TECHNICAL FIELD

The present invention relates to the control of fungal infections in plants and plant products.

BACKGROUND ART

Biological control using antagonistic microorganisms to reduce postharvest fruit decay has attracted attention in recent years because of public concern about chemical applications to food and the development of resistance in pathogens to chemicals (9, 20) . A number of antagonists have been reported to be effective against various fruit pathogens (10, 17, 18) . A Bacillus subtilis strain B-3 has been tested under semi-commercial conditions against peach brown rot, showing good control ability and being generally compatible with waxes (12) .

Although Pseudomonas cepacia is known as an incitant of sour skin rot in onions (19) , two strains of _______ cepacia have been found to be active in controlling blue mold and gray mold on apple and pear (10) and green mold caused by Penicillium dicritatum Sacc. on citrus fruit

(21) , respectively. Another strain of this bacterium was found to promote P _^. diαitatum infection on postharvest oranges (7) . Protective ability by different strains of Pseudomonas under commercial conditions and the response of fruit to the bacteria have not been reported.

Description of the Invention Abbreviations: PAL = phenylalanine ammonia-lyase; 2,4-D = 2,4-dichlorophenoxyacetic acid.

The present invention relates to the use of an isolate of _______ cepacia (ID 2131, described herein) which is capable of providing significant control of Pencillium diαitatum under both laboratory and packing line conditions .

The present invention also relates to a novel bacterial isolate (ID 2859, described herein) related to

Pseudomonas σlathei capable of providing this activity. These isolates should also be of use in protection against other fungal plant pathogens such as Fusarium crookwellense. Colletotrichium linde uthianum, P. italicum. and fungal agents which are tropical plant pathogens.

Isolates ID 2859 and ID 2131 were deposited with the Australian Government Analytical Laboratories of 1 Suakin Street, Pymble 2073, New South Wales, Australia in accordance with the provisions of the Budapest Treaty on 25 July 1991 under accession numbers N91/41766 and N91/41765 respectively.

Unlike ID 2131, ID 2859 does not appear to produce antibiotics and does not appear to rely on antibiotic production for its protective effect against fungal attack.

According to a first aspect of the present invention there is provided a novel Pseudomonas strain herein identified as ID 2859. According to a second aspect of the present invention there is provided a method for protecting a plant or plant product from fungal infection which method comprises exposing the plant or plant product to ID 2131 or ID 2859. Preferably the plant or plant product is citrus fruit.

More preferably the citrus fruit is post-harvest oranges.

Preferably the fungal infection is infection by _______ digitatum.

The ID 2131 or ID 2859 is applied to the plant or plant product at a concentration of at least 10 ⁵ colony forming units/ml (cfu/ml) .

Preferably the concentration is 10 ⁸ cfu/ml to 10 ⁹ cfu/ml.

More preferably the concentration is 10 ⁹ cfu/ml. Typically, for post-harvest oranges the method comprises: washing the oranges with water; spraying the

fruit with bacterial cells from a bacterial culture; and allowing the fruit to dry. The method is typically followed by sorting the fruit and then transporting or storing the fruit. Prior to sorting oranges are usually waxed, however,the wax "Britseal" (Milestone Chemicals Pty Ltd, Melbourne, Australia) has been shown to reduce the effectiveness of the bacterial treatment and therefore may be replaced with other waxing materials. Alternative waxes which may be used include Shellac, Carnauba polyethylene and resin based waxes, (all available from Castle Chemicals Pty Ltd Australia) .

Alternatively a combination of bacteria and a growth regulator such as 2,4-D, preferably at a concentration of 250 ppm may be applied to the oranges.

This treatment overcomes the effect of wax such as Britseal which can then be applied after drying.

Typically the growth regulator is one which principally has auxin activity. In an alternative version, instead of spraying the fruit on the packing line, the fruit may be dipped in the bacterial cells from a bacterial culture either in the presence or absence of growth regulator, prior to loading onto the packing line. According to a third aspect of the present invention there is provided a novel antibiotic substance produced by ID 2131.

In order to improve the properties of this antibiotic, such as for instance its stability, derivatives of the antibiotic can be prepared using standard chemical techniques.

