole (1) and dichlorodioxin(2).

Table 2. Removal of different concentrations of 4-chloroani-sole by fungus biomass immobilized in loofa sponge (FBILS)discs after 7 days.

4-chloroanisole, initial

concentration (mg l)1)

4-chloroanisole removed

(mg g)1 FBILS disc)

5 14.1±0.9

10 25.4±1.2

20 45.8±2.8

Fig. 4. Effect of contact time on the removal of 4-chloroani-sole from aqueous solution. FBILS-discs were shaken at100 rpm for 10 days in an aqueous solution of 10 mg l)1 4chloroanisole at 35 �C. This concentration (and the othersused) is well below the maximum solubility of 4-chloroanisolein water (237 mg l)1, Lun et al. 1995). FBILS were removedafter different periods of contact by simple decantation. Hun-dred millilitre of the decanted culture medium was mixed withan equal volume of diethyl ether in a separating funnel. Theorganic layer was removed, reduced to 10% volume and anal-ysed for 4-chloroanisole by GC using a WCOT (Wall CoatedOpen Tubular) fused silica capillary column, 30 m� 0.25 mmID. Concentrations of 4-chloroanisole were determinedagainst calibration standards using accurately obtained solu-tions of the organic material in diethyl ether. Dry weight offungal biomass was determined after drying in an oven at70 �C overnight.

1322

radation intermediates such as anisole or phe-nols, however, the control studies show that thisremoval does not occur by an adsorption pro-cess. It is known that P. chrysosporium will de-grade the aromatic ring structure of chlorinatedaromatic compounds to carboxylic acids and car-bon dioxide (Valli et al. 1992), so this is the like-ly pathway. Such compounds would not bedetected using the current experimental tech-nique. These results clearly demonstrated theability of FBILS to remove the chlorinatedorganic compound from aqueous solution. Fur-thermore, the absence of aromatic degradationproducts demonstrates added value in the pro-cess, as degrading chlorodioxins simply todioxins would only lead to a small decrease intoxicity. Degradation of the whole structure tocarboxylic acids represents a much cleaner over-all remediation process.

Conclusions

Loofa sponge is an effective immobilization ma-trix for the entrapment of fungal hyphae to pro-duce the fungal biomass immobilized loofasponge (FBILS). FBILS have a high capacity toremove both the toxic metal and chlorinated or-ganic compounds from aqueous solution. Highbioremoval capacity, good mechanical strength,ease of handling, high porosity and low costavailability of the immobilization matrix are fea-tures which lend this system to practical applica-tions for the removal of metals and chlorinatedaromatic compounds from industrial effluents.

References

Aloysius R, Karim MIA, Ariff AB (1999) The mechanismof cadmium removal from aqueous solution by non-metabolizing free and immobilized live biomass of Rhizopusoligosporus. World J. Microbiol. Biotechnol. 15: 571–578.

Arica MY, Kacar Y, Gene O (2001) Entrapment of white rotfungus Trametes versicolor in ca-alginate beads: prepa-ration and biosorption kinetic analysis for cadmiumremoval from aqueous solution. Bioresource Technol. 80:121–129.

de Rome L, Gadd GM (1987) Copper adsorption by Rhizopusarrhizus, Cladosporium resinae and Penicillium italicum.Appl. Microbiol. Biotechnol. 26: 84–90.

Hu MZC, Reeves M (1997) Biosorption of uranium byPseudomonas aeruginosa strain CUS immobilized in a novelmatrix. Biotechnol. Prog. 13: 60–70.

Kratochvil D, Volesky B (1998) Advances in the biosorption ofheavy metals. Trends Biotechnol. 16: 291–302.

Leenen EJTM, Dos Santos VAPM, Grolle KCF, Tramper J,Wijffels RH (1996) Characteristics of and selection criteriafor cell immobilization in wastewater treatment. Water Res.30: 2985–2996.

Lopez A, Lazaro N, Morales S, Marques AM (2002) Nickelbiosorption by free and immobilized cells of Pseudomonasfluorescens 4F39: a comparative study. Water Air Soil Poll.135: 157–172.

Lun R, Shiu WY, Mackay D (1995) Aqueous solubilities andoctanol water partition coefficients of chloroveratroles andchloroanisoles. J. Chem. Eng. Data 40: 959–962.

Mahan CA, Holcombe JA (1992) Immobilization of algae cellson silica gel and their characterization for trace metalpreconcentration. Anal. Chem. 64: 1933–1939.

McHale AP, McHale S (1994) Microbial biosorption of metals:potential in the treatment of metal pollution. Biotechnol.Adv. 12: 647–652.

Perez RR, Benito GG, Miranda MP (1997) Chlorophenoldegradation by Phanerochaete chrysosporium. BioresourceTechnol. 60: 207–213.

Plette ACC, Benedetti MF, Riemsdjik WH (1996) Competitivebinding of protons, calcium, cadmium and zinc to isolatedcell walls of a gram-positive soil bacterium. Environ. Sci.Technol. 30: 1902–1910.

Reddy GVB, Gelpke MDS, Gold MH (1998) Degradation of2,4,6-trichlorophenol by Phanerochaete chrysosporium:involvement of reductive dechlorination. J. Bacteriol. 180:5159–5164.

Steel RGD, Torrie JH (1996) Principles and Procedures ofStatistics: A Biometrical Approach, 3New York:McGraw-Hill.

Trujillo EM, Sprinti M, Zhuang H (1995) Immobilizedbiomass: a new class of heavy metal ion exchangers. In:Senguptal AK, ed. Ion Exchange Technology: Advances inPollution Control, Pennsylvania: Technomic PublishingCompany Inc, pp. 225–271.

Valli K, Wariishi H, Gold MH (1992) Degradation of 2,7-dichlorodibenzo-p-dioxin by the lignin-degrading basid-omycete Phanerochaete chrysosporium. J. Bacteriol. 174:2131–2137.

1323

A metal ion as a cofactor attenuates substrate inhibition in the enzymaticproduction of a high concentration of D-glutamate using N-acyl-D-glutamateamidohydrolase

Kazuaki Yoshimune, Ai Hirayama & Mitsuaki Moriguchi*Department of Applied Chemistry, Faculty of Engineering, Oita University, Dannoharu 700, 870-1192, Oita,Japan*Author for correspondence (Fax: +81-97-554-7890; E-mail: mmorigu@cc.oita-u.ac.jp)

Received 5 May 2005; Revisions requested 2 June 2005; Revisions received 27 June 2005; Accepted 27 June 2005

Key words: D-aminoacylase, D-aspartate, D-glutamate production, N-acyl-D-aspartate amidohydrolase, N-acyl-D-glutamate amidohydrolase

Abstract

N-Acyl-D-glutamate amidohydrolase (D-AGase) was inhibited by 94 % when 1 mol/l N-acetyl-DL-glutamate was used as a substrate. The addition of 1 mM Co2+ stabilized D-AGase. Moreover, the sub-strate inhibition was weakened to 88% with the addition of 0.4 mM Co2+ to the reaction mixture. AlthoughD-AGase is a zinc-metalloenzyme, the addition of Zn2+ from 0.01 to 10 mM did not increase the D-glutamicacid production in the saturated substrate. Under optimal conditions, 0.38 M D-glutamic acid was obtainedfrom N-acyl-DL-glutamate with 100% of the theoretical yield after 48 h.

Introduction

N-Acyl-D-amino acid amidohydrolases catalyzethe hydrolysis of N-acyl derivatives of various D-amino acids to D-amino acids and fatty acids andcan be used to resolve DL-amino acids. They aredivided into three types according to their sub-strate specificities. D-Aminoacylase (D-ANase)acts on N-acyl derivatives of various neutral D-amino acids. D-ANase from Alcaligenes xylosoxy-dans subsp. xylosoxydans A-6 (Alcaligenes A-6)has been applied as a commercial enzyme (D-ami-noacylase ‘‘Amano’’) for the optical resolutionproduction of neutral D-amino acids (Wakayamaet al. 2003). N-Acyl-D-glutamate amidohydrolase(D-AGase) and N-acyl-D-aspartate amidohydro-lase (D-AAase) are specific for N-acyl-D-gluta-mate and N-acyl-D-aspartate, respectively(Wakayama & Moriguchi 2001).

The D-ANase of Alcaligenes A-6 contains2.3 g atom Zn2+ per mol (Wakayama et al.

2000). However, the activity is inhibited by 92%by 1 mM Zn2+ (Moriguchi et al. 1993b). Laiet al. (2004) reported that a large excess of Zn2+

strongly inhibits D-ANase of Alcaligenes faecalisDA1 because it changes the conformation of itsactive center. D-AGase of Pseudomonas sp. strain5f-1 is zinc-metalloenzyme, and the activity isinhibited by 47% by 2 mM Zn2+ (Wakayamaet al. 1995b). The activity of D-AAase fromAlcaligenes A-6 is also inhibited by 73% by2 mM Zn2+ (Moriguchi et al. 1993a). Theseinhibitions may be of the same manner as that ofD-ANase of Alcaligenes faecalis DA1. Co2+ canoften substitute for Zn2+ to restore the activityof the apo-form of the zinc-metalloenzyme(Vallee & Galdes 1984). Moreover, D-AGase ofPseudomonas sp. strain 5f-1 is stabilized by Co2+

(Wakayama et al. 1995b). Although D-AGase isinhibited by Zn2+, Co2+ has no inhibitory effecton the enzyme at the concentration that stabilizesthe enzyme (Wakayama et al. 1995b). Thus,

Biotechnology Letters (2005) 27: 1325–1328 � Springer 2005DOI 10.1007/s10529-005-0480-3

Co2+ can be a useful additive for stabilizingzinc-metalloenzymes, such as D-AGase.

In industrial productions using enzymes, highconcentrations of substrate are often reacted fora long time to increase the production yield andrecovery of the product. Thus, industrial enzymesare required to have a high activity and stabilityin the presence of high concentrations of its sub-strate and product. Previously, we reported thatthe production levels of D-AGase of AlcaligenesA-6 in Escherichia coli are increased by the coex-pression of molecular chaperones (Yoshimuneet al. 2004). Despite the increasing activity of D-AGase, the enzyme is unable to produce a highconcentration of D-glutamic acid in a high yield.This may be due to the substrate inhibition andstability of D-AGase in the presence of high con-centrations of its substrate. Here we report howto remove substrate inhibition in an industrialproduction using an enzyme.

Materials and methods

Materials

Plasmid pKGSD2 encoding D-AGase (Yoshi-mune et al. 2004) and plasmid pETAD1 encod-ing D-AAase (Wakayama et al. 1995c) wereprepared as described previously. N-Acetyl-DL-glutamate and N-acetyl-DL-aspartate werepurchased from Sigma. All other chemicals werefrom Wako Pure Chemicals.

Enzyme preparation

The recombinant D-AGase or D-AAase was over-produced in E. coli using pKGSD2 or pETAD1,respectively, as previously described (Wakayamaet al. 1995c; Yoshimune et al. 2004). E. coli in10 mM potassium phosphate buffer (pH 7.0) wasdisrupted by sonication, and cell debris was re-moved by centrifugation. The supernatant ob-tained was used as the enzyme solution.

Assay of enzyme activity

The activities of D-AGase and D-AAase were as-sayed by measuring the D-amino acid formed.The reaction mixture (0.2 ml) contained 100 mM

potassium phosphate buffer (pH 7.0), 40 mM N-acetyl-DL-amino acid and enzyme (10 ll). Afterincubation at 30 �C for 10 min, the reaction wasstopped by adding 0.1 ml 0.25 M NaOH. The D-amino acid formed was measured by the 2,4,6-trinitrobenzenesulfonic acid (TNBS) method(Fields 1972). One unit of the enzyme was de-fined as the amount of enzyme that catalyzed theformation of 1 lmol D-amino acid per min.

Using substrates above 0.1 M, reaction mix-tures were adjusted to pH 8.5 with 25% (w/v)NaHCO3. After incubation at 30 �C for 1 h, thereaction was stopped by boiling for 3 min. Theinsoluble substrate and product were dissolvedwith the addition of water. The concentration ofthe D-amino acid in the solution was determinedby the TNBS method.

Determination of the concentration of productand substrate

The D-amino acid concentrations were deter-mined by the TNBS method. For the determina-tion of the concentration of N-acetyl-DL-aminoacid, the acid was hydrolyzed by the correspond-ing N-acyl-D-amino acid amidohydrolase, andthe D-amino acid produced was measured withthe TNBS method. The concentration of satu-rated N-acetyl-DL-amino acid was defined as theconcentration in the supernatant of a solutioncontaining 1.5 M N-acetyl-DL-amino acid and25% (w/v) NaHCO3 at 30 �C. The solubility ofthe N-acetyl-DL-amino acids was increased byneutralization with sodium hydrogencarbonate.The solubility of N-acetyl-DL-glutamate or N-acetyl-DL-aspartate in 25% (w/v) NaHCO3 wasabout 0.96 or 0.65 M, respectively.

Results and discussion

Effects of the substrate concentrationson the activities of N-acyl-D-amino acidamidohydrolases

The activities of D-AGase and D-AAase wereinhibited by 83% and 94 %, respectively, with1 M N-acetyl-DL-amino acids. On the other hand,the activities of D-AGase and D-AAase wereunaffected by their products, 0.4 M D-glutamic

1326

acid and 0.4 M D-aspartic acid, respectively, inthe presence of 1 M N-acetyl-DL-amino acids.These results suggest that product inhibition didnot take place and the productivities of D-glu-tamic and D-aspartic acids increase by the atten-uation of substrate inhibition.

Effect of Co2+ on the production of D-glutamicacid in the saturated substrate

Figure 1 shows the effect of Co2+ on the produc-tion of D-glutamic acid. Its productivity was in-creased more than three times by 1 mM Co2+ in1.25 M N-acetyl-DL-glutamate. In the absence ofCo2+, D-glutamate was not produced after 3 h.On the other hand, a small increase of D-gluta-mate was observed from 3 h through 24 h in thepresence of Co2+ (data not shown). The increaseof production level were considered probablecause of the stabilization by Co2+. D-AGase wasstabilized by Co2+ at 30 �C (Table 1) but not at40 �C (data not shown). Co2+ might slightlychange the conformation of D-AGase to stabilizeit at 30 �C. Activity of D-AGase was doubled by0.4 mM Co2+ with 0.96 M N-acetyl-DL-glutamate(data not shown). However, the D-AGase activitywas not affected by 0.4 mM Co2+ when 40 mM

N-acetyl-DL-glutamate was used as a substrateconcentration (standard assay condition; datanot shown). Co2+ might change the conforma-tion of the active center of D-AGase to weakenthe inhibition of the substrate. Although D-AG-ase is considered to be a zinc-metalloenzyme(Wakayama et al. 1995a), Zn2+ did not affectthe production of D-glutamic acid in the satu-rated substrate (data not shown). The productionlevel of D-glutamate was not increased by 1 mM

Mg2+, Mn2+, Fe2+, Fe3+, Ni2+, Cu2+ or Ba2+

(data not shown). Zn2+ of zinc-metallo enzymecan be often replaced by Co2+ or Mn2+. Fur-thermore, L-aminoacylase (Toogood et al. 2002)and thermolysin (Holland et al. 1995) are acti-

vated by Co2+. Co2+ can be a useful additivefor activating and stabilizing zinc-metalloenzymes. Neither Co2+ nor Zn2+ increased theproduction level of D-aspartic acid by D-AAasein the saturated substrate.

D-Amino acid production by N-acyl-D-aminoacid amidohydrolases

Figure 2 shows the time course of the productionof D-glutamic acid with the use of D-AGase in1 mM Co2+ and 0.77 M N-acetyl-DL-glutamate.Only 20% of theoretical yield was obtained up toapproximately 14 h of the reaction, probably be-cause of the substrate inhibition. After 20 h ofthe reaction, the amount of D-glutamic acid rap-idly increased due to the lower concentration ofthe substrate. Under the same condition, D-

Fig. 1. Effect of Co2+ on the production of D-glutamic acid.A reaction mixture containing 1.25 M N-acetyl-DL-glutamate,25% (w/v) NaHCO3, 125 mM potassium phosphate buffer(pH 7.0), and 2 U/ml D-AGase was incubated with variousconcentrations of Co2+ for 24 h at 30 �C. The reaction wasstopped by boiling for 3 min.

Table 1. Effect of various metal ions on the stability of D-AGase.

Metal ion No addition Mg2+ Mn2+ Fe2+ Fe3+ Co2+ Ni2+ Cu2+ Zn2+ Ba2+

Concentration (mM) 1 1 1 1 0.2 1 2 4 1 1 1 1

Residual activity (%) 46 49 0 2.4 9.7 42 98 62 31 1.4 0 52 43

The enzyme solutions containing 100 mM potassium phosphate buffer (pH 7.0) and 2 U D-AGase/ml in the presence of various metalions were incubated at 30 �C for 12 h. The residual D-AGase activities were measured.

1327

aspartic acid was produced at 70% of the theo-retical yield for 48 h (Figure 3). A slight increasein D-aspartic acid was observed after 24 h of thereaction, probably due to the inactivation of theenzyme. D-AAase must be stabilized for the pro-duction of D-aspartic acid with 100 % of the the-

oretical yield in this condition. Further studieswill be required to understand the mechanisms ofCo2+ action on D-AGase.

References

Fields R (1972) The rapid determination of amino groups withTNBS. Method Enzymol. 25: 464–468.

Holland DR, Hausrath AC, Juers D, Matthews BW (1995)Structural analysis of zinc substitutions in the active site ofthermolysin. Protein Sci. 4: 1955–1965.

Lai WL, Chou LY, Ting CY, Kirby R, Tsai YC, Wang AHJ,Liaw SH (2004) The functional role of the binuclear metalcenter in D-aminoacylase. J. Biol. Chem. 279: 13962–13967.

Moriguchi M, Sakai K, Katsuno Y, Maki T, Wakayama M(1993a) Purification and characterization of novel N-acyl-D-aspartate amidohydrolase from Alcaligenes xylosoxydanssubsp. xylosoxydans A-6. Biosci. Biotech. Biochem. 57: 1145–1148.

Moriguchi M, Sakai K, Miyamoto Y, Wakayama M (1993b)Production, purification, and characterization of D-amino-acylase from Alcaligenes xylosoxydans subsp. xylosoxydansA-6. Biosci. Biotech. Biochem. 57: 1149–1152.

Toogood HS, Hollingsworth EJ, Brown RC, Taylor IN, TaylorSJC, McCague R, Littlechild JA (2002) A thermostableL-aminoacylase from Thermococcus litoralis: cloning,overexpression, characterization, and applications in bio-transformations. Extremophiles 6: 111–122.

Vallee BL, Galdes A (1984) The metallobiochemistry of zincenzymes. Adv. Enzymol. Relat. Areas. Mol. Biol. 56: 283–430.

Wakayama M, Ashika T, Miyamoto Y, Yoshikawa T, SonodaY, Sakai K, Moriguchi M (1995a) Primary structure ofN-acyl-D-glutamate amidohydrolase from Alcaligenes xylos-oxydans subsp. xylosoxydans A-6. J. Biochem. 118: 204–209.

Wakayama M, Miura Y, Oshima K, Sakai K, Moriguchi M(1995b) Metal-characterization of N-acyl-D-glutamate ami-dohydrolase from Pseudomonas sp. strain 5f-1. Biosci.Biotech. Biochem. 59: 1489–1492.

Wakayama M, Watanabe E, Takenaka Y, Miyamoto Y, TauY, Sakai K, Moriguchi M (1995c) Cloning, expression, andnucleotide sequence of the N-acyl-D-aspartate amidohydro-lase gene from Alcaligenes xylosoxydans subsp. xylosoxydansA-6. J. Ferment. Bioeng. 80: 311–317.

Wakayama M, Moriguchi M (2001) Comparative biochemistryof bacterial N-acyl-D-amino acid amidohydrolase. J. Mol.Catal. B: Enzym. 12: 15–25.

WakayamaM, Yada H, Kanda S, Hayashi S, Yatsuda Y, SakaiK, Moriguchi M (2000) Role of conserved histidine residuesin D-aminoacylase from Alcaligenes xylosoxydans subsp.xylosoxydans A-6. Biosci. Biotechnol. Biochem. 64: 1–8.

Wakayama M, Yoshimune K, Hirose Y, Moriguchi M (2003)Production of D-amino acid amidohydrolase and its struc-ture and function. J. Mol. Catal. B: Enzym. 23: 71–85.

Yoshimune K, Ninomiya Y, Wakayama M, Moriguchi M(2004) Molecular chaperones facilitate the soluble expressionof N-acyl-D-amino acid amidohydrolases in Escherichia coli.J. Ind. Microbiol. Biotechnol. 31: 421–426.

Fig. 2. Time course of D-glutamic acid production. A reactionmixture containing 0.77 M N-acetyl-DL-glutamate, 0.2 g/ml so-dium hydrogencarbonate, a 380 mM potassium phosphatebuffer (pH 7.0), 1 mM Co2+, and 2 U D-AGase was incu-bated at 30 �C.

Fig. 3. Time course of D-aspartic acid production. A reactionmixture containing 0.77 M N-acetyl-DL-aspartate, 0.2 g/NaH-CO3/ml and 380 mM potassium phosphate buffer (pH 7.0),and 2 U D-AAase was incubated at 30 �C.

1328

Characterization of an extracellular serine protease genefrom the nematophagous fungus Lecanicillium psalliotae

Jinkui Yang1, Xiaowei Huang1, Baoyu Tian1, Hui Sun1, Junxin Duan2, Wenping Wu2 &Keqin Zhang1,*1Laboratory for Conservation and Utilization of Bio-resources, Yunnan University, 650091, Kunming,P. R. China2Research and Development Center of Novozymes in China, 100085, Beijing, P. R. China*Author for correspondence (Fax: +86-871-5034878; E-mail: kqzhang111@yahoo.com.cn)

Received 27 April 2005; Revisions requested 24 May 2005; Revisions received 27 June 2005; Accepted 29 June 2005

Key words: gene cloning, Lecanicillium psalliotae, serine protease, similarity comparison

Abstract

The gene encoding a cuticle-degrading serine protease was cloned from three isolates of Lecanicilliumpsalliotae (syn. Verticillium psalliotae) by 3¢ and 5¢ RACE (rapid amplification of cDNA ends) method. Thegene encodes for 382 amino acids and the protein shares conserved motifs with subtilisin N and peptidaseS8. Comparison of translated cDNA sequences of three isolates revealed one amino acid polymorphism atposition 230. The deduced protease sequence shared high degree of similarities to other cuticle-degradingproteases from other nematophagous fungi.

Introduction

Extracellular enzymes are important virulencefactors in nematophagous and entomophagousfungi (Segers et al. 1994, Tunlid et al. 1994,Bonants et al. 1995, Joshi et al. 1995). Lopez-Llorca (1990) isolated a serine protease P32 fromPochonia suchlasporia (syn. Verticillium suchlas-porium) and found it was involved in egg pene-tration of nematode. Subsequently, twoproteases, VCP1 and PIP, were isolated from ne-matophagous fungi Pochonia chlamydosporia(syn. Verticillium chlamydosporium) and Paecilio-myces lilacinus, respectively (Segers et al. 1994,Bonants et al. 1995). Similar extracellular prote-ases also had been found in entomophagous fun-gi (St Leger et al. 1992, Joshi et al. 1995).Moreover, collagenase and chitinase have beenidentified from nematophagous fungi Arthrobot-

rys amerospora, Po. chlamydosporia and Po.suchlasporia (Schenck et al. 1980, Tikhonov et al.2002).

Lecanicillium psalliotae is a nematophagusfungus with commercial potential for thebiocontrol of root knot and cyst nematodes. Itproduces an alkaline serine protease, Ver112,during infection of the saprophytic nematodePanagrellus redivivus. Ver112 had been purifiedfrom culture filtrates of L. psalliotae. The N-ter-minal amino acid sequence has been submitted toSwiss-Prot (accession number Q68GV9). In thisreport, we described the cloning of an alkalineserine protease from L. psalliotae by the 3¢ and 5¢RACE method, the analysis of the primary ami-no acid sequence of protease Ver112 from threeisolates, and comparison with other cuticle-degrading serine proteases isolated from differentnematophagous and entomopathogenic fungi.

Biotechnology Letters (2005) 27: 1329--1334 � Springer 2005DOI 10.1007/s10529-005-0482-1

Materials and methods

Microorganisms and culture conditions

Three isolates (112, 602 and 608) of nematopha-gous fungus Lecanicillium psalliotae used in thisstudy were originally isolated from field soilsamples in Yunnan Province; strain 112 has beendeposited in the China General MicrobiologicalCulture Collection Center. Fungi were culturedin PD (potato/dextrose) medium at 26 �C withshaking at 200 rpm for 3 days.

Escherichia coli DH 5a was used in all DNAmanipulations and grown in Luria--Bertani med-ium containing (per liter): 10 g tryptone,10 g NaCl, 5 g yeast extract, and 16 g agar.

Genomic DNA and total RNA extraction

Mycelium were collected by filtration in a steril-ized filter funnel and ground to a fine powder inliquid N2. DNA was extracted according to themethod of Zhang et al. (1996).

Total RNA extraction was done according tothe manual of TRIzol Reagent (Invitrogen,America), and RNA was stored at )70 �C.

Amplification of 3¢ and 5¢ nucleotide sequence

A partial cDNA of Ver112 was obtained by 3¢RACE kit (Invitrogen, America) using a degener-ate primer, SERP3-1 5¢-ACNCARCARCARGGNGCNAC-3¢, which was designed according tothe N-terminal amino acid residues of the proteaseVer112. The first strand cDNA and target cDNAwere synthesized according to the manual of 3¢RACE system for rapid amplification of cDNAends. 5¢ RACE was conducted as described in themanual of 5¢ RACE system for rapid amplifica-tion of cDNA ends using two gene-specific prim-ers derived from the 3¢ RACE product, R5-15¢-AGTCTTGGACTCCGATGGTG-3¢, and R5-2 5¢-TGGGAGATGCGAGTAAGTC -3¢.

Amplification of the Ver112 chromosomal geneand full-length cDNA

Two gene-specific primers, FP 5¢-CTGATTAT-CAACAAGATGCGTC-3¢ and RP 5¢-TTACGTGGCGCCGTTGAAGGC-3¢, were designedaccording to the PCR fragments of 3¢ and 5¢

RACE, genomic DNA and the first strand ofcDNA was used as template, respectively. TargetDNA and cDNA were amplified by a touch-down program (Kim et al. 2003).

The cDNA and genomic sequences were com-pared using the DNAman software package(Version 5.2.2, Lynnon Biosoft, Canada).

Cloning and sequencing

The PCR products were purified from a 1% aga-rose gel using a DNA fragment purification kitver 2.0 (Takara, Japan) and subcloned intopGEM-T Vector (Promega, America). White col-onies were randomly selected and purified usingthe plasmid DNA purification kit (Qiagen, Ger-man) and the plasmid DNA was sequenced usingan ABI 3730 autosequencer (Perkin--Elmer,America) with four fluorescent dyes. Thesequencing primers were T7 and SP6 universalprimers (Takara, Japan). Sequence data wereanalyzed using DNAman software package. Se-quence identity was compared with other cuticle-degrading protease gene using the GenBankdatabase.

Sequence analysis

Database searches were performed usingBlastX (http://www.ncbi.nlm.nih.gov/BLAST).Signal sequence prediction was performedusing Signal P (http://www.cbs.dtu.dk/services/signalP) (Henrik et al. 1997). Multiple sequencealignments were performed using DNAmansoftware package. Proteins were examined forconserved motifs using Pfam (http://pfam. wus-tl.edu/hmmsearch. shtml) (Garcia-Sanchez et al.2004). N-linked glycosylation sites were pre-dicted by NetNGlyc (http://www.cbs.dtu.dk/ser-vices/NetNGlyc/).

Results

Cloning of the cuticle-degrading serine protease

Under the conditions described above, a 500 bpPCR product (Figure 1) was successfully ampli-fied by 5¢ RACE and sequencing indicated thatthe PCR fragment contained a putative start co-

1330

don (ATG). One thousand one hundred and fiftyand 1350 bp fragments (Figure 1) were amplifiedby using, respectively, the first strand of cDNAand genomic DNA as template from three iso-lates of L. psalliotae. These fragments were alsocloned and sequenced. The combined nucleotidesequence for the partial DNA and cDNA were1640 and 1440 bp, respectively.

Sequence analysis

The sequence of Ver112 comprised an ORF,which contained three introns and four exons. Itencoded a polypeptide of 382 amino acid resi-dues with a Mr of 39.654, which shared con-served motifs with subtilisin N and peptidase S8.Comparison of Ver112 with other serine prote-ases from nematophagous fungi revealed that itwas typical of fungal serine proteases, which pos-sessed a pre-pro-peptide structure. It has a signalpeptide (15 amino acids) consisting of the initialmethionine, a core of seven hydrophobic resi-dues, a helix-breaking residue (proline), and fourhydrophobic residues before a signal peptidasecleavage site (Ala-Leu-Ala). Comparison of thededuced amino acid sequence with the N-termi-nal sequence of Ver112 revealed that the matureprotein started at residue 103, and the final resi-due of the pro-peptide was an asparagine (N),position in 102. Each intron began with GT andended with AG, which was a common feature offungal introns and had been observed in the ser-ine protease gene from Acremonium chrysogenum(Isogai et al. 1991). The mature protein consistedof 280 amino acids.

Comparison of the nucleotide sequences ofVer112 from three isolates of L. psalliotae re-

vealed that they were very conservative, thenucleotide sequences from L. psalliotae 112 and608 were identical, and there were four nucleo-tide residues different from L. psalliotae 602, twoof them located at the second intron, and twoother variable nucleotides located at different ex-ons, which resulted in one amino acid polymor-phism at position 230, arginine (A) changed toglycine (G). Like VCP1, Ver112 lacks anyN-linked glycosylation site (Asn-X-Ser/Thr).These nucleotide sequences have been submittedto GenBank, under accession numbers AY692148 (112 and 608) and AY870806 (602).

Comparison of Ver112 with other serineproteases isolated from nematophagousand entomopathogenic fungi

These cuticle-degrading proteases shared somesimilar biochemical properties of low molecularmass and being inhibited by PMSF (phen-ylmethylsulfonylfluoride) (Table 1). However, PIIand Aozl isolated from nematode-trapping fungiA. oligospora had lower pI and higher molecularmasses than other proteases from nematopha-gous and entomopathogenic fungi.

The databank search showed that Ver112shared extensive similarities to fungal membersof the subtilisin family of serine proteases(Figure 2). The deduced amino acid sequence ofthe Ver112 showed 39.6%, 41.7%, 62.8%,75.7%, 57.0%, 61% and 58.2% identity, respec-tively, to Aozl (Arthrobotrys oligospora), PII (A.oligospora), PIP (Pic. lilacinus), Pr1 (Beauveriabassiana), PrA (Metarhizium anisopliae), Prk(Tritirachium album), and VCP1 (Po. chlamydos-poria). The signal peptide and pro-region cleav-

Fig. 1. Result of PCR amplification. Lanes 1, 2 and 4 -- result of 5¢ RACE amplification; lanes 3 and 5 -- DNA marker (Ladder100 bp, Promega, America); lanes 6 and 9 -- PCR fragment amplified with genomic DNA as template; lanes 7, 8 and 10 -- PCRfragment amplified with the first strand of cDNA as template.

1331

age sites of them were conserved and the firstamino acid of mature proteases was alanine.They shared the conservation of the aspartic acid(Asp143)--histidine (His173)--serine (Ser328) (inVer112) catalytic triad. The two blocks of side-chains that form the sides of the substrate-bind-ing S1 pocket in subtilisin occur in regions ofhigh similarity and consist of Ser236Leu237Gly238and Ala262Ala263Gly264, respectively, in Ver112.Furthermore, the highly conserved Asn265 (inVer112) is important in subtilisin for stabilizationof the reaction intermediate formed during prote-olysis (Kraut 1977).

Discussion

Extracellular serine proteases have been isolated,cloned and purified from several nematophagousand entomopathogenic fungi. From Table 1 andFigure 2, these cuticle-degrading serine proteasesfrom different nematophagous and entomopatho-genic fungi may be divided into two categoriesaccording to the difference of biochemical char-acterization and primary sequence. Class I is iso-lated from nematode-trapping fungi and haslower pI, and class II is isolated from nematode-

parasitic or egg-parasitic fungi and has higher pI.However, whether the differences of biochemicalcharacterization between classes II and I areimportant for the ability of the enzymes to de-grade components of the nematode cuticle andeggshell, respectively, and whether the differencesis connected to their mode of infection are cur-rently not known.

The high degree of similarities between extra-cellular serine proteases from different nemato-phagous and entomopathogenic fungi suggestthat they may derive from a common ancestralsubtilisin-like protease gene. Cloning of Ver112provides a good foundation for future investiga-tion of infection mechanism and improvementthe pathogenicity of nematophagous and ento-mopathogenic fungi.

Acknowledgement

We are grateful to Dr. Dilantha Fernando inthe University of Manitoba, Canada, and Dr.Li Haipeng in the Ludwig-Maximilians-Univer-sitat Munchen, Germany, for their invaluablecomments and revising the manuscript. The

Table 1. Partial characterization of cuticle-degrading serine proteases isolated from different nematophagous and entomopathogenicfungi.

Protease Fungus Molecular mass (kDa) Inhibitor of protease pI Reference

PII Arthrobotrys oligospora 35 PMSFa 4.6 Tunlid et al. (1994)

Aozl Arthrobotrys oligospora 38 PMSF, SSIb 4.9 Zhao et al. (2004)

VCPl Pochonia chlamydosporia 33 PMSF 10.2 Segers et al. (1994)

P32 Pochonia suchlasporia 32 PMSF, pCMBc -- Lopez-Llorca LV (1990)

Ver112 Lecanicillium psalliotae 32 PMSF -- GenBank (AAU01968)

PIP Paecilomyces lilacinus 33 PMSF 10.2 Bonants et al. (1995)

PrA Metarhizium anisopliae 25 PMSF 10.2 St Leger et al. (1992)

Pr1 Beauveria bassiana 32 PMSF 10.0 Joshi et al. (1995)

aPMSF, phenylmethylsulfonylfluoride.bSSI, Streptomyces subtilisin inhibitor.cpCMB, p-chloromercuric benzoic acid.

Fig. 2. Alignment of subtilase amino acid sequences from Arthrobotrys oligospora (PII and Aozl), Paeciliomyces lilacinus (PIP),Beauveria bassiana (Pr1), Metarhizium anisopliae (PrA), Tritirachium album (Prk), Pochonia chlamydosporia (VCP1) and Lecanicilli-um psalliotae (Ver112). The GenBank accession numbers are CAA63841, AAM93666, AAA91584, AAK70804, CAB64346, P06873,CAD20578 and AAU01968, respectively. Areas shaded in black are conserved regions (100% similarity), areas shaded in gray arehigh degree similarity (more than 50% similarity) and unshaded areas are regions of variability between the proteases. , indicatesPutative signal-sequence cleavage site; . indicates Proregion cleavage site. m indicates the aspartic acid (Asp143)--histidine (His173)--serine (Ser328) (in Ver112) catalytic triad. The underlined region is the substrate-binding S1 pocket in subtilisin.

c

1332

1333

work was funded by the projects from Minis-try of Science and Technology of China (ap-proved No. 2002BA901A21) and Departmentof Science and Technology of Yunnan Prov-ince (approved No. 2004C0001Z, No.2003C0003Q).

References

Ahman J, EK B, Rask L, Tunlid A (1996) Sequence analysisand regulation of a gene encoding a cuticle-degrading serineprotease from the nematophagous fungus Arthrobotrysoligospora. Microbiology 142: 1605--1616.

Bonants PJ, Fitters PF, Thijs H, Den BE, Waalwijk C,Henfling JW (1995) A basic serine protease from Paecilomy-ces lilacinus with biological activity against Meloidogynehapla eggs. Microbiology 141: 775--784.

Garcia-Sanchez A, Cerrato R, Larrasa J, Ambrose NC, ParraA, Alonso JM, Hermoso-de-Mendoza M, Rey JM, MineMO, Carnegie PR, Ellis TM, Masters AM, Pemberton AD,Hermoso-de-Mendoza J (2004) Characterization of anextracellular serine protease gene (nasp gene) from Derma-tophilus congolensis. FEMS Microbiol. Lett. 231: 53--57.

Henrik N, Engelbrech TJ, Brunak S, Von HG (1997)Identification of prokaryotic and eukaryotic signal peptidesand prediction of their cleavage sites. Protein Eng. 10: 1--6.

Isogai T, Fukagawa M, Kojo H, Kohsaka M, Aoki H,Imanaka H (1991) Cloning and nucleotide sequences of thecomplementary and genomic DNA for the alkaline proteasefrom Acremonium chrysogenum. Agric. Biol. Chem. 55: 471--477.

Joshi L, St Leger RJ, Bidochka MJ (1995) Cloning of a cuticle-degrading protease from the entomopathogenic fungus,Beauveria bassiana. FEMS Microbiol. Lett. 125: 211--218.

Kim SK, Lee SB, Kang TJ, Chae GT (2003) Detection of genemutations related with drug resistance in Mycobacteriumleprae from leprosy patients using Touch-Down (TD) PCR.FEMS Immunol. Med. Microbiol. 36: 27--32.

Kraut J (1977) Serine proteases structure and mechanism ofcatalysis. Annu. Rev. Biochem. 46: 331--358.

Lopez-Llorca LV (1990) Purification and properties of extra-cellular proteases produced by the nematophagous fungusVerticillium suchlasporium. Can. J. Microbiol. 36: 530--537.

Schenck S, Chase T, Jr, Rosenzweig WD, Pramer D (1980)Collagenase production by nematode-trapping fungi. Appl.Environ. Microbiol. 40: 567--570.

St Leger RJ, Frank DC, Roberts DW, Staple RC (1992)Molecular cloning and regulatory analysis of the cuticle-degrading protease structural gene from the entomopatho-genic fungus Metarhizium anisopliae. Eur. J. Biochem. 204:991--1001.

Segers R, Butt TM, Kerry BR, Peberdy JF (1994) Thenematophagous fungus Verticillium chlamydosporium pro-duces a chymoelastase-like protease which hydrolyses hostnematode proteins in situ. Microbiology 140: 2715--2723.

Tikhonov VE, Lopez-Llorca LV, Salinas J, Jansson HB (2002)Purification and characterization of chitinases from thenematophagous fungi Verticillium chlamydosporium andVerticillium suchlasporium. Fungal Genet. Biol. 35: 67--78.

Tunlid A, Rosen S, EK B, Rask L (1994) Purification andcharacterization of an extracellular serine protease from thenematode-trapping fungus Arthrobotrys oligospora. Micro-biology 140: 1687--1695.

Zhang D, Yang Y, Castlebury LA, Cerniglia CE (1996) Amethod for the large scale isolation of high transformationefficiency fungal genomic DNA. FEMS Microbiol. Lett. 145:261--265.

Zhao ML, Mo MH, Zhang KQ (2004) Characterization of aneutral serine protease and its full-length cDNA from thenematode-trapping fungus Arthrobotrys oligospora. Mycolo-gia 96: 16--22.

1334

Biotechnology Letters (2005) 27: 1335 � Springer 2005DOI 10.1007/s10529-005-1074-9