Production of Glomus intraradicesPropagules, an Arbuscular MycorrhizalFungus, in an Airlift Bioreactor

M. Jolicoeur,1 R. D. Williams,1 C. Chavarie,1 J. A. Fortin,2 J. Archambault3

1Ecole Polytechnique de Montreal, Department of Chemical Engineering,Biopro Research Centre, P.O. Box 6079 Centre-Ville Station, Montreal,Quebec, Canada, H3C 3A7; telephone: (514) 340-4711, ext. 4525; fax: (514)340-4159; e-mail: mario.jolicoeur@ polymtl.ca2Institut de Recherche en Biologie Vegetale, Universite de Montreal, 4101Sherbrooke St. East, Montreal, Quebec, Canada, H1X 2B23Universite du Quebec a` Trois-Rivie`res, Engineering School, Department ofChemical Engineering , P.O. Box. 500, Trois-Rivie`res, Quebec, Canada, G9A5H7

Received 10 May 1998; accepted 14 October 1998

Abstract: This work addresses the symbiotic culture ofthe arbuscular mycorrhizal (AM) fungus Glomus intrara-dices with Daucus carota hairy roots transformed byAgrobacterium rhizogenes, in two submerged culturesystems: Petri dish and airlift bioreactor. AM fungi playan active role in plant nutrition and protection againstplant pathogens. These fungi are obligate biotrophs asthey depend on a host plant for their needs in carbohy-drates. The effect of the mycorrhizal roots inoculum-to-medium volume ratio on the growth of both symbiontswas studied. A critical inoculating condition was ob-served at ~ 0.6 g dry biomass (DW) ? L1 medium, abovewhich root growth was significantly reduced when usinga low-salt minimal (M) liquid medium previously devel-oped for hairy root-AM fungi co-culture. Below criticalinoculum conditions the maximum specific root growthand specific G. intraradices spore production rates of0.021 and 0.035 d1, respectively, were observed for Petridish cultures. Maximum spore production in the airliftbioreactor was ten times lower than that of Petri dishcultures and obtained with the lowest inoculum assessed(0.13 g DW ? L1 medium) with 1.82 105 4.05 104

(SEM) spores (g DW inoculum)1 (L medium)1 in 107 d.This work proposes a second-generation bioprocess forAM fungi propagule production in bioreactors. 1999John Wiley & Sons, Inc. Biotechnol Bioeng 63: 224232, 1999.Keywords: Daucus carota; Glomus intraradices; Endo-mycorrhizae; AM fungi; propagules; vesicles; spores;hairy roots; bioreactors

INTRODUCTION

The widespread use of conventional fertilizers has signifi-cantly increased crop yields. Chemical fertilizers, however,have detrimental effects on the environment due to highlevels of run-off resulting in contamination of streams andgroundwater. An attractive alternative is the use of the natu-

ral plant growth stimulant arbuscular mycorrhizal (AM)fungi (Wood and Cummings, 1992). These fungi are obli-gate symbionts that co-evolved with vascular plants (Simonet al., 1993). Over two-thirds of plants species are naturallyliving in symbiosis with endomycorrhizal fungi (Newshamet al., 1995). This plant-AM fungi association is of highinterest as it has been shown to improve plant nutrient up-take (Smith, 1980), growth (Bethlenfalvay, 1992), as well asresistance to pathogens (Andrade et al., 1997; St-Arnaud etal., 1997, 1995, 1994) and drought (Arora et al., 1991;Khasa et al., 1990; Trappe et al., 1984). Thus, a source ofAM fungal inocula combined with appropriate field man-agement would permit efficient and ecologically improvedagricultural practices.

Few companies offer AM fungi inocula for agriculturaland horticultural uses. Endomycorrhizal inocula are avail-able in the form of spores (propagules formed outside theroot) and root pieces containing vesicles (propagulesformed inside the root tissue) (Sylvia and Jarstfer, 1992).They are added directly into the surrounding soil of seedsand plantlets. Fungal spores can also be pelletized withseeds (Hall, 1978). Once in the soil, the germination of AMfungi propagules can be induced by specific stimulatingfactors such as CO2 and root exudates (Becard and Piche,1989) in a humid environment. The elongation of germinat-ing hyphae is driven by the propagule energy reserves untilit finds and colonizes a plant root, and draws sugars fromthe plant (Pfeffer et al., 1998; Mosse and Hepper, 1975).

Currently, production methods of AM fungi are based ontheir natural proliferation process. Whole plants are inocu-lated with propagules, a mixture of mycorrhizal root piecesand spores, and grown on various soils or substrates (Mosseand Thompson, 1984; Hung and Sylvia, 1988; Dugassa etal., 1995), aeroponically (Hung and Sylvia, 1988), hydro-ponically (Dugassa et al., 1995), or by using the nutrientliquid film technique (Mosse and Thompson, 1984).

Originally developed for growing plants (Zobel et al.,Correspondence to: M. Jolicoeur

1999 John Wiley & Sons, Inc. CCC 0006-3592/99/020224-09

1976), these production systems are operated under poorlycontrolled conditions. Problems encountered include con-tamination and variability of product quality. Furthermore,it remains unclear whether these systems can produceenough endomycorrhizal inocula to meet the large potentialworld demand.

Initial attempts to co-culture these symbionts asepticallyhave involved surface-sterilized excised roots as fungal host(Mosse and Hepper, 1975). Later, Agrobacterium rhizo-genes transformed roots (Tepfer and Tempe, 1981) wereused as hosts for Endomycorrhizae fungi (Glomus mosseaeand Gigaspora margarita) propagation (Mugnier andMosse, 1987). After modifications of the culture mediumand growth conditions, Becard and Fortin (1988) obtained asignificant growth of G. margarita hyphae and spore pro-duction in Petri dish cultures using Daucus carota hairyroots and a solidified minimal (M) medium based onWhites medium. Recently, this symbiotic AM fungus andcarrot hairy root co-culture system has been substantiallyimproved in terms of spore production of Glomus intrara-dices using dual-compartment Petri dishes (St-Arnaud et al.,1996). The main interest of mycorrhizal hairy roots as fun-gal host is that they grow faster than nontransformed rootsand do not require exogenous plant growth regulators.Therefore, this system represents an attractive approach forthe large scale production of AM fungi inocula in bioreac-tors under controlled conditions.

Several studies have reported the culture of hairy roots invarious bioreactor configurations for secondary metaboliteproduction (Toivonen 1993). However, none of them in-volved the production of endomycorrhizae in bioreactorsystems. Hairy roots have been grown under various sub-merged (Kondo et al., 1989; Taya et al., 1989; Hilton andRhodes, 1990; Whitney, 1992; McKelvey et al., 1993; Nuu-tila et al., 1995) and mist conditions (sprayed liquid me-dium) (Ramakrishnan et al., 1994; McKelvey et al., 1993;Whitney, 1992; Dilorio et al., 1992; Wilson et al., 1990) aswell as in a rotating drum bioreactor (Kondo et al., 1989).The use of packing to support root growth in a mist biore-actor was shown to be efficient for root aeration and distri-bution (Ramakrishnan et al. 1994) but appears to be prob-lematic during harvesting. Direct mechanical mixing wasfound detrimental to long roots, yielding short viable roots(Takayama and Takizawa, 1994) and increased cell debris(Kondo et al. 1989). Mixing shear is also expected to bedetrimental to the rigid chitin-based cell wall of endomy-corrhizal hyphae.

In this context, the objectives of this study were: first, toevaluate the potential of submerged cultures for the growthof the endomycorrhizal fungus G. intraradices on D. carotahairy roots, and second, to study the performance of anairlift-type bioreactor configuration, previously found suit-able for the growth of hairy roots, for the production of AMfungi propagules. This bioreactor presents interesting scale-up potential and offers flexible operational and culture con-ditions.

MATERIALS AND METHODS

Hairy Root and AM Fungi Cultures

Hairy root cultures of D. carota were obtained as described(Becard and Fortin, 1988) and subcultured monthly bytransfer of ~ 1 g wet weight (WW) (0.06 g dry weight (DW))in 80 mL modified Whites medium (Becard and Fortin,1988) in 250-mL Erlenmeyer shake flasks agitated at 120rpm and maintained at 23 1C in continuous light.

Cultures of the AM fungus G. intraradices with D. carotahairy roots were originally started from spores isolated fromsoil as described by Chabot et al. (1992). Subculture of theroot-fungus pair was performed by transfer of colonizedroot pieces to fresh solid M medium (Becard and Fortin,1988) (20 mL of media solidified with 0.4% Gellan Gum(ICN Biochemicals, Cleveland, OH)) in Petri dishes every23 months. All Petri dishes were incubated in the dark at26 1C.

Bioreactor inocula were prepared in Petri dishes by ad-dition of a 1 1 0.5 cm gel piece containing entrappedspores and hyphae produced in solid cultures, to 0.5 g WW(wet weight) nonmycorrhizal hairy root, obtained in shakeflasks, in 20 mL of liquid M medium for 3 months. Prior tobioreactor inoculation, dishes were examined visually toconfirm colonization (presence of extraradical spores andhyphae). Mycorrhizal roots were harvested without the in-oculum gel piece, cut into ~ 1-cm pieces with a scalpel, andtransfered into the bioreactor. Using the same type of in-oculum, Petri dish liquid cultures were also carried out ascontrol cultures. The spore content of the inocula was of2300 600 spores (g DW)1.

Airlift Bioreactor

The airlift bioreactors were made of 1.2-L total volume (11cm ID 13 cm height) autoclavable polycarbonate jars(Nalgene, Sybron International, Rochester, NY) with amodified cover (Fig. 1). A central glass draught tube (ID 42.54 cm) was installed 2 cm above the bottom and below theliquid level. Medium circulation and oxygenation were pro-vided by a gas flow rate of 5 mL min1 through a porous (2mm) stainless steel sparger which generated fine bubbles atthe bottom of the draught tube. This gas flow rate yielded aninitial oxygen mass transfer coefficient (kLao) of 8 h1. Astainless steel screen mesh (mesh size of 20) was installed 3cm above the bottom to cover the annular section in order toprevent root circulation with the fluid drag. Sterile air filters(bacterial air vent, Gelman Sciences, Ann Arbor, MI) and aliquid condenser ensured sterility and minimal liquid lossesby evaporation throughout the cultures duration. A dis-solved oxygen probe (polarographic: Ingold, Urdorf, Swit-zerland) was installed in the bioreactor as illustrated in Fig.1. Cultures in this bioreactor were performed using an initialvolume of 500 mL of liquid M medium.

Medium and bioreactors were steam sterilized separatelyfor 35 min (121C, 1 bar). All bioreactor cultures were fed

JOLICOEUR ET AL.: PRODUCTION OF GLOMUS INTRARADICES PROPAGULES IN AN AIRLIFT BIOREACTOR 225

with air enriched with 2% CO2. Cultures were grown at 23 1C under continuous light.

Analytical

Sampling and Harvesting

Media from all bioreactor and Petri dish cultures weresampled at harvest. Liquid samples were filtered (0.45 mmand stored at 20C for further analysis. Roots were notsampled during bioreactor cultures but harvested at the endof each experiment. Clinging liquid was removed by plac-ing the roots between absorbent towels (Kimwipes, Kim-berly-Clark, USA) and applying gentle pressure. There wasno damage to the roots. Biomass wet weight and residualliquid medium volume were measured and retained for fur-ther analysis. Fresh root samples were withdrawn for dry(24 h at 80C) weights as well as for level of colonization.

Measurement of Propagule Production

Propagule production was evaluated as spore number andAM fungi density per root length. A fraction of the biomassharvested was used for spore counts. This fraction wasground in a blender (20 s1 for 30 s) in 500 mL H2O.Samples of 20-mL of the resulting suspension were placedin Petri dishes for spore number determination. A secondfraction of the biomass harvested was then used for theevaluation of root colonization using chlorazol black Estaining (Brundrett et al., 1984). Stained roots were ob-served under microscope by 2-mm sections, the optical fieldfor a 20 magnification. The percentage of 2-mm segmentscontaining vesicles and/or intraradical hyphae over the totalsection number observed was determined. The average

number of vesicles per 2-mm segment containing more thanone vesicle was determined as the vesicle density. Viabilityof propagules was verified by culturing harvested sporesand root pieces (~ 1 cm) on solid M medium in Petri dishes.After 3 weeks, a plateau was reached for new fungalgrowth. The percentages of spores and root pieces showingnew external hyphal development were reported as the vi-ability of the propagule sample assayed. The colonizationpotential of these propagules was not verified.

Extracellular Nutrient Analysis

Carbohydrates were analysed by an HPLC system consist-ing of a Waters model 6000A pump, a Gilson model 231/401 automatic injector, a Gilson Model 132 refractive indexdetector, and a Hewlett-Packard Model HP3394A integra-tor. Separation was achieved using a Bio-Rad HPX-87Ccolumn maintained at 80C and water as the mobile phaseflowing at a rate of 1.0 mL min1.

Major ions were analysed using a Dionex (DionexCanada Ltd, Oakville, Canada) HPLC system equipped witha gradient pump, an automated sampler, and a pulsed elec-trochemical detector in conductivity mode, all controlledwith Dionex Al-450 software. Both anions and cations wereanalysed at 23 1C. Anions were separated using a 4 250 mm IONPAC AS4A-SC column, a guard column(IONPAC AG4A-SC), and an anion self-regeneration sup-pressor (ASRS-1) to improve the signal-to-noise ratio. Themobile phase consisted of an aqueous bicarbonate buffer(1.8 mM Na2CO3/1.7 mM NaHCO3) solution flowing at arate of 2.0 mL min1. This allowed separation of all majoranions present in the liquid M medium samples in 10 min.Cations were separated using a 4 250 mM IONPAC CS-12 analytical column, a guard column (IONPAC CG-12Guard column), and a cation self-regenerating supressor(CSRS-1) to reduce signal-to-noise ratio. The mobile phasewas an aqueous methanesulphonic acid (20 mM) solutionflowing at a rate of 1.0 mL min1, which was suitable toseparate the cations within 10 min.

RESULTS

The study of endomycorrhizal symbiotic cultures in biore-actors presented a few major analytical difficulties. For in-stance, the interconnected AM fungus/hairy root networkforming the growing biomass showed high heterogeneity inspore number and hyphae distribution in the root bed and inthe degree of root colonization. Representative sampling ofthe biomass grown in bioreactors, as well as in Petri dishes,could be achieved only with great difficulty without dis-turbing the fungus/root network or dismantling the culturevessel. Consequently, whole cultures were harvested tomeasure biomass growth and propagule production. In thiscontext, Petri dish cultures were carried out in triplicatewhile more complex bioreactor cultures were not system-atically repeated.

A second difficulty involved defining a common basis of

Figure 1. Airlift bioreactor configuration.

226 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 63, NO. 2, APRIL 20, 1999

comparison for production levels between Petri dish andairlift bioreactor submerged culture systems. For these twoculture systems, the culture volume was taken to be theliquid medium volume. Therefore, propagule production re-

sults were normalized to the mycorrhizal roots inoculum (gDW) and initial liquid medium volume. Production resultsare thus expressed as propagule yields with respect to in-oculation conditions.

Liquid Cultures in Petri Dishes

The growth of mycorrhizal hairy roots in Petri dishes couldbe best described as an exponential growth (Fig. 2A) with aspecific growth rate (DW) of 0.021 d1 (Table I). An ex-ponential growth behaviour is proposed with respect to thebranching growth pattern of hairy roots. Most roots re-mained immersed in the liquid phase while a few root tipsgrew above the liquid surface. New fungal hyphae and roottips were observed within days following inoculation. Thequantity of new lateral roots was much lower than normallyseen on solid M medium cultures (not shown). Maximumbiomass was reached 77 d after inoculation which coincidedwith cessation of nitrate and potassium uptake (Fig. 3A) andresulted in a maximum growth index of 2.16 0.43 (TableI) which declined somewhat thereafter. The wet-to-dry (W/D) biomass ratio of 9.85 0.10 was similar for all cultures.No significant difference was observed between the Petridish and airlift bioreactor.

Phosphate was the only measured nutrient observed to bedepleted (within experimental error) from the medium,which occurred within the first 10 d (Fig. 3A). Nitrate andpotassium ions were partially consumed at similar specificuptake rates (Table I).

Sucrose hydrolysis occurred within 24 d from inoculation(Fig. 3B), with glucose being preferentially consumed overfructose. Carbohydrates were not limiting in any culture,including bioreactors. A biomass yield of 0.39 g DW (gcarbohydrates)1 was observed with respect to total carbo-hydrate consumption.

Roots remained mycorrhized throughout the culture du-ration, as 25 to 50% of total root length contained hyphal

Table I. Growth parameters of D. carota hairy roots mycorrhized with the arbuscular mycorrhizalfungus G. intraradices in Petri dish cultures using liquid M medium.*

Inoculuma (g DW ? L1) 0.60 0.11m (d1) 0.021 0.0037d r2 4 0.88Growth indexb (g ? DW ? (g DW)1) 2.16 0.434YX/PO4 (g DW ? mmol1) 25.71 3.88eYX/NO3 (g DW ? mmol1) 0.33 0.026f r2 4 0.87YX/K (g DW ? mmol1) 0.64 0.056f r2 4 0.84YX/S (total)c (g DW ? g1) 0.39 0.016f r2 4 0.84rNO3 (d1) 0.024 0.0017f r2 4 0.98rK (d1) 0.028 0.0031f r2 4 0.94

aMycorrhizal hairy root inoculum dry mass per medium liquid volume.bHarvested to inoculum mycorrhizal hairy root dry mass ratio.cTotal carbohydrates.dCalculated for the 13 to 48 d culture time period.eCalculated as the mycorrhizal root dry weight produced to the initial phosphate content of the

culture medium.fCalculated for the 13 to 48 d culture time period.*Specific rates are the maxima values for the specified culture time period. Errors are 95%

confidence interval values (SE).

Figure 2. Dual culture performances of D. carota hairy roots mycorrhi-zed with G. intraradices in liquid M medium in Petri dish. (A) MycorrhizalD. carota hairy root growth, shown as growth index (harvested dry mass/inoculated dry mass). (B) Normalized spore productions: in spores ? g1DW of inoculum ? L1 medium. Cultures for inocula of 0.60 0.11 gDW ? L1 medium. Error bars indicate standard deviations.

JOLICOEUR ET AL.: PRODUCTION OF GLOMUS INTRARADICES PROPAGULES IN AN AIRLIFT BIOREACTOR 227

structures although a negligible number of vesicles was pre-sent (Table II). Intraradical propagule viability was con-firmed by new hyphae emerging from 33% of harvestedroot pieces when placed on fresh medium. Spore formationfollowed following inoculation. The quantity of new lateralroots was much lower than normally seen on mycorrhizalroot growth. Spore production rate could be described by anexponential law (Fig. 2B) since G. intraradices producesnew spores in parallel to growth and branch formation(Chabot et al., 1992). A specific spore production rate of0.035 d1 was found in the exponential phase. A maximumof 2.51 106 9.86 104 spores (g DW inoculum)1 (Lmedium)1 or 600 230 spores per Petri dish were pro-duced after 83 d (Fig. 2 and Table II). An average of 60%of harvested spores were viable (Table II).

Airlift Bioreactor Cultures

A key factor of this type of dual culture concerns the par-ticular growth and proliferation patterns of each symbioticpartner with respect to the objective of the bioprocess underdevelopment, namely the production of viable endomycor-rhizal propagules (spores, hyphae, and intraradicalvesicles). In this context, bioreactor cultures were carriedout to determine the best inoculation conditions favouringboth mycorrhizal hairy root growth and AM fungi prop-agule production. Hairy root inoculum-to-medium volumeratios from 0.13 to 1.51 g DW (L medium)1 were assessed.This inoculum range was conditioned by the minimal me-dium used as discussed below. The duration of the culturesin the airlift bioreactor was determined from visual obser-vation of the roots and the culture medium. Cultures were

Figure 3. Nutrient consumption in Petri dish liquid cultures for an in-oculum of 0.60 0.11 g DW ? L1 medium. (A) Major ions: (j) nitrate;(s) potassium; (m) phosphate. (B) Carbohydrates: (h) fructose; (d) glu-cose; (n) sucrose. Each point is the average of three cultures. Error barsindicate standard deviations.

Table II. Comparative G. intraradices propagules production process performances.*

Culture method Host plant

Spore production Root mycorrhization

Time(weeks) References#/Plant % Germination % Colonization

% Withviable fungus

Aeroponic Bahia grass NA NA 50% NA 12 Sylvia and Hub-bel (1986)

Aeroponic Bahia grass 630 NA 1728% 1214 Hung and Sylvia(1988)

Aeroponic Sweet potato 4500 NA 21% NA 1214

Pot (sand) Linseed NA NA 75% NA 13 Dugassa et al.(1995)

Hydroponic NA NA 80% NA 13

#/L of M mediumDual compartment

Petri dishaCarrot hairy roots 500,000 67% NA NA 16 St-Arnaud et al.

(1996)Petri dish (liquid)b Carrot hairy roots 30,000 11,500 60% 2550% 33%d 12 This workAirlift bioreactorc Carrot hairy roots 12,400 800 58% 2575% 30%d 12 This work

aActual spore production in a two-compartment Petri dish as described by St-Arnaud et al. (1996).bActual spore production in a standard Petri dish filled with 0.02 L liquid M medium and 0.015 g DW inoculum.cActual spore production in a 1-L airlift bioreactor filled with 0.5 L liquid M medium and 0.065 g DW inoculum.dPercentage of 1 cm pieces of mycorrhiza producing hyphal growth on subculture.*NA, non-available. Error data are 95% confidence interval values.

228 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 63, NO. 2, APRIL 20, 1999

harvested two weeks after the roots and the liquid mediumcolours turned to brown. To our knowledge, no moregrowth for both the AM fungus and the roots occured fromthat point (not published). Cultures were harvested after 83to 107 d.

The bioreactor cultures performed using inocula from0.13 to 0.61 g DW L1 mycorrhizal D. carota hairy rootsgrew well with growth indices from 9.8 to 2.0, respectively,after 107 and 83 d (Fig. 4A). Roots growth showed fewerlateral branches than found in Petri dish liquid cultures.New root apices were well distributed throughout the liquidvolume and even protruded from the liquid surface. Appar-ently, the downward medium circulation in the annulus re-gion induced by gas bubbling failed to counterbalance thetransformed roots negative geotropism. Dissolved oxygenconcentration remained over 80% of air saturation for allairlift cultures.

An increase in the inoculum to 1.51 g DW L1 mediumresulted in a decrease of the growth index to 1.28 after 83 d.A critical inoculum concentration of ~ 0.6 g DW L1 me-dium was thus observed, above which the growth of my-corrhizal hairy roots was significantly limited (Fig. 4A).This inoculum ratio range was termed the high inoculum

range. Lowering the inoculum-to-medium ratio improvedbiomass growth indices (harvested root DW (inoculumDW)1).

Normalized (Fig. 4B) and actual spore productions (Fig.4C) were different with respect to inoculum-to-medium ra-tio. Normalized spore productions were found to be in-versely proportional to the inoculum-to-medium ratio. Lowinocula conditions yielded 1.82 105 2.86 104 (SEM)spores produced (g DW inoculum)1 (L medium)1 as com-pared to a value of 2.06 104 8.63 103 (SEM) for thehigh inoculum range.

Conversely, average actual productions seemed to be in-dependent of inocula conditions with 5.03 103 4.26 103 (SD). However, spore productions obtained in the airliftbioreactor were significantly lower than Petri dish cultureswith (2.51 106) (9.86 104) (SEM) spores (g DWinoculum)1 (L medium)1 (Fig. 3B). The biomass yield onphosphate for low inocula was higher than for Petri dishcultures with 33.6 2.5 g DW/mM PO4 as compared to25.71 3.88 g DW/mmol PO4 (Table I) while decreasing to15 g DW/mM1 PO4 for an inoculum of 1.51 g DW L1.

The roots remained mycorrhized within 2575% of theirlength (Table II) showing negligible vesicle content. Thequality of the propagules produced was similar to thosefrom Petri dishes with a spore germination level of 58%,and 30% of harvested root pieces showed new hyphalgrowth when recultured on fresh medium.

DISCUSSION

Growth of Mycorrhizal Hairy Roots inSubmerged Cultures

This work focused on the study of the performance of anairlift bioreactor, which could be scaled up to industrial size,for the production of endomycorrhizal propagules. Cultureconditions in the airlift system resulted in the growth of alow density root bed that did not cause medium channellingand stagnation, as observed by McKelvey et al. (1993) forcompact root beds, nor favoured, unfortunately, high root-to-root contact. Mycorrhizal hairy roots grew well, follow-ing an inverse relationship between the growth index andthe inoculum concentration (Fig. 4A). However, problemsof nonhomogeneous medium distribution with high root beddensities leading to insufficient nutrient renewal zones havebeen shown to occur in bioreactor cultures (Whitney, 1992).An important culture parameter is thus the nutrient refresh-ment rate to the root biomass (Dilorio et al., 1992) in orderto support the requirements of metabolic activity. However,the similar growth indices obtained in Petri dish and airliftbioreactor for similar inocula suggests that this problem didnot occur. Nonhomogeneous medium distribution may alsobe involved in the growth inhibition of high inocula cul-tures, but the suddenness of the effect (Fig. 4A) is not likelyto be due to hydrodynamic and mass transfer limitations.Plausible reasons for this sudden cessation are insufficient

Figure 4. Effect of inoculating conditions on symbiont growth in sub-merged cultures. (A) Mycorrhizal D. carota hairy root growth index (asdescribed in Fig. 2). (B) Normalized spore production (as described in Fig.2). (C) Actual spore production per culture volume: (h) Petri dish cultures;(m) arlift bioreactor cultures. Error bars are 95% confidence interval val-ues.

JOLICOEUR ET AL.: PRODUCTION OF GLOMUS INTRARADICES PROPAGULES IN AN AIRLIFT BIOREACTOR 229

levels of nutrients to favour growth or a possible inhibitorylevel of root exudates under high inocula conditions. It iswell known that a minimal root inoculum mass is requiredto initiate new cultures (Mosse and Hepper, 1975), but,unfortunately, no studies have been reported on the effect ofvery high levels of inocula.

The specific biomass growth rate observed in this studywas significantly lower than that reported by Taya et al.(1989) (0.24 d1) for nonmycorrhizal D. carota hairy rootcultures in an airlift bioreactor using MS medium (3% su-crose). This can be explained by the use of different D.carota root lines and medium as well as the effect of thefungus on root growth. The lower sugar and salt concentra-tions of the M medium have been shown to lower thegrowth rate of strawberry hairy roots in shake flask culturesfrom 0.16 to 0.05 d1 (calculated from Nuutila et al., 1995).

Mycorrhizal Hairy Roots Nutrition

The low-salt M medium was previously developed by Be-card and Fortin (1988) for hairy root-AM fungi co-culture.The authors clearly showed the requirement of a low phos-phate concentration (0.04 mM) for the arbuscular fungi G.margarita to colonize carrot hairy roots. This medium al-lowed both symbionts to grow in Petri dish and airlift bio-reactor submerged cultures. However, the culture systemsand the use of mycorrhizal hairy roots in this study pre-vented a separate characterization of the major nutrient re-quirements and uptakes of each symbiotic organism sinceboth grew simultaneously with intracellular nutrient ex-changes between symbionts. Therefore, the uptake rates andbiomass yields discussed in this study are overall values.Use of overall values may explain the high variability in thespecific uptake rates of nitrate and potassium which aregenerally related to growth (Lee, 1982; Lefebvre and Glass,1982; Clarkson and Hanson, 1980). Moreover, Petri dishcultures showed a lower phosphate (23%) to biomass yieldthan that observed for airlift cultures (Table I). Biomassyield with respect to total carbohydrates presented a lowervalue (0.39 g DW/g) for Petri dish cultures using mycor-rhizal hairy roots than the generally accepted value of 0.5 gDW/(g carbohydrates) for a variety of uncolonized hairyroot lines regardless of the culture system and medium (Uo-zumi and Kobayashi, 1994; Uozumi et al., 1993). Theseresults may also suggest a more important maintenance con-tribution relative to global nutrition at higher initial biomassconcentrations. This is currently under investigation.

Complete uptake (within experimental error) of phos-phate before the end of biomass growth was expected sincethis has already been reported for strawberry hairy root(Nuutila et al., 1995) as well as undifferentiated plant cellsuspensions cultures (Jolicoeur et al., 1992). Subsequentgrowth is driven from cytosolic and vacuolar phosphatestores (plants: Ashihara and Tokoro, 1985; Bieleski, 1973;fungi: Harrison and van Buuren, 1995; Beever and Burns,1980; endomycorrhizal hairy roots: Jolicoeur, 1998). Rootgrowth and propagule production stopped at the depletion of

a limiting nutrient, likely intracellular phosphate, as evi-denced by simultaneous cessation of nitrate, potassium andcarbohydrate uptake.

AM Fungi Propagule Production

The submerged culture conditions imposed by the Petri dishand airlift bioreactor systems used in this study were notdetrimental to the mycorrhizal state of inoculated mycor-rhizal D. carota hairy roots. Root colonization was con-firmed by staining and growth of external hyphae from rootpieces cultured on fresh medium. Airlift cultures yieldedroot colonization increases, from 1050% for the inoculatedbiomass to final levels of 2575%, which approach thoseobtained (7580%) using linseed as host plant for pot andhydroponic cultures (Dugassa et al., 1995) (Table II). Fur-thermore, the intraradical propagules produced in the Petridishes were of the same quality (viability) as those fromairlift bioreactors with 33% and 30% of the harvested rootpieces showing new hyphal growth, respectively. Moreover,internal fungal growth was mainly limited to hyphal andarbuscular structures with few vesicles distributed in 510%of total root length. Thus, submerged cultures (Petri dishand airlift bioreactor) showed no significant increases inintraradical fungal vesicle levels as compared to inocula(Table II), which suggest that this condition did not stimu-late the AM fungus G. intraradices to produce vesicles incarrot hairy roots.

Harvested spores in Petri dish and airlift bioreactors wereall of the same quality, with germination rates of 58 and60%, respectively. However, the significant difference ob-served in production levels between these two culture sys-tems may be related to their respective culture conditionswhich mainly differed in hydrodynamic conditions, root-to-root contact level and gas phase renewal rate. In the case ofthe airlift system, continuous liquid circulation may havebeen inhibitory to fungal growth as this organism usuallygrows under static conditions in soil. Furthermore, in highinocula culture conditions, the roots may have producedexcessive exudates which are suspected to adversely affectspore formation (St-Arnaud et al., 1996). Similarities be-tween aseptic culture conditions and the symbionts naturalniche, such as soil water content, may also be importantparameters to consider. In this study, carrot roots weregrown under submerged culture conditions which differfrom field conditions.

During this study, significant variations in spore produc-tion and colonization levels were observed for Petri dish andairlift bioreactor cultures (Fig. 4B and C). Similar resultshave been reported for the mass production of G. intrara-dices inocula using entire plants (Table II). The heteroge-neity of mycorrhizal root inocula may be involved in thisphenomenon. However, the use of higher inocula levels inbioreactors, which should have decreased inoculum vari-ability between cultures failed to overcome this problem ofproduction variability (Fig. 4B and C). These results show-ing high variability in propagules production may indicate

230 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 63, NO. 2, APRIL 20, 1999

the requirement of exhaustive studies of the effect of im-portant parameters such as root age, colonization level andfungal mass on the inocula quality. Finally, another impor-tant parameter that needs to be considered is host plant/fungal compatibility (Gianinazzi-Pearson, 1984; Smith,1980). This is evidenced by differences in spore productionand colonization levels between sweet potato and bahiagrass in aeroponic culture (Table II) suggesting that carrothairy roots may not be the optimal host plant to use for G.intraradices propagule production. However, the symbiontswere studied in this work as they showed high spore pro-duction in dual compartment Petri dishes (St-Arnaud et al.,1996).

In comparing various technologies for spore production,Petri dish cultures appear to be a simple and efficient solu-tion for laboratory-scale production levels. The fungal ma-terial is easy to harvest and free of contaminants. In thiscontext, the dual compartment Petri dish system (St-Arnaudet al. 1996) represents a suitable approach. This simple sys-tem displayed spore production of more than ten times thoseobserved for the Petri dish cultures using liquid M mediumdiscussed in this study. For industrial scale production,however, these small systems clearly will be insufficient tomeet world market demand at reasonable prices. Resultsfrom this work showed that this co-culture technology canalso be carried out under submerged culture conditions us-ing conventional airlift bioreactor configuration which al-lowed significant fungal propagule production. To ourknowledge, this is the first time that a second generationproduction process of endomycorrhizal propagules is pre-sented. The airlift bioreactor offer valuable scale-up poten-tial that can eventually be considered for large industrialscale production. However, more work has to be done onthe culture conditions.

The authors thank Bio-Capital Inc. and Premier Tech Ltd fortheir financial help. Particularly, we thank M. J. Santoire, L.Nantais, and L. Lavoie for their technical help and A. Bitton forreviewing this document.

REFERENCES

Ashihara H, Tokoro T. 1985. Metabolic fate of inorganic phosphate ab-sorbed by suspension cultured cells of Catharanthus roseus. J PlantPhysiol 118:227233.

Andrade G, Mihara KL, Linderman RG, Bethlenfalvay GJ. 1997. Bacterialfrom rhizosphere and hydrosphere soils of different arbuscularmy-corrhizal fungi. Plant Soil 192:7179.

Arora DK, Rai B, Mukerji KG, Knudsen GR. 1991. Handbook of AppliedMycology. Vol 1: Soil and Plants. New York: Marcel Dekker.

Becard G, Fortin JA. 1988. Early events of vesiculararbuscular mycor-rhiza formation on Ri-T-DNA transformed roots. New Phytol 108:211218.

Becard G, Piche Y. 1989. Fungal growth stimulation by CO2 and rootexudates in vesicular-arbuscular mycorrhizal symbiosis. Appl EnvironMicrobiol 55:23202325.

Beever RE, Burns DJW. 1980. Phosphorus uptake, storage and utilizationby fungi. In: Woolhouse HW, editor. Advances in Botanical Research.London, UK: Academic Press. Vol 8, p 127219.

Berta G, Fusconi A, Trotta A, Scannerini S. 1990. Morphogenetic modi-

fications induced by the mycorrhizal fungus Glomus strain E3 in theroot system of Allium porrum L. New Phytol 114:207215.

Bethlenfalvay GJ. 1992. Mycorrhizae and Crop Productivity. 1. In: My-corrhizae in Sustainable Agriculture. Am Soc Agric, Spec Publ 54:127.

Bieleski RL. 1973. Phosphate pools, phosphate transport, and phosphateavailability. Annu Rev Plant Physiol 24:225252.

Brundrett MC, Piche Y, Peterson RL. 1984. A new method for observingthe morphology of vesiculararbuscular mycorrhizae. Can J Bot 62:21282134.

Chabot S, Becard G, Piche Y. 1992. Life cycle of Glomus intraradices inroot organ culture. Mycologia 84:315321.

Clarkson DT, Hanson JB. 1980. The mineral nutrition of higher plants.Annu Rev Physiol 31:239298.

Dilorio AA, Cheetham RD, Weathers PJ. 1992. Growth of transformedroots in a nutrient mist bioreactor: Reactor performance and evalua-tion. Appl Microbiol Biotechnol 37:457462.

Dugassa DG, Grunewaldt-Stocker G, Schonbeck F. 1995. Growth of Glo-mus intraradices and its effect on linseed (Linum usitatissimum L.) inhydroponic culture. Mycorrhiza 5:279282.

Gianinazzi-Pearson V. 1984. Hostfungus specificity, recognition andcompatibility in mycorrhizae. In: Verma DPS, Hohn TH, editors. PlantGene Research: Genes involved in microbeplant interactions. NewYork: Springer-Verlag. p 225253.

Hall IR. 1978. Soil pellets to introduce vesiculararbuscular mycorrhizalfungi into soil. Soil Biol Biochem 11:8586.

Harrison MJ, van Buuren M. 1995. A phosphate transporter from themycorrhizal fungus Glomus versiforme. Nature 378:626629.

Hilton MG, Rhodes MJC. 1990. Growth and hyoscyamine production ofhairy roots cultures of Datura stramonium in a modified stirred tankreactor. Appl Microbiol Biotechnol 33:132138.

Hung LLL, Sylvia DM. 1988. Production of vesiculararbuscular mycor-rhizal fungus inoculum in aeroponic culture. Appl Environ Microbiol54:353357.

Jolicoeur M. 1998. Optimisation dun procede de production dechampignons endomycorhiziens en bioreacteur. PhD thesis, ecolePolytechnique de Montreal, Montreal, Quebec, Canada.

Jolicoeur M, Chavarie C, Carreau PJ, Archambault J. 1992. Developmentof a helical-ribbon impeller bioreactor for high-density plant cell sus-pension culture. Biotechnol Bioeng 39:511521.

Khasa P, Furlan V, Fortin JA. 1990. Effets de differentes espe`ces dechampignons endomycorhiziens sur la croissance de dix espe`ces deplantes tropicales au Zare. Tropicultura 8:159164.

Kondo O, Honda H, Taya M, Kobayashi T. 1989. Comparison of growthproperties of carrot hairy root in various bioreactors. Appl MicrobiolBiotechnol 32:292294.

Lee RB. 1982. Selectivity and kinetics of ion uptake by barley plantsfollowing nutrient deficiency. Ann Bot 50:429449.

Lefebvre DD, Glass ADM. 1982. Regulation of phosphate influx in barleyroots: Effects of phosphate deprivation and reduction of influx withprovision of orthophosphate. Physiol Plant 54:199206.

McKelvey SA, Gehrig JA, Hollar KA, Curtis WR. 1993. Growth of plantroot cultures in liquid and gas-dispersed reactor environments. Bio-technol Prog 9:317322.

Mosse B, Hepper C. 1975. Vesiculararbuscular mycorrhizal infections inroot organ cultures. Physiological Plant Pathol 5:215223.

Mosse B, Thompson JP. 1984. Vesiculararbuscular endomycorrhizal in-oculum production. I. Exploratory experiments with beans (Phaseolusvulgaris) in nutrient flow culture. Can J Bot 62:15231530.

Mugnier J, Mosse B. 1987. Vesiculararbuscular mycorrhizal infections intransformed Ri T-DNA roots grown axenically. Phytopathol 77:10451050.

Newsham KK, Fitter AH, Watkinson AR. 1995. Multi-functionnality andbiodiversity in arbuscular mycorrhizas. TREE 10:407411.

Nuutila AM, Vestberg M, Kauppinen V. 1995. Infection of hairy roots ofstrawberry (Fragaria Ananassa Duch.) with arbuscular mycorrhizalfungus. Plant Cell Rep 14:505509.

Pfeffer PE, Shachar-Hill Y, Becard G, Rolin D, Douds DD Jr. 1998.

JOLICOEUR ET AL.: PRODUCTION OF GLOMUS INTRARADICES PROPAGULES IN AN AIRLIFT BIOREACTOR 231

Nutrient transport and metabolism in the life cycle of arbuscular my-corrhizae as examined by NMR spectroscopy. In: Flores HE, LynchJP, Eissenstat D, editors. Radical Biology. Rockville, MD: AmericanSociety of Plant Physiologists. Vol 18, p 178187.

Ramakrishnan D, Salim J, Curtis WR. 1994. Inoculation and tissue distri-bution in pilot-scale plant root culture bioreactors. Biotechnol Tech8:639644.

St-Arnaud M, Hamel C, Caron M, Fortin JA. 1994. Inhibition of Pythiumultimum in roots and growth substrate of mycorrhizal Tagetas patulacolonized with Glomus intraradices. Can J Plant Pathol 16:187194.

St-Arnaud M, Hamel C, Vimard B, Caron M, Fortin JA. 1995. Alteredgrowth of Fusarium oxysporum f. sp. chrysanthemi in an in vitro dualculture system with the vesicular arbuscular mycorrhizal fungus Glo-mus intraradices growing on Daucus carota transformed roots. My-corrhiza 5:431438.

St-Arnaud M, Hamel C, Fortin JA. 1996. Enhanced hyphal growth andspore production of the arbuscular mycorrhizal fungus Glomus intra-radices in an in vitro system in the absence of host roots. Mycol Res100:332338.

St-Arnaud M, Hamel C, Vimard B, Caron M, Fortin JA. 1997. Inhibitionof Fusarium oxysporum f. sp. dianthi in the non-VAM species Dian-thus caryophyllus by co-culture with Tagetes patula compagnionplants colonized by Glomus intraradices. Can J Bot 75:9981005.

Simon L, Bousquet J, Levesque R, Lalonde M. 1993. Origin and diversi-fication of endomycorrhizal fungi and coincidence with vascular landplants. Nature 363:6769.

Smith SSE. 1980. Mycorrhizas of autotrophic higher plants. Biol Rev55:475510.

Sylvia DM, Hubbell DH. 1986. Growth and sporulation of vesicu-lararbuscular mycorrhizal fungi in aeroponic and membrane sys-tems. Symbiosis 1:259267.

Sylvia DM, Jarstfer AG. 1992. Sheared roots inocula of vesicular mycor-rhizal fungi. Appl Environ Microbiol 58:229232.

Takizawa N, Takayama S. 1994. The method and culture system for largescale culture of plant roots using aerationagitation bioreactor: Strat-

egies to scale-up. In: Abstracts of the VIIIth International Congress ofPlant Tissue and Cell Culture, Florence, Italy. p 263.

Taya M, Yoyoma A, Kondo O, Kobayashi T, Matsui C. 1989. Growthcharacteristics of plant hairy roots and their cultures in bioreactors. JChem Eng Jpn 22:8488.

Tepfer DA, Tempe J. 1981. Production dagropine par des racines formeessous laction dAgrobacterium rhizogenes, souche A4. C R Acad ScParis 292:153156.

Toivonen L. 1993. Utilization of hairy root cultures for production ofsecondary metabolites. Biotechnol Prog 9:1220.

Trappe JM, Molina R, Castellano M. 1984. Reactions of mycorrhizal fungiand mycorrhiza formation to pesticides. Annu Rev Phytopathol 22:331359.

Uozumi N, Kobayashi T. 1994. Application of Hairy Root and Bioreactors.In: Ryu DDY, Furusaki S, editors. Advances in Plant Biotechnology.New York: Elsevier. p 307337.

Uozumi N, Kohketsu K, Kobayashi T. 1993. Growth and kinetic param-eters of Ajuga hairy root in fed-batch culture on monosaccharide me-dium. J Chem Tech Biotechnol 57:155161.

Whitney PJ. 1992. Novel bioreactors for the growth of roots transformedby Agrobacterium rhizogenes. Enzyme Microbiol Technol 14:1317.

Wilson PDG, Hilton MG, Meehan PTH, Waspe CR, Rhodes MJC. 1990.The cultivation of transformed roots from laboratory to pilot plant:Current plant science and biotechnology in agriculture. In: NijkampHJJ, Van der Plas LHW, Van Aartrijk J, editors. Progress in PlantCellular and Molecular Biology. Dordrecht, The Netherlands: KluwerAcademic Publishers. p 700705.

Wilson JM. 1984. Comparative development of infection by three vesicu-lararbuscular mycorrhizal fungi. New Phytol 97:413426.

Wood T, Cummings B. 1992. Biotechnology and the future of VAM com-mercialization. In: Mycorrhizae in Sustainable Agriculture. Am SocAgric, Spec Publ 54:468487.

Zobel RW, Dei Tredici P, Torren JG. 1976. Method for growing plantsaeroponically. Plant Physiol 57:344346.

232 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 63, NO. 2, APRIL 20, 1999