According to a fourth aspect of the present invention there is provided a method for protecting a plant or plant product from fungal infection which method comprises exposing the plant or plant product to the antibiotic of the third aspect.

According to a fifth aspect of the. present invention there is provided an antifungal composition comprising an

effective amount of strain ID 2131 or ID 2859 together with an agriculturally acceptable carrier or diluent compatible with maintaining viability of the strain.

The composition may additionally comprise an effective amount of a growth regulator, which is typically a regulator having principally auxin activity.

According to a sixth aspect of the present invention, there is provided an antifungal composition comprising an effective amount of the antibiotic substance of the third aspect or a derivative of it together with an agriculturally acceptable carrier or diluent.

The isolate of _______ cepacia ID 2131 ^' inhibits germination of P_j_ digitatum and shows strong ability to protect wounded postharvest oranges from }____ digitatum infection. The protective ability of this bacterial strain is closely related to cell concentration from 1.8 x 10 ⁵ to 10 ⁹ cfu/ml (r=0.867) . Treatment with £ _^ cepacia cells on unwounded Washington Navel and Valencia oranges under packing line conditions also significantly reduced Pj_ digitatum decay. Brown pigmentation appeared around some wound sites when the cell concentration was higher than 10 ⁷ cfu/ml and PAL activity around injured sites was suppressed together with an inhibition of lignin synthesis. According to a seventh aspect of the present invention there is provided a variant of strain ID 2131 in which the genes regulating the expression of the deleterious compound responsible for suppressing wound healing have been deleted. This variant could also be used in the methods of the invention.

Brief Description of the Drawings Fig. 1 shows the effect of Pseudomonas cepacia concentration and time interval on control of Penicillium digitatum infection of Valencia oranges. ( A A ) , P___ cepacia applied 2 hr before P _^ . digitatum;

( Δ __ ) , ______ cepacia applied with _?____ digitatum;

( ø ) , _^ cepacia applied 2 hr after _______ digitatum;

( o _© ) , P., _. cepacia applied 24 hr after _______ digitatum;

Vertical bar represents significant difference by Waller- Duncan's Bayesian K ratio LSD at k=100.

Fig. 2 shows the effect of Pseudomonas cepacia on PAL activity of wounded flavedo at healing sites of Valencia oranges at 25 or 30°C. Water or P_j_ cepacia (1.3 x 10 ⁹ cfu/ml) was applied to wounds made with sandpaper. Fruit were incubated at 96% relative humidity.

( A ____> ) , water then 25°C;

C .§ ) , water then 30°C; ( Λ ) , v___ cepacia then 25°C;

(Q Q) , P_^ cepacia then 30°C;

Vertical bars represent the standard errors about the means of two replicates.

Fig. 3 shows a scanning electron micrograph of P_ _^ cepacia ID 2131 cells on a wound site.

Fig. 4 shows that where cells of P _^ cepacia ID 2131 are abundant on a wound site, spores of P_ _s _ digitatum do not germinate.

Fig 5 shows that where cells of ______ cepacia ID 2131 are sparse on a wound site 9___ digitatum spores germinate.

Fig. 6 shows a trace of the purification of the effective antibiotic fraction from _____ cepacia ID 2131 cells collected from the reversed phase column on HPLC equipped with normal phase column.

CoIumniNova-pak silica (4 μm)

8 x 100 mm Solvent: CHG _j hexaπe (60:40) How Rate: 1 ml/min Chan Speed: 30 cm/ r Detector (UV): 254 nm

Peak 2 and 3 = solvent Peak 1 = effective fraction

Fig. 7 shows a 400 MHz ^X H NMR spectrum of the C ₁₈ column bleed which was run in solvent CD ₂ Cl ₂ for the antibiotic from ^~ ___ cepacia ID 2131 cells.

Fig. 8 shows a 400 MHz ¹ H NMR spectrum of the active sample which was run in solvent CD ₂ Cl ₂ for the antibiotic from P_j_ cepacia ID 2131 cells.

Fig. 9 shows the off-resonance decoupled ¹³ C spectrum of the antibiotic compound. The solvent was CD ₂ Cl ₂ and the spectrum produced at 100 MHz.

Fig. 10 shows the main fragments of the antibiotic that may form during electronic impact mass spectrometry which are supported by high resolution mass spectrometry.

(The ionic structures shown may not be the actual structures formed during the fragmentation process) .

Fig. 11 shows the effect of the purified antibiotic produced by strain ID 2131 on PAL activity at healing sites of wounded flavedo Valencia orange at 30°C. Water or the antibiotic (50μg/ml) was applied to wounds made with sandpaper. Fruit were incubated at 30°C with 96% relative humidity. Vertical bars represent the standard errors about the means of two replicates.

Fig. 12 shows the general structure of some terpenoid antibiotics isolated from microorganisms or plants.

Fig. 13a shows stereoscanning electron micrographs showing sparse cells of the strain ID 2859 of Pseudomonas glathei (upper) and germinated spores of Penicillium digitatum (lower) at wound sites on Washington navel oranges having been incubated at 25°C for 48 hr.

Fig. 13b shows stereoscanning electron micrographs showing the abundant bacterial cells (upper) and ungerminated fungal spores (lower) at wounded sites on fruit having been incubated at 30°C for 24 hr and then for a further 24 hr at 25°C.

Fig. 14 shows the effect of temperature on inhibition of Penicillium digitatum infection by strain ID 2859 of Pseudomonas glathei on wounded Washington navel oranges.

The fruit were wounded to 3 mm deep and inoculated with P. digitatum (6.7 x 10 ^5' spores/ml) , and then 2 hr later treated with the bacterium ID 2859 (1.7 x 10 ⁹ cfu/ml) . The fruit were either directly incubated at 25°C for 5 days or held at 30°C for 24 hr before being transferred to 25°C and incubated for a further 5 days before the symptoms were recorded. The different letters above each graph-bar represent significant differences between

treatments by Waller-Duncan's Bayesian K ratio LSD at k=100.

Best Method of Carrying out the Invention

The isolates described herein can be cultured under the conditions described in the following Examples, and can be applied to plants or plant products using techniques such as those described in the following

Examples.

Compositions of the invention may be prepared using standard techniques and carriers and diluents as used in preparation of agricultural compositions. Where the composition includes a growth regulator standard regulators may be used. Examples include 2,4-D and gibberellic acid. Where the compositions comprise the bacterial isolates, the carriers and diluents are selected from those compatible with maintaining the viability of the bacteria.

Example 1 Inoculum. A bacterial strain was isolated from the surface of Washington Navel oranges (Citrus sinensis Osback) at the Somersby research station, NSW Agriculture and Fisheries. After streak-separation and subculture on malt-extract agar (MEA) , the organism was identified from its fatty acid profile on a gas chromatography Microbial Identification System (Aerobe Library version 3.0, Microbial ID Inc, Newark DE) at NSW Agriculture and Fisheries Biological and Chemical Research Institute and produced the following fatty acid profile:

SUBSTITUTE SHEET

The strain was identified as Pseudomonas cepacia with the following identification results

TSBA [Rev 3.0] Pseudomonas 0.888

P. cepacia 0.888 P. c. GC subgroup B 0.888

P. gladioli 0.850

P. g. gladioli 0.850

It closely matched P _^ _ cepacia according to the library profiles (0.888, ID 2131) .

The bacterial cells were incubated routinely on MEA at 25°C for 5 days. The cells were collected with a sterilized loop and suspended in distilled water. The concentration was determined as colony forming units per ml.

P. digitatum spores were collected from decayed Washington Navel oranges and cultured routinely on MEA at 25°C. The spores for experiments were produced by wound- inoculating orange fruit, incubating the fruit at 25°C for 6 days and harvesting the conidia by brushing them into sterile distilled water containing 0.05% wetting agent 'Lisspol' . The suspension was filtered through 4 layers of muslin and adjusted to the required concentration. In vitro tests on the fungus. Tests on ___χ digitatum spore germination on slides were carried out by the same method as described in reference 7. For tests on agar plates 1 ml of Pj _. cepacia cell suspension (1.8 x 10 ¹⁰ cfu/ml) was placed on the centre of MEA in Petri dishes and incubated at 25°C for 24 hr. The plates were then dusted with __P___ digitatum spores and incubated at 25°C for 48 hr before measuring diameters of inhibition zones around the bacterium. Then, pieces of agar in the inhibition zones and 0.5 cm away from bacterial colonies were removed and transferred onto fresh medium (MEA) . After being left at 5°C overnight the plates were dusted with _P___ digitatum spores. After incubation for another 48 hr at 25°C, diameters of inhibition zones around the

agar pieces were measured.

Fruit and their inoculation. Washington Navel and Valencia oranges were harvested from Somersby station, NSW Agriculture and Fisheries. After being washed and sorted into uniform size, the fruit were sterilized with 80% ethanol and injured to a depth of 3 mm at four equidistant points around the equator. Generally, fruit were treated by pipetting 20 μl of a suspension of P. cepacia cells onto each injured site and 2 hr later, by pipetting 20 μl _____ digitatum spores. Alternatively as in Table 1-2, the bacterial cells were applied as sprays or dips to the whole fruit. As shown in Fig 1, fruit were treated by pipetting 20 μl _____ cepacia at different concentrations, and suspensions of P^. digitatum spores were dropped into wounded sites at the time intervals of 24 hr earlier, 2 hr earlier, coincident or 2 hr later relative to P_ _^ cepacia or water treatment .

The experiments were arranged in a completely randomized design. Each treatment was replicated three times with ten fruit for each replication. Amounts of infection were measured by assessing presence or absence of green mold symptoms after 5 days at each of the four inoculation sites. Results were measured as the percentage of inoculation sites showing symptoms and data were transformed as arcsines. The transformed data were analysed for variance and the means were compared using a Waller-Duncan's Bayesian K ratio LSD at k=100 (16) .

Fruit and their treatment on a packing line. Washington Navel and Valencia oranges were carefully harvested from Somersby station, NSW Agriculture and Fisheries. Fruit were sorted into uniform sizes and divided into groups according to experiments .

P. cepacia cells were collected from MEA and transferred into nutrient broth. After incubation at 25°C for 72 hr the cultures were centrifuged at 10,000 g for approximately 15 minutes and the bacterial cells were resuspended in distilled water for spraying onto fruit.

Washington Navel oranges were treated on the packing

line by spraying with guazatine (500 μg/ml as Panoctine 40% a.i. from Shell Chemicals, Australia) or P_ _^ cepacia cell suspension at 10 ⁹ cfu/ml, each mixed with a spore suspension of P_,. digitatum. After treatment, the fruit were stored at ambient temperature for 40 days.

Valencia oranges were treated similarly on the packing line except that imazalil (500 μg/ml as Fungaflor 75% a.i. from Janssen Cilag, Sydney, Australia) was used to replace guazatine and the plant growth regulator 2,4-D amine 50 at 250 μg/ml (Amalgated P/L Australia) was tested alone or mixed with P_;_ cepacia cell suspension in some treatments. A suspension of 5.6 x 10 ⁶ spores/ml of P. digitatum was sprayed onto fruit after the treatments. The fruit were stored at 12°C for 70 days. A water-based wax 'Britseal' (Milestone Chemicals Pty Ltd, Melbourne, Australia) was used on some batches of both Washington Navel oranges and Valencia oranges throughout the experiments, being applied immediately after spraying fungicide or the bacterial suspension. The experiments were arranged in a completely randomized design. Each treatment was replicated three times with sixty fruit for Washington Navel oranges and fifty for Valencia oranges per replication. Amounts of infection were measured by assessing presence or absence of green mold symptoms on individual fruit. Results were expressed and analysed as described above.

Examination of lignin deposition and PAL activity.

Valencia oranges were artificially injured by rubbing with 60-grit sandpaper and treated by pipetting a 40 μl cell suspension (1.2 x 10 ⁹ cfu/ml) of P_,_ cepacia onto each wound. Control fruit received the same amount of distilled water. The treatment was replicated twice with ten fruit for each replication. Fruit were then held in plastic bags (RH>96%) at 25°C and 30°C, respectively.

Fruit were examined for lignin deposition by using toluidine blue 0 and phloroglucinol-HCl . Segments were removed from injured sites and fixed in 95% ethanol for

16 hr. Ten hand-sections were made from each segment and stained with either toluidine blue O or phloroglucinol- HC1. A total of 10 segments were examined per treatment. PAL activity was measured by methods previously described (8,13) . Five g flavedo tissues from the injured area were removed and held in 100 ml cold acetone at -15°C for 2 hr. Acetone was changed twice during this period. After homogenising in a Sorvall Omni-mixer for 5 min at the highest speed, homogenates were filtered through Whatman no. 1 papers and the residues were washed with 100 ml cold acetone. Residues from the filter paper were allowed to dry at room temperature for one hr and then in a desiccator overnight at 4°C. The residues were stored at -15°C until used. PAL assay was carried out in 10 ml sodium borate buffer (0.1M, pH 8.8) . Concentration of cinnamic acid was measured at 269 nm on a Cary 2200 spectrophotometer. PAL activity was expressed as μkat/mg protein. One katal is defined as the amount of enzyme converting 1 mole of substrate per second. Protein was determined by the method described by Bradford (2) . The experiment was repeated once. Observation under scanning electron microscope.

Washington Navel orange were wounded 3 mm deep and treated with _____ cepacia cells at 10 ⁹ cfu/ml. Two hours later, the wound site was challenged with _______ digitatum spores at 10 ⁵ spores/ml. After the fruit were incubated at 25°C for 48 hr, the tissues at wound sites were removed (0.8 x 0.2 x 0.4 cm) and fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.3) for 1 hr. After being postfixed in 1% 0s0 ₄ in the same buffer for 2 hr, the material was then dehydrated through an ethanol series, critical point-dried, sputter-coated with gold and examined under a scanning electron microscope. Results The germination of _____ digitatum spores was strongly inhibited by P _^ cepacia on glass slides and agar surfaces (Table 1-1) . Similar inhibition was brought about by substances detected around the bacterial cells in agar.

These substances could be transferred from one agar to another.

Each of the methods of applying _______ cepacia cells onto wound sites on oranges gave substantial control of P_ _^ _ digitatum applied 2 hr later to the same site (Table 1- 2) .

The protective ability of P_ _s _ cepacia was influenced by its cell concentration and time interval between P. cepacia treatment and _?___ digitatum inoculation (Fig. 1) . When applied 24 hr after inoculation with _____ digitatum, P. cepacia was ineffective. When applied 2 hr after or at the same time, E\_ cepacia gave significant protection at high concentrations. Application 2 hr before inoculation increased the effectiveness of protection at each concentration of the bacterium. Generally, increasing the bacterial concentration above an order of 10 ⁷ cfu/ml substantially decreased P _^ _ digitatum decay.

Concentrations of bacteria higher than 10 ⁷ cfu/ml caused brown pigmentation in citrus cells around some wounds (Table 1-3) . A light brown colour appeared 3 days after wound inoculation, and dark brown another day later.

Figure 3 shows a scanning electron micrograph of . cepacia cells on a wound site. Where cells of _____ cepacia were abundant, spores of P _.^. digitatum had not germinated

(Figure 4) . Where the cells were sparse, |\_ digitatum spores had germinated (Figure 5) .

A plant growth regulator 2,4-D, commonly used on citrus fruit for long term storage and for keeping the button alive, did not inhibit ]_____ digitatum infection and when mixed with P_j_ cepacia cells it did not significantly affect biocontrol on artificially wound inoculated fruit (Table 1-4) .

PAL activity underwent major increases during the healing period at wound sites, but these increases were strongly prevented in the presence of ______ cepacia (Fig.

2 ) . Lignin synthesis was also inhibited around wounds treated with P_;_ cepacia as revealed by stain responses to

toluidine blue 0 and phloroglucinol-HCl (Table 1-5) .

Tables 1-6 and 1-7 give results of trials under packing line conditions using carefully handled and unwounded fruit, showing significant biocontrol of P _^ _ digitatum in unwaxed fruit . The effectiveness of biocontrol evident on unwaxed fruit seemed to be prevented by waxing. Adding 2,4-D into the P _^. cepacia suspension appeared to overcome the deleterious effect of waxing and was also beneficial in unwaxed fruit . The isolate of P____ cepacia ID 2131 was strongly effective against P _^ . digitatum infection on postharvest oranges under both laboratory and packing line conditions. Satisfactory protection by this strain was achieved with sufficient bacterial cells surrounding P___ digitatum spