Levels and identities of nonrhizobial microorganisms found in commercial legume inoculant made with nonsterile peat carrier

Perry E. Olsen, Wendell A. Rice, Lucien M. Bordeleau, A.H. Demidoff, and Mandy M. Collins

Abstract: Sixty samples of commercial North American legume inoculants manufactured for sale in 1994 using nonsterile peat as carrier were tested for Rhizobium (or Bradyrhizobium) content and nonRhizobium biological contaminant load. Products of three major producers of such inoculants for sale in Canada were examined. Viable Rhizobium content varied from 5.6 x lo5 to 8.1 x 109 cellslg, while the contaminant load varied from 1.8 x lo8 to 5.5 x 1010 cfulg. Most of the inoculants contained more nonrhizobial organisms than they did rhizobia. Identifications were made of the most numerous nonrhizobial bacteria occurring in 100 samples of inoculants collected in 1993 and 1994. The most commonly identified contaminant was Xanthomonas maltophilia. Pseudomonas aeruginosa, Klebsiella pneumoniae, and Enterobacter cloacae were also found at high levels in some products. Contaminant organisms capable of inhibiting rhizobial growth in plate culture were found in the products of all three manufacturers.

Key words: Rhizobium, contaminant, inoculant.

Resume : Soixante Cchantillons d'inoculants du commerce pour ICgumineuses de 1'Arnbique du Nord, fabriquCs pour la vente en 1994 a I'aide de tourbes non-stCriles comme sources d'inoculum, ont CtC test& pour leur teneur en Rhizobium (ou Bradyrhizobiurn) et pour la prksence de contaminants biologiques non-rhizobiens. Les produits de trois fabricants majeurs de tels inoculants vendus au Canada ont fait l'objet d'analyses. La teneur en bactCries rhizobiennes viables a fluctuC de 5,6 x lo5 2 8,l x 109 celluleslg et les contaminants ont variC de 1,8 x 108 B 5,s x 101° ufclg. La majorit6 des inoculants contenaient plus d'organismes non-rhizobiens que rhizobiens. La plupart des bactCries non-rhizobiennes prksentes dans 100 Cchantillons d'inoculants prClevCs en 1993 et 1994 ont kt6 identifi6es. Le contaminant le plus commun a kt6 le Xanthomonas maltophilia. Dans certains des produits, les bactkries Pseudomonas aeruginosa, Klehsiella pneumoniae et Enterobacter cloacae ont aussi CtC trouvCes h des niveaux ClevCs. Les produits des trois manufacturiers contenaient des organismes contaminants capables d'inhiber la croissance des rhizobiums cultivCs sur plaques.

Mots cle's : Rhizobium, contaminant, inoculant. [Traduit par la rCdaction]

Inoculation of legume crops with nitrogen-fixing Rhizobium or Bradyrhizobium bacteria is a common agronomic practice designed to ensure and maximize biological nitrogen fixation. Such inoculation involves the placing of nodule-forming bac- teria on or near the seed at the time of planting. A variety of rhizobial carriers are used to facilitate seed inoculation. These include liquid formulations, dry clay powder, and powdered or granular peat. Rhizobial powdered peat inoculants are prepared using either sterilized peat or nonsterile peat as the carrier.

I Received June 9, 1995. Revision received September 29, 1995. Accepted October 3, 1995.

P.E. Olsenl and W.A. Rice. Agriculture and Agri-Food Canada, Northern Agriculture Research Centre, Beaverlodge, AB T O H OCO, Canada. L.M. Bordeleau. Agriculture and Agri-Food Canada, Soils and Crops Research and Development Centre, Sainte-Foy, QC G l V 2J3, Canada. A.H. Demidoff and M.M. Collins. Agriculture and Agr-Food Canada, Northern Agriculture Research Centre, Beaverlodge, AB T O H OCO, Canada.

Author to whom all correspondence should be addressed.

