J Sci Food Agric 1990, 50, 127-140

Agriculture Group Symposium Microbial Inoculants in Agriculture

The following are summaries of papers presented at a joint meeting of the Agriculture Group and the Biotechnology Group of the Society of Chemical Industry held on 14 February 1989 at the Society of Chemical Industry, 14-15 Belgrave Square, London SWlX 9 P S . The papers published here are entirely the responsibility of the authors and do not reflect the views of the Editorial Board of the Journal of the Science of Food Agriculture.

Microbial Inoculants: A Promise Deferred?

P Bernard Tinker Terrestrial and Freshwater Sciences Directorate, Natural Environment Research Council, Swindon, Wiltshire SN2 IEU, UK

Over many years there have been attempts to use microbiological inocula to improve crop growth and agricultural productivity. It is a contention of this paper that, when surveyed overall, the results are deeply unsatisfactory, in that each group of new materials is greeted with enthusiasm but usually fades into disappointment and neglect.

Among the earliest such interests were nitrobacterin and phosphobacterin, which were very popular during the 1930s. Despite many claims from within the USSR, and careful tests by USDA and Australian scientists, it was finally concluded that no dependable and repeatable positive results could be obtained.

The main type of inoculum which has established itself as a major product is Rhizobium. It is well established that certain crops require inoculation, because of the low population of the specific Rhizobium species which are required for particular crops. Despite this well established commercial success, it is notable that the huge amount of research work on Rhizobium over the last 15 years has produced little of applied value. Strains of Rhizobium in commercial production and distribution are still substantially similar to those used before that time, and have been obtained by simple selection and testing from natural isolates.

Mycorrhizal inocula have generally proved disappointing on a field scale. Numerous experiments have shown extremely large responses to inoculation, but usually when carried out in sterilised soil most often in the glasshouse. The potential for the future may be better, the most promising application probably being in tree nursery inoculation. A large range of other organisms have been considered, and are

127

J Sci Food Agric (50) (1990)-62 1989 Society of Chemical Industry. Printed in Great Britain

I28 Agriculturr Group Symposium

now popular subjects of research. Pseudomonads have been suggested for root disease control and Azotobacter for nitrogen production in the rhizosphere.

The reasons for this rather disappointing history are discussed, together with prospects for the future.

Fate of Microbial Inoculants in Soil

Martin Wood Department of Soil Science. University of Reading, London Road, Reading GRI 5AQ, U K

In order to survive, multiply and be effective in soil, an introduced microorganism should be tolerant of chemical factors such as acidity and physical factors such as extremes of temperature. It should also be able to move through soil to its site of action. Bacteria are the main organisms of current interest, and the required destination for most inoculants is the plant root. At present it is impossible to predict the fate of an organism introduced into soil. This paper therefore considers some of the factors which may influence the survival, multiplication and movement of bacteria in soil, in particular the soil around plant roots.

Bulk soil measurements of acidity, salinity and water content are of little use in understanding the fate of microorganisms in soil, because they survive in microsites which may be quite different from the bulk soil conditions. Microsites are also extremely difficult to measure. The soil around plant roots may differ by as much as 2.5 pH units from the non-rhizosphere soil, will have a much lower concentration of phosphate, and may be wetter or drier than the bulk soil. A successful inoculant must be able to tolerate these localised conditions. There may also be a higher population of predatory protozoa and nematodes in the rhizosphere, which will reduce the population of introduced microorganisms.

For appreciable movement of bacteria through soils there must be enough water- filled pores of the required diameter to provide a continuous pathway. Movement will be restricted at water potentials of approximately - 100 kPa, which is much higher than potentials which cause plants to wilt. Motile organisms (those possessing flagella) may have a better chance of meeting a plant root, and chemotactic responses may assist in avoiding unfavourable conditions and locating a plant root. Bacteria may also move relatively rapidly by mass flow through cracks and channels formed by earthworms and dead roots. This may play an important part in determining the fate of bacteria introduced into soil.

The fickle nature of most current soil microbial inoculants such as Rhizobium sp and Pseudornonas sp is probably a reflection of the complexity and hostility of the environment into which they are introduced. Improvements in the reliability of soil inoculants will depend upon a greater understanding of soil microbial ecology. In the meantime the development of soil inoculants will remain empirical.

Microhiul irioculants in agriculture 129

Biological Control of Diseases Using Microbial Inoculants

Richard Campbell Department of Botany, University of Bristol, Bristol BS8 lUG, UK

Microbial inoculants for disease control have existed, and been available commercially on a small scale, for many years. They have not, however, been widely used in general agriculture. The main reason for this is that a need for them has not been perceived by the industry, and such research as has been done has had difficulty in developing inocula which have consistent control of disease. There has recently been an increased interest in biocontrol agents, and research and development effort has increased.

Most of the research programmes have concentrated on soilborne diseases, especially of intensively grown crops, or on situations where conditions especially favour introduced inocula. Areas of interest include: (1) disease where there is RO chemical control or varietal resistance available at present; (2) horticultural crops where sterilised or specially prepared composts may be used and where plants are transplanted; (3) diseases where entry is by defined infection routes (eg pruning wounds on tree).

There are many possible modes of action of biocontrol agents, but those which produce antibiotics, siderophores or lytic enzymes are receiving most attention at present.

Future developments are likely to concentrate on integrated control, using chemicals and agriculture practices to enhance biological control or allow establishment of inocula. Genetically engineered organisms with enhanced biocontrol ability and improved colonisation and survival are possible. The problem is to determine the exact characteristics, which could be manipulated genetically, to confer these properties on the organism.

The concept of biological control is attractive, but there are clearly problems which were not foreseen in early work. These problems are not insuperable, but much more information is needed on basic microbial ecology before the behaviour of biocontrol agents can be monitored and predicted in natural and agro- ecosystems.

The Development and Use of Viral Insecticides

David H L Bishop NERC Institute of Virology, Mansfield Road, Oxford OX1 3SR, UK

A variety of biological agents for the control of insect pests have been developed in the twentieth century. Both bacterial toxins, such as Bacillus thuringiensis toxin, and insect-specific viruses have found favour as alternatives to chemical insecticides,

130 Agriculture Group Symposium

particularly for environmentally sensitive cultivars (such as forests that are in water catchment areas).

Unlike many chemical insecticides these biological agents are specific and limited in their effect to particular (or a few) insect species. There are, however, disadvantages to the extensive use of viral insecticides. For example, they are of limited use against pest complexes that include susceptible and non-susceptible species. Viruses are much slower than chemicals, so that crop damage may occur while the viruses produce an effect. On the whole, biological control agents are more expensive to manufacture than chemicals. Despite these drawbacks there is a role for biological insecticides both as stand-alone agents and in concert with others in integrated programmes of pest management.

Three areas illustrate the use of viral insecticides: the control of the leipidopteran pest Panolis Jammea, pine beauty moth, in Scotland; the control of the coleopteran Oryctes rhinoceros beetles in palm groves in the Seychelles; and the control of the hymenopteran Neodiprion sertijer, pest pinesawfly, in Scottish forests. Effective control procedures have been developed for these pests by staff of the Institute of Virology.

The opportunities for the development of novel insecticides using genetically engineered viruses are described, including insecticides that can control pests that are of concern not only to agriculture but also to human health.

Past, Present and Future Uses of Rhizobiurn Inoculants

Penny R Hirsch AFRC IACR Rothamsted Experimental Station, Harpenden, Herts AL5 UQ. UK

The beneficial effects of including legumes in crop rotations was known to the Romans and was exploited in European agriculture long before it was realised that they enriched soil through fixation of atmospheric di-nitrogen by symbiotic Rhizobium bacteria in their root nodules. The first commercial inoculants containing pure cultures of Rhizobium specific for different legumes were produced at the end of the last century. Since then research on the Rhizobium legume symbiosis has led both to widespread use of crop inoculation in appropriate circumstances, and also to proposals of how to produce improved Rhizobium through manipulation of genes involved in symbiotic N, fixation.

The use of genetically engineered bacteria as crop inoculants has already caused controversy in the USA. Fears voiced both by the public and by the scientific community have led to the formulation of guidelines and legislation by advisory and regulatory bodies in many countries to oversee the deliberate release of genetically manipulated organisms in the field. Of particular concern is the possibility that novel genes cloned into an inoculant could transfer to native strains and spread through the indigenous population, even if the inoculant itself does not persist. Such hybrids could have potentially undesirable properties (depending upon the nature of the novel gene) and would be adapted to survive in the environment.

Microbial inoculants in agriculture 131

In an experiment to investigate the possible occurrence of horizontal gene transfer from introduced to native bacteria in the field, a self-transmissible plasmid in Rhizobium leguminosarum biovar viceae was marked with Tn5, a transposon determining resistance to the antibiotic neomycin, and used to inoculate peas on the Rothamsted farm. The marked Rhizobium was classified as ‘genetically manipulated’ since Tn5 was derived from another genus of bacteria, although no in- vitro genetic engineering was involved in its construction. Therefore clearance was obtained from the HSE Advisory Committee on Genetic Manipulation before the experiment could proceed.

The transposon marker proved invaluable for monitoring survival of the inoculant strain itself, but no evidence for transfer of the marker to other rhizobia isolated from root nodules has been obtained in the two years following release of the inoculant. Results in the laboratory using sterile soil to which the inoculant and different recipient strains were added indicated that transfer would not be detected when fewer than lo6 per gram of soil of either parent were present. It is difficult to extrapolate from laboratory results to predict what might occur in the field. However, the fact that no horizontal transfer of the Tn5 marker gene was observed implies that there are no special circumstances in the field which greatly enhance genetic interaction between rhizobia.

These findings may help in the future in assessing the risks of using Rhizobium as agricultural inoculants when strains genetically manipulated to improve symbiotic performance are available.

Prospects for Practical Use of Inoculum of Vesicular-Arbuscular Mycorrhizal Fungi

David P Stribley AFRC Institute of Arable Crops Research, Rothamsted Experimental Station, Harpenden, Herts AL5 UQ, UK

Experiments made under controlled conditions have demonstrated beyond doubt that, per unit length of absorbing organ, vesicular-arbuscular (VA) mycorrhizas take up phosphorus (P) from the soil solution much faster than do non-symbiotic roots, over a wide range of P concentrations in the soil. Consequently, mycorrhizal plants may attain maximum yields at reduced inputs of phosphate fertiliser. Uptake of some trace metals may also be enhanced by colonisation of root systems by VA mycorrhizal fungi, but nearly all other effects of mycorrhizas on the physiology of the host, for example on water relations and disease resistance, are an indirect consequence of improved P nutrition.

Field experiments with various crops on various soils have demonstrated that inoculation with VA mycorrhizal fungi can increase growth of crops on soil low in phosphate. No unique effects of VA mycorrhizas on crop physiology have been convincingly demonstrated, nor are they to be expected. Small-grain cereals have usually responded poorly to inoculants. Artificial inoculation is effective

I32 Agriculture Group Symposium

presumably because it increases the rate of colonisation of roots by the symbiotic fungi. There may be differences between strains of mycorrhizal fungi in efficacy of P uptake, but this is not fully understood. Artificial inoculation may not always be necessary to achieve maximum benefit since suitable husbandry practices may maximise the natural inoculum.

Most field experiments with inoculants of VA mycorrhizal fungi have been highly empirical and have given inexplicably variable results. The economic implications and precise aims of these experiments have rarely been stated explicitly. The only realistic benefit of the use of mycorrhizal inoculum in large-scale agriculture is to reduce inputs of phosphate fertilisers. In advanced agriculture this is probably of little immediate consequence, but it may be of value to undeveloped systems.

A Compost System for Microbial Release and Recovery Experiments

Robert C Brooks," Terence R Fermorb and Alan J McCarthy" "Department of Genetics and Microbiology, Life Sciences Building, University of Liverpool, PO Box 147, Liverpool L69 3BX. UK; bDepartment of Microbiology, AFRC Institute of Horticultural Research, Littlehampton, West Sussex, BN17 6LP, UK

There is now a requirement for versatile model ecosystems into which different microbial host/vector systems can be released and monitored under different environmental conditions. The authors are using laboratory-acquired data to construct a scaled-up compost system which will have predictive value. Composting is essentially an aerobic process in which plant biomass is degraded by a heterogeneous population of microorganisms. Composts for mushroom cultivation are prepared by a defined procedure comprising a number of phases in which actinomycetes and bacilli are very active. Since these microorganisms are of industrial and potentially agricultural importance, Phase I1 compost has been selected as a substrate to provide a rigorous test of gene survival and transfer mediated by released recombinants.

The system is based on a thermostatically regulated heating box which can hold up to 50 kg compost at temperatures up to 70°C. Compost is held in a wire basket (80 x 8Ox 80 cm) above a water reservoir through which air is pumped when necessary. In this way, the compost can be maintained in an aerobic, moist and consequently active state. A second level of containment is provided by the room itself, which has been fitted with an air lock, and the whole containment facility can be steam-sterilised and chemically disinfected at the end of each experiment. The compost and containment room can be separately monitored for temperature and humidity. Air samples from the compost provide for carbon dioxide and ammonia determinations, and potentially gas chromatography. The microbiological effectiveness of the containment can be monitored using an Andersen sampler. Compost sampling regimes for analysis of the microflora have been devised.

A steady core temperature was easily maintained, and manipulation of composting temperature using this system was achieved. However, maintaining

Microbial inoculants in agriculture 133

microbial activity by controlling aeration and humidity presented a number of problems. In initial experiments the compost rapidly became desiccated and CO, levels in the core rose dramatically. A polythene jacket and periodic aeration with humidified air led to a c 70% reduction in core C 0 2 levels but the surface layers of the compost continued to dry out rapidly. A water loss of approximately 1 litre day- ' was estimated, and a periodic misting system is being developed on this basis. In laboratory experiments with inoculated sterile compost, significant pH fluctuations have been observed and interpreted as being due to ammonia production via deamination. pH in the scaled-up system has yet to be monitored, but it has not been possible to detect ammonia in the active core of the Phase I1 compost. The authors have not detected any rise in the numbers of airborne actinomycetes and bacilli outside the containment room, where normal background counts of 100-200 propagules m-3 air were detected. This compares favourably with numbers detected inside the containment room during experiments (13-4.0 x lo3 m-3 air).

The system described has now been developed to the point where long-term experiments can be attempted. The compost microflora will be characterised using techniques specifically designed for this purpose. These experiments will be followed by release and recovery studies using a stable pigmented variant of the compost actinomycete species, Saccharomonospora uiridis. The relationship between the results of these studies and the authors' laboratory data on release and recovery of this organism will be an important priority. In this stepwise manner, the authors will proceed to an investigation of the survival and transfer of actinomycete genes introduced into compost, initially using recombinants derived from the S uiridis variant and later using Bacillus spp.

Immunofluorescence Studies with a Flavobacterium Isolated from Soil

Julie Mason and Richard G Burns Biological Laboratory, University of Kent, Canterbury, Kent CT2 7NJ, U K

The successful exploitation of microbial soil amendments in agriculture and for environmental decontamination will depend upon sensitive methods of measuring inoculant survival, proliferation and dispersal. The authors are investigating various immunological techniques for the identification and enumeration of microorganisms in order to monitor their fate on release into the environment. A powerful method of detecting microorganisms involves preparing monoclonal antibodies which recognise cell-surface antigens and are species specific. This approach was adopted with a Flavobacterium species isolated from soil and currently being used in survival, growth and movement studies in laboratory and field experiments.

Monoclonal antibodies (MAB) were raised using a crude cell wall/membrane preparation of Flavobacterium P25. Hybridomas producing antibodies to the

134 Agriculriirr Group Symposium

antigen(s) were screened using an enzyme-linked immunosorbant assay (ELISA). MABs B4, B7 and C6 were chosen to use in immunofluorescence studies. An indirect immunofluorescence technique using a secondary antibody labelled with a fluorescent marker (FITC) was used to detect the Flauobacterium. The antibodies were seen to target a surface component of the bacterium. Cross-reactivity of the antibodies with other Gram-negative bacteria (including Flauobacterium species), Gram-positive bacteria and soil fungi was investigated but none has been found.

Immunochemical procedures are being used to identify the surface component(s) (ie LPS or protein) targeted by the antibodies. It is the authors' intention to use the monoclonal antibodies to detect, locate and quantify Flauobacterium added to soils and in the rhizosphere.

A Broad-host-range Plasmid Marker System for Monitoring the Fate of Genetically Engineered Microorganisms in Soil Model Environments

Craig Winstanley," J Alun W Morgan," Fiona C Raitt," Jonathan P Carter," Roger W Pickup,b J Gwynfryn Jonesh and Jon R Saunders" "Genetics and Microbiology Department, University of Liverpool, PO Box 147. Liverpool L69 3BX, UK; *Freshwater Biological Association, Windermere Laboratories, Ambleside, Cumbria LA22 OLP. UK

Little is known about the fate of genetically engineered microorganisms (GEM) or their recombinant DNA in the event of accidental or deliberate release into soil or other natural environments. It is therefore important to study model release systems involving GEM in order to assess the possible hazards and environmental consequences involved.

The authors have developed a versatile plasmid marker system to assess the survival of GEM and the transfer of recombinant DNA. The basic test system involves detection of the xylE gene and its product (catechol2,3 oxygenase; C230) of the TOL plasmid pWW0. On the marker plasmids xylE is expressed from the strong lambda promoter PR. A simple colour test enables the detection of colonies that express C230 activity. After spraying with 1 :< w,/v catechol solution, such colonies produce a yellow coloration due to the conversion of catechol to 2- hydroxymuconic semialdehyde. The release hosts and marker plasmids can be detected by a variety of biological and physical methods. These include conventional culture with selection for plasmid-encoded phenotypic markers, DNA hybridisation, direct enzyme assay and immunological assays. Either whole xylE gene probes or oligonucleotide probes specific for xylE, xylE mRNA or host RRNA can be used to study the fate of both the host and the plasmid. Antibodies to purified C230 can be used in ELISA to detect as few as lo3 cells in 100 ml filtered water samples. An immunomagnetic capture strategy has been developed involving the use of a monoclonal antibody specific for the flagella of the model release host Pseudomonas putida PaW340. This antibody can be attached to magnetic polystyrene beads (Dynabeads) and used to extract target cells from environmental

Microbial irroculurr~s i t , agriculture 135

samples. O n plasmid pLVlOl0 the expression of xy!E is unregulated, whereas on pLV1013 the expression of the gene is regulated by the presence of the lambda temperature-sensitive repressor gene cI857. C230 product is induced by elevation of temperature, thus enabling the analysis of any effects caused by adverse metabolic burden due to over-expression of xylE. Since pLVlOl0 and pLV1013 both contain the broad-host-range replication functions of the IncQ plasmid pKT230, it has been possible to introduce the two marker plasmids into a number of different hosts to assess the performance of the system. High-level expression of xylE from pLVlOl0 and inducible C230 activity from pLV1013 have been observed in a wide range of Gram-negative bacteria.

It has been found that the stability of the over-expressing construct pLVlOl0 varies not only between but also within bacterial species, and that there are species- specific differences in the regulation of the marker systems. These observations suggest that DNA contained in released GEMS, if it has the potential to transfer, may not act in a predictable manner in the natural environment.

A Novel Detection System for Genetically Manipulated Microbial Inoculants in Soil

Elizabeth M Meiklejohn, Audrey Meikle, L Anne Glover, Kenneth Killham and James I Prosser Departments of Genetics and Microbiology, Biochemistry and Soil Science, University of Aberdeen. Aberdeen. UK

Full realisation of the potential for use of genetically engineered microbial inoculants in agriculture requires the development of accurate and sensitive systems for the detection of cells and of introduced genetic material. Here are reported preliminary results on the development of luminescence-based techniques which provide sensitive in-situ estimation of microbial cell concentrations in the soil. The techniques have been tested using Escherichia coli containing a plasmid-borne lux cassette which codes for the enzyme luciferase, with light output measured by luminometry .

Luminescent strains of E coli DH1 and MM294 were obtained by insertion of two plasmids pUCD607 and pBTK5 containing the lux cassette, originally obtained from Vibrio fischeri and also coding for ampicillin resistance. Luminescence during growth of each strain in liquid medium was studied by inoculation into L-broth, followed by incubation at 30°C on a rotary shaker. Light output was measured in 2- ml samples on a LKB Model 1251 luminometer with the output integrated over a 30-s period. Luminescence in soil was investigated by inoculation of each strain into a sterilised sandy loam soil contained in a soil column. The column was supplied continuously with minimal medium containing sodium acetate as carbon source. Samples were removed at regular intervals to determine light output by luminometry, and viable cell number using the dilution plate count.

Variation in light output during growth in liquid culture differed between the two plasmids. E coli strain DH 1 containing plasmid pUCD607 showed maximal light output during the exponential growth phase, but light output was greatest in E coli MM294 containing plasmid pBTK5 during the stationary phase. In liquid medium, luminometry permitted quantitative detection of cell concentrations approaching 1 cell ml-'. In soil, sensitivity was reduced by quenching of light by soil particles. Maximum sensitivity was obtained immediately after inoculation of soil, and quantitative detection of approximately 10 cells g- was possible. The relationship between cell concentration and light output during subsequent growth in soil varied with the strain used and growth phase as in liquid culture.

These preliminary data demonstrate the sensitivity and speed of the luminescent- based techniques and their ability to determine microbial cell concentration in situ. Results also indicate the possibility of distinguishing growing and non-growing populations in the soil, by use of plasmids providing different profiles of light output. Detection of luminescent strains of bacteria in the soil may also be achieved by image-enhanced microscopy, gene probing and luminescence of colonies following plating on to solid media. Luminescence-based detection technology therefore provides a range of techniques capable of sensitive, rapid and in-situ detection of genetically engineered strains and genetic material.

Immunological Detection of Genetically Manipulated Streptomycetes in Soil

Anil Wipat," Elizabeth M Wellingtonb and Venetia A Saunders' "School of Natural Sciences, Liverpool Polytechnic, Liverpool L3 3AF, U K ; bDepartment of Biological Sciences, University of Warwick, Coventry CV4 7AL. U K

Streptomycetes are commonly found in soil where they are thought to exist mainly as spores. They play a part in the recycling of plant material and maintenance of soil fertility. Furthermore, members of the group produce agriculturally important herbicides and insecticides. Efficient gene cloning systems have been developed for some species and may be used for large-scale production of foreign proteins. Genetically manipulated streptomycetes are therefore potential candidates for intentional or accidental release into soil environments. A prerequisite for release is the availability of sensitive and selective techniques for monitoring the behaviour of genetically manipulated microorganisms and their genetic material in the environment. The authors are currently investigating immunological methods for detection and recovery of genetically manipulated streptomycetes released into a soil environment. Antibodies have been raised initially to spore-surface antigenic determinants of the intended release host Streptomyces liuidans 1326, a member of the Streptomyces species-group cluster 21. Members of this cluster are frequently isolated from soil.

Polyclonal antisera have been raised in mice and rabbits by immunisation with intact spores, and monoclonal antibodies have been produced by in-vivo immunisation of mice followed by fusion of spleen cells with an immortal cell line.

Microbial inoculants in agriculture 137

Spore antiserum from mice reacted both with spores and with mycelium from the immunising strain, and cross-reacted to some extent with other members of cluster 21 when tested by ELISA. Western analysis of spore proteins from several strains of this cluster reveals the presence of several common antigenic determinants. A number of monoclonal-antibody-producing clones that secrete antibodies against various epitopes of the S lividans spore surface have been obtained. One such clone which has been further characterised produces IgGclass antibodies against a spore-surface polypeptide. The antibody reacts against a number of cluster-21 members when tested by ELISA. However, preliminary analysis suggests that the epitope for this antibody does not occur on the surface of other non-cluster-21 representatives. Such an antibody may thus prove useful as a means of cluster-21 identification and recovery.

Studies are in progress to optimise the use of both polyclonal antisera and monoclonal antibodies for recovery and identification of the release host. Methods under investigation include magnetic capture by using antibodies coupled to magnetic polystyrene beads (Dynabeads, DYNAL UK Ltd), ELISA applied to both soil isolates and directly to soil samples, and immunofluorescence techniques for direct observation of Streptomyces spores in soil.

Post-harvest Biocontrol of Anthracnose Disease of Mangoes

Irene Koomen," John C Dodd," Mike J Jeger' and Peter Jeffries" "Biological Laboratory, University of Kent, Canterbury, Kent CT2 7NJ, UK; bOverseas Development & Natural Resources Institute, 56-62 Grays Inn Road, London WClX 8LU, UK

Mango (Mangifera indica L) is grown in over 87 countries with India accounting for about 65% of total world production. Mangoes have been grown on a 'home- garden' basis in most countries, but countries such as the Philippines have begun to plant commercial orchards and have doubled exports of mango over the last 20 years, making them an important revenue crop. Anthracnose is the commonest and most serious disease that mangoes can suffer. Yield losses through this disease can severely affect production, particularly when the disease is 'quiescent', ie although the fungus is present at an early stage in fruit development it does not cause damage until after the fruit has ripened. This postharvest disease is particularly damaging as the producers have already paid for harvesting, packaging and transport to the market. The disease is caused by the fungus Colletotrichum gloeosporioides (teleomorph: Glomerella cingulata). A collaborative research programme has been established between the University of Kent and ODNRI to devise better control strategies for this disease. The programme involves epidemiological studies and cytological investigations, as well as chemical and biological approaches to disease suppression. This presentation concerns the latter aspect of the programme.

A variety of bacteria, yeasts and filamentous fungi have been isolated from the surface of Kenyan pickling mangoes and Peruvian mangoes cv Hayden, and from mango leaves from Sri Lanka, Thailand, India, Uganda and Tanzania. All isolates

138 Agriculture Group Symposium

(c 600 in total) have been tested for their inhibitory potential against C gloeosporioides on malt extract agar. Inhibition of the pathogen varied between no inhibition (even slight stimulation) up to 70% inhibition.

A second-stage screen was then carried out in which all antagonists that showed over 20% inhibition in the first test were tested for their effect on germination of C gloeosporioides spores on cellophane overlying tap water agar. Out of a total of 11 1 bacteria and yeasts tested, 35 bacteria and 18 yeasts prevented germination of conidia of C gloeosporioides.

The third stage of screening involved a simulated commercial post-harvest treatment in which mature mangoes were dipped in a suspension of pathogen spores incubated for 15 h and then dipped in a suspension of cells (bacteria, yeasts) or spores (filamentous fungi) of the antagonist. The mangoes were then stored at 25°C for up to 14 days to allow full anthracnose development. Four bacteria and two yeasts have maintained significant disease control in all screens so far, and tested under commercial conditions in the Philippines in 1989.

Production and Characterisation of Monoclonal Antibodies Raised against Soil Fungi

Margaret Marshall, K Gull and P Jeffries Biological Laboratory, University of Kent, Canterbury, Kent CT2 7NJ, UK

Monoclonal antibodies (Mabs) have been raised against cell wall antigens from the soil fungi Neurospora crassa and Paxillus involutus. The object is to obtain specific antibodies for use in the detection, discrimination and eventual quantification of fungal inoculants in natural ecosystems. Mab assays have been the authors’ method of choice over conventional polyclonal assays because of the potential for a continuous supply of antibody of known high specificity. Previous authors have reported that polyclonal antisera raised against fungi and tested by immunofluorescence or ELISA are notoriously non-specific.

The methodology involved three or more injections of Balb/c mice with clean cell wall fragments (500 pg per injection), or with formaldehyde-fixed conidia (1 x lo7 conidia per injection). Spleenocytes were fused by standard protocols with Sp2-0 myeolma cells 4 days after a final booster injection. Resulting hybridomas were screened for antibody production by ELISA using cell wall fragments as targets. Positive hybridomas (OD > 1.0) were further screened by immunofluorescence using cell wall fragments and spores, and only hybridomas that produced differing patterns of fluorescence were further subcloned.

Three Mabs raised against N crassa and one raised against P involutus have been further characterised. Antibody S3B3, an IgM, reacts strongly by immunofluorescence with a component in the inter-septa1 region of hyphae and conidial chains. The antibody cross-reacts with an epitope in a similar locality in Penicillium chrysogenum but not against a range of other fungi. Antibody S4D1,

Microbial inoculants in agriculture 139

another IgM, reacts with an antigen present on the surface of conidia, germ tubes and hyphae of N crassa. The epitope is found predominantly in the al-3-linked glucan fraction of cell walls of N crassa. In addition al-3-linked glucans of Schizophyllum commune and Aspergillus nidulans react positively by ELISA with this antibody. Immunofluorescence patterns produced with this antigen differ among the fungi tested. For example, in P chrysogenum the hyphae and conidiophores fluoresce but not the spores and phialides. In contrast, in Amanita muscaria the spores fluoresce but not the vegetative hyphae. Antibody S1E5 reacts strongly with an antigen present on the surface of germ tubes of N crassa but not on spore surfaces. Antibody PAXI, raised against clean cell wall fragments of P involutus, reacts to a surface component of Paxillus hyphae, and also with N crassa to a similarly located but different epitope from that of S3B3.

In conclusion, the authors report the success of protocols necessary to raise monoclonal antibodies against fungal cell wall antigens. However, the first antibodies raised in this way are not useful for detection of specific isolates of fungi in the environment but will be used to investigate distribution of particular epitopes across fungal groups. More stringent specificity might be achieved by immunising with purified specific fungal components or by immunosuppression of dominant antigens by cyclophosphamide.

Streptomycete Inoculants in Soil: Growth and Genetic Interactions

Paul Herron," Neil Cresswell," Elizabeth M Wellington" and Venetia A Saundersb "Department of Biological Sciences, University of Wanvick, Coventry CV4 7AL, U K ; bSchool of Natural Sciences, Liverpool Polytechnic, Liverpool L3 3AF, UK

Soil microcosms were used to investigate plasmid and phage mediated gene transfer under sterile and non-sterile conditions, with various nutrient amendments. Methods were also developed to improve the efficiency of inoculant recovery from soil to provide a procedure for detection of one spore per gram of soil.

All experimental work was done using members of the S griseoruber group including S lividans, the genetically well characterised host for cloning vectors. For plasmic transfer experiments pIJ673 was used, a derivative of pIJ101, with the antibiotic resistance genes aph, tsr and vph inserted and used as markers for detection of transfer of this multicopy transmissible plasmid. Transconjugants were readily detected two days after inoculation in sterile soil. Transfer frequencies were lower than in vitro although, in a cross between S violaceolatus ISP5438 as donor and S lividans as recipient, soil microcosms produce transconjugants at a lower frequency than in vitro possibly due to antibiotic inhibition of S liuidans. Plasmid transfer was detected in non-sterile unamended soil 7 days after inoculation at a frequency of 1-09 x lo-'. This demonstrates the considerable potential for gene transfer between streptomycetes in soil. Restriction digests of plasmid isolated from transconjugants showed that pIJ673 was maintained in its original form.

140 Agriculture Group Symposium

Mathematical models have been tested to determine distribution on growth form of inoculants, and preliminary data indicate a good fit to observed interactions but non-random distribution of spores is likely. SEM has been used to investigate growth in soil, and resulp indicate that germination and mycelial development began after 18 h but sporulation occurred after 3 4 days in sterile amended soil. Actinophage-mediated gene transfer was examined by inoculating KC301 (4C3 1 carrying tsr gene) into soil together with spore suspensions of S liuidans 1326. Lysogens were first detected after 2 days. The presence of prophage was confirmed by probing lysogens with 32P-labelled KC301 DNA. Free phage declined rapidly and could not be detected after 39 days.

Traditional methods of bacterial recovery from soil were used in the detection of transconjugants and lysogens. Shaking with saline and direct plating of soil extracts gave poor recovery of inoculum and poorly reproducible results at low population densities in soil. An alternative approach was developed involving dispersing soil with iminodiacetic ion exchange resin in various eluents followed by several centrifugation steps to allow all of the inoculum in a large volume of soil to be concentrated into a small volume of liquid. Maximum recovery was found using an eluent of PEG/sodium deoxycholate mixture. Ten spores per 100 g soil could be detected using this method.