Much of North America's rhizobial inoculant is prepared by the addition of rhizobial cells grown in sterile nutrient broth to nonsterile peat powder. This results in a moist (40-45% water by weight) peat, which is then normally incubated at 25-30°C for 5-21 days to allow development of maximum rhizobia cell numbers. Inoculants made in this way have long been known to carry a considerable viable nonrhizobial contaminant load, but the contaminating organisms have been generally ignored. We have previously shown that most inoculants prepared with nonsterile peat contain more contaminant organisms than they do the desired rhizobia (Olsen et al. 1995). In the case of one lot of clover inoculant sold nationwide in Canada and the United States in 1993, the dominant contaminant was both an opportunistic human pathogen and actively inhibitory, in plate culture, to growth of the rhizobial species the inoculant was intended to carry. Rhizobia in this inoculant were undetectable using the official Canadian regulatory most-probable-number (MPN) test protocol (fewer than 1.0 x 105 rhizobia cellslg).

The objectives of the work reported here were to further define the degree to which high levels of nonrhizobial organ- isms are found in commercial rhizobial inoculant and to gen- erate some understanding of the types of organisms commonly present.

Can. J. Microbiol. 42: 72-75 (1996). Printed in Canada / IrnprimC au Canada

Can

. J. M

icro

biol

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

Sim

on F

rase

r U

nive

rsity

on

11/1

4/14

For

pers

onal

use

onl

y.

I Notes

I I Packages of commercial inoculant were routinely collected

in Canada in 1994 as part of the Agriculture and Agri-Food Canada Legume Inoculant and Pre-Inoculated Seed Product

1 Testing Program, which conducts regulatory inoculant analysis at the Agriculture and Agri-Food Canada Research Centres in Beaverlodge, Alta., and ~ainte-FO~, Que. All inoculants were tested for both Rhizobium (or Bradyrhizobium) and contami- nant viable cell numbers before the expiry dates given on the packages. The inoculants tested were all moist peat powder products manufactured using nonsterile peat carrier. The inocu- lants were manufactured by the major sellers of such inoculant in Canada: LiphaTech Inc., Milwaukee, Wis.; MicroBio RhizoGen Corp., Saskatoon, Sask., and Urbana Laboratories, Research Seeds Inc., St. Joseph, Mo. The products were vari- ously intended for the inoculation of alfalfa (Medicago spp.), peas (Pisum sativum), birdsfoot trefoil (Lotus corniculatus), lentil (Lens culinaris), beans (Phaseolus vulgaris), clovers (Trifolium spp.), or soybean (Glycine max).

All inoculants were evaluated for viable rhizobial count using either the MPN plant nodulation grow-out test or by colony immunoblot analysis (Anonymous 1991). Contaminant counts were obtained by aerobic plate count on yeast extract - mannitol - agar (YEMA) (Vincent 1970), tryptic soy agar (TSA) (Difco), or TSA containing 5% sheep's blood. Counts were made following 72-96 h incubation at 30 or 35OC.

Three basic contaminant types were observed: bacteria, actinomycetes, and fungi. Identification was attempted for contaminants isolated from samples collected in 1993 (Olsen et al. 1994) and in 1994 for this study. Contaminant bacterial colonies were isolated by a single colony pick from growth on TSA medium. Isolates were preserved by freezing aliquot of tryptic soy broth (TSB) culture with 10% glycerol at -135°C. Owing to limits on time and expense, identification work was focused on bacterial contaminants. No attempt was made at actinomycete identification and only limited attempts were made at fungal identification. Bacterial contaminants were identified using the Analytical Profile Index (API) (bioMerieux Vitek, Inc., Hazelwood, Mo.) and (or) the Biolog Microstation System (Release 3.50) (Biolog Inc., Hayward, Calif.). Gram, flagella, and spore stains, colony and microscopic morphology examinations, and motility, catalase, and oxidase tests were also performed. Our isolate identifications of Pseudomonas aeruginosa, Enterobacter cloacae, Klebsiella pneumoniae, and Klebsiella oxytoca were confirmed by a regional hospital medical microbiology laboratory (Queen Elizabeth I1 Hospital, Grande Prairie, Alta.) and other isolate identifications were confirmed by the Agriculture and Agri-Food Canada, Central Plant Health Laboratory, Ottawa, Ont.

Antagonism between bacterial contaminant isolates and rhi- zobial strains was determined by growth of the contaminant on a rhizobial lawn. Rhizobium meliloti (NRG185 and NRG34) and Rhizobium leguminosarum bv. trifolii (NRG405) strains were grown in broth culture for 24-72 h in YEMB, added (1 pL/mL) to molten YEMA at 39OC, and poured into petri plates. Following solidification of the seeded agar plates, con- taminant isolates were inoculated onto the agar surface at the centre of the rhizobia-seeded plates. Plates were then incubated for 48-72 h at 28°C and observed for zones of inhibition of rhizobial growth around the growing contaminant colony. Bacterial contaminants that inhibited growth of one or more species of Rhizobium were isolable from products manufac-

tured by each of the three companies. One of the isolated antagonistic organisms (Pseudomonas aeruginosa) strongly inhibited growth of the R. leguminosarum bv. trifolii strain tested but had no apparent effect on the two R. meliloti strains tested. No attempt was made to detect rhizobial antagonism by actinomycetes, although variable reactions have been reported. Panthier et al. (1979) and Pugashetti et al. (1982) reported that many commonly found soil actinomycetes demonstrate antagonism to rhizobial growth, while Antoun et al. (1978) found that some actinomycetes can exercise biological control of legume disease without affecting rhizobia.

Viable Rhizobium and contaminant count results are shown in Table 1 for each of the 60 inoculants tested. All of the samples contained at least 108 contaminant cfu/g. Twelve percent of the samples contained between lOsand lo9 contaminant cfulg, 80% contained greater than lo9 contaminant cfulg, and 8% contained greater than 101° contaminant cfu/g. While the num- ber of contaminants found was consistently high (1.8 x 108 - 5.5 x 101° cfu/g), the rhizobial content ranged from low to high (5.6 x 105 - 8.1 x 109 cellslg). These results confirm our pre- vious observations (Olsen et al. 1994).

The number of obviously different (colony morphology or colour) types of bacterial contaminants seen from individual inoculant samples plated at the 10-7 dilution ranged from 2 to 11. Overall numbers of contaminants by type was in the following order: bacteria (109- 1010 cfu/g), actinornycetes (107- 109 cfu/g), and fungi (10" lo7 cfu/g). The most common fungal isolates were identified as species of Penicillium. Other isolates were identified as species of Alternaria, Fusarium, and Pythium. All of these genera contain potentially plant patho- genic species or strains. Preliminary pathogenicity tests with alfalfa showed that 11 of 30 different fungal isolates caused some degree of root necrosis. No attempt was made to evaluate rhizobial antagonism with the fungal isolates, although nega- tive effects on rhizobial growth by fungi have been reported (Pugashetti et al. 1982).

Many of the isolated bacterial contaminants were not confi- dently identifiable using either the Biolog or API systems. Only about 30% of the isolated contaminants were confidently iden- tified and some of these were identified only to the genus level. Several contaminants occurring at more than lo6 cfu/g were identified and confirmed (Table 2). The most commonly recov- ered and identified contaminant from these commercial inocu- lant samples was Xanthomonas maltophilia.

The potentially negative effects posed by the presence of high levels of contaminating organisms in legume inoculants made with nonsterile carrier have been previously described (Olsen et al. 1995). These include the possibilities of patho- genicity (to humans, animals, plants, or rhizobia), reduced inoculant effectiveness, and difficulty in applying rapid qual- ity control procedures to the products. Several countries (Australia, Brazil, France, England) have policies relating to permissible levels of contaminants in legume inoculants (Olsen et al. 1994). France enforces a policy that no contaminants be present in legume inoculant sold there (Catroux 1991). One of the reasons given for this policy is the possibility of the pres- ence of human or plant pathogens (Catroux and Amarger 1992). The Canadian government agency responsible for regulation of legume inoculants also has regulations relating to contaminants in rhizobial legume inoculants sold in Canada. These regula- tions state that "Any contaminant organisms present in the

Can

. J. M

icro

biol

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

Sim

on F

rase

r U

nive

rsity

on

11/1

4/14

For

pers

onal

use

onl

y.

Can. J. Microbiol. Vol. 42, 1996

Table 1. Sample identification number, viable rhizobium cell number, and contaminant cell number for 60 samples of commercial inoculants collected in 1994.

Rhizobial Contaminant cells/g cfulg

IDa inoculant inoculant

Company A

B94-10 2.3~109 2 . 0 ~ 1 0 ~ B94-34 1.9~109 4.3x109 B94-41 2.2~107 7.4x109 B94-57 1.9x109 3.8~109 B94-59 3.0~109 3.3~109 B94-7 1 8.1~109 4.1 x109 B94-104 7.3~108 7.3x109 B94- 122 2.6x108 5.0x108 B94-148 8.2x107 1.4~109 Q94-3 1 >2.0xl09 7.0~109 Q94-37 >2.0xl09 7.1 xi09 494-39 1 . 1 ~ 1 0 ~ 2.9x109 494-40 >2.0x109 1.4x1010 Q94-43 >2.0xl09 7.8x109 494-64 7.5x108 6.1 xlO9 494-90 >2.0xl09 5.6x109 Q94-91 >2.0xl09 4.3x109 494-97 >2.0xl09 6.0~109 494- 127 >2.0x109 4.7~109 Q94-128 >2.0xl09 9.4x109

Company B B94-02 3 . 8 ~ 1 0 ~ 1.1~109 B94-06 5.5~10' 5.2x108 B94-12 2.1x108 7.8x108 B94-17 5.9x108 3.4~109 B94-36 2.7x108 5.2~108

I B94-46 1.9x108 1.6~109 B94-51 8.9x108 2.4x109

I B94-54 5.4x108 1.2~109 B94-83 6.0~105 4.2x109 B94-102 4.1 ~ 1 0 7 2.5 xlo9 B94- 109 7.1 xlo7 8 . 0 ~ 1 0 ~ B94- 11 8 5.4~107 6.0x108 494-05 1.7~109 2.1 ~ 1 0 9 494-06 >2.0~109 3.3~109 494-07 4.8x108 1.7~109 494-38 5 . 9 ~ 1 0 ~ 2.7x109 494-41 >2.0x109 1.2x1010 Q94-42 6 . 7 ~ 1 0 ~ 2.9x109 Q94-48 >2.0xl09 3.Ox1O9 494-50 >2.0x109 5.9~109 Q94-5 1 6.5~108 8 . 5 ~ 1 0 ~ 494-70 >2.0xl09 4.5~109 Q94-7 1 2.2~108 2.5x109 Q94-72 8 . 3 ~ 1 0 ~ 3.3x109 494-85 52x106 1.0~10"-' 494-101 >2.0x109 1.1 xlo9 Q94- 102 8.1x108 3.1 xlO9 494-109 >2.0xl09 7.2~109 Q94-117 >2.0xl09 3.1 x109 Q94-129 >2.0xl09 1.3~1010 494-131 >2.0~109 5.5~1010

Table 1. (concluded)

Rhizobial Contaminant cellslg cfulg

ID a inoculant inoculant

Company C

B94-09 2.2~109 1 . 3 ~ 1 0 ~ B94-28 2.2~109 1.8x109 B94-70 1.3x108 3.1 x109 B94-78 2.0x10s 2.8~109 B94-88 1.4~109 5.1 xlo9 B94-89 1 . 1 ~ 1 0 ~ 3.4x109 B94-100 3.6x108 6.6x109 B94-113 1.6~107 1.8x108 B94-129 5 . 6 ~ 1 0 ~ 6.4x109

aB designation indicates sample tested in Beaverlodge, Alta, and Q designation indicates sample tested in Sainte-Foy, QC.

Table 2. Nonrhizobial bacterial contaminants identified in 100 samples of nonsterile peat base legume inoculants collected in 1993 and 1994.

Found at > lo8 cfu/g Found at lo6-lo8 cfu/g

Acinetobacter spp. Alcaligines xylosoxydans ssp. Bacillus spp. denitrzpcans Cornomonas acidovorans Citrobacter freundii Enterobacter cloacae Klebsiella oxytoca Enterobacter agglomerans Klebsiella planticola Enterobacter spp. Ochrobactrum anthropi Klebsiella pneumoniae Staphylococcus spp. Klebsiella terrigena Pseudomonas aeruginosa Sphingobacterium multivorum Sphingobacterium spp. Xanthomonas maltophilia

inoculants must not have any detrimental effect on the active species" and "the number of viable cells of microbial species, other than the desired nodule inducing Rhizobium species, are at a level not likely to affect the viability or performance of the desired species" (Anonymous 1986). These regulations have not been enforced owing to lack of information regarding the extent and nature of the nonrhizobial micro-organisms present or of the interactions involved.

Anumber of the organisms isolated from commercial legume inoculants sold in Canada during the 1993 and 1994 seasons are significant opportunistic human pathogens. These include Pseudomonas aeruginosa, Enterobacter cloacae, Klebsiella pneumoniae, Klebsiella oxytoca, Xanthornonas maltophilia, and Citrobacterfieundii. Although generally considered nor- mal soil inhabitants, bacteria such as Pseudomonas aeruginosa and Xanthomonas maltophilia have the "potential for adverse health effects due to exposure to high concentrations of micro- bial products during production or environmental application" (George et al. 1993). Klebsiella pneumoniae is an important

Can

. J. M

icro

biol

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

Sim

on F

rase

r U

nive

rsity

on

11/1

4/14

For

pers

onal

use

onl

y.

Notes

cause of nosocomial and community-acquired infections (Farmer I11 and Kelly 1991). While many species of Entero- bacteriaceae cause extraintestinal human infections, K. pneu- monia, K. oxytoca, and E. cloacae are, in fact, part of a small group that accounts for most of these infections (Farmer and Kelly 1991). Citrobacter freundii has been described as a "well-recognized human pathogen that causes a variety of infections" (Kelly et al. 1985).

Sphingobacterium multivorum has recently been reported to accept, by conjugation, the symbiotic plasmid from R. legumi- nosarum bv. trifolii and to express it in terms of nodulating ability of clover but without functional nitrogen fixation (Fenton and Jarvis 1994). This raises the possibility that a fast growing transconjugant Sphingobacterium could become dominant in an inoculant and actively create ineffective nodules. It also raises the interesting possibility that some "ineffective rhizobia" may in fact be ineffective nodule- producing nonrhizobia.

Approximately 25% of the inoculants tested were found to contain, at high levels, at least one identifiable opportunistic human pathogen. Some inoculants contained two or more different species of pathogens. The potential for human expo- sure to opportunistic pathogens in rhizobial inoculants may be exacerbated by the fact the organisms are commonly present at high concentration in products which are powdery, are subject to wind blow and air transport, and will be opened and applied in bulk during seed inoculation. Based on information from this and previous studies, new Canadian registration requirements

4 will require that inoculant products produced using nonsterile carrier materials carry package labels warning users to avoid

1 product inhalation and contact with skin or eyes. The deliberate and thoughtful introduction of beneficial

microorganisms into the environment for the achievement of specific purposes, such as the inoculation of legume seed with Rhizobium bacteria, is of unquestioned value. The routine provision to the public of such bacteria, however, in a milieu

I that includes variable and unknown numbers and types of microorganisms in a potentially unsafe "microbial soup" deserves re-examination. It is true that most rhizobial inocu- lants have been prepared in nonsterile carrier in essentially the

I same fashion for at least 75 years without apparent human health problems (Beringer and Bale 1988). However, what were once acceptable practices are often found to be unaccept- able today, either for human health or other environmental reasons. One of the three manufacturers of the nonsterile carrier

I inoculants included in this study has since switched to the production of sterile-carrier inoculants.

1 The manufacture of legume inoculants using sterile bags and carrier is technically simple and normally results in products with consistently high Rhizobium cell numbers. Irradiating the peat and bags may raise production cost but allows for signifi- cantly increased quality control testing by plate count, thus I reducing MPN plant-nodule grow-out testing (which normally requires 4 weeks to obtain a result). This should result in 1 products of higher quality and effectiveness. Use of sterile carriers would prevent the potential for in-the-bag inhibition of

rhizobial growth by antagonist organisms and should also eliminate the threat of exposure of either the host crop or human operators to accidental pathogens.

References

Anonymous. 1986. Office consolidation fertilizers act R.S., c. F-9. Fertilizers regulations. Feed and Fertilizer Division, Government of Canada, Ottawa.

Anonymous. 1991. The methods of testing legume inoculant and pre- inoculated seed products. Fertilizers act, section 23, regulations. Feed and Fertilizer Division, Government of Canada, Ottawa.

Antoun, H., Bordeleau, L.M., Gagnon, C., and Lachance R.A. 1978. Actinomycetes antagonistes de champignons et n'affectant pas le Rhizobium meliloti. Can. J. Microbiol. 24: 558-562.

Beringer, J.E., and Bale, M.J. 1988. The survival and persistence of genetically-engineered micro-organisms. In The release of geneti- cally-engineered micro-organisms. Edited by M. Sussman, C.H. Collins, F.A. Skinner, and D.E. Stewart-Tull. Academic Press, London. pp. 29-46.

Catroux, G. 1991. Inoculant quality standards and controls in France. In Report of the expert consultation on legume inoculant production and quality control. Edited by J.A. Thompson. Food and Agricul- tural Organization of the United Nations, Rome. pp. 113- 120.

Catroux, G., and Amarger, N. 1992. Rhizobia as soil inoculants in agriculture. In Release of genetically engineered and other micro- organisms. Edited by J.C. Fry and M. J. Day. Cambridge University Press, Cambridge. pp. 1-13.

Farmer, J. J., LII, and Kelly, M.T. 1991. Enterobacteriaceae. In Man- ual of clinical microbiology. 5th ed. Edited by A. Ballows, W.J. Hausler, Jr., K.L. Herrmann, H.D. Isenberg, and H. J. Shadomy. American Society for Microbiology, Washington, D.C. pp. 360-383.

Fenton, M., and Jarvis, B.D.W. 1994. Expression of the symbiotic plasmid from Rhizobium leguminosarum biovar trifolii in Sphingo- bacterium multivorum. Can. J. Microbiol. 40: 873-879.

George, S.E., Kohan, M.J., Gilmour, M.I., Taylor, M.S., Brooks, H.G., Creason, J.P., and Claxton, L.D. 1993. Pulmonary clearance and inflammatory response in C3H/HeJ mice after intranasal exposure to Pseudomonas spp. Appl. Environ. Microbiol. 59: 3585-3591.

Kelly, M.T., Brenner, D.J., and Farmer, J.J., 111. 1985. Entero- bacteriaceae. In Manual of clinical microbiology. 4th ed. Edited by E.H. Lennette, A. Ballows, W.J. Hausler, Jr., and H.J. Shadomy. American Society for Microbiology, Washington, D.C. pp. 263-277.

Olsen, P.E., Rice, W.A., Bordeleau, L.M., and Biederbeck, V.O. 1994. Analysis and regulation of legume inoculants in Canada: the need for an increase in standards. Plant Soil, 161: 127-134.

Olsen, P.E., Rice, W.A., and Collins M.M. 1995. Biological contami- nants in North American legume inoculants. Soil Biol. Biochem. 27: 699-701.

Panthier, J. J., Diem, H.G., and Dommergues, Y. 1979. Rapid method to enumerate and isolate soil actinomycetes antagonistic towards rhizobia. Soil Biol. Biochem. 11: 443-445.

Pugashetti, B.K., Angle, J.S., and Wagner, G.H. 1982. Soil micro- organisms antagonistic to Rhizobium japonicum. Soil Biol. Bio- chem. 14: 45-49.

Vincent, J.M. 1970. A manual for the practical study of the root- nodule bacteria. International Biological Programme Handbook No. 15. Blackwell, Oxford.

Can

. J. M

icro

biol

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

Sim

on F

rase

r U

nive

rsity

on

11/1

4/14

For

pers

onal

use

onl

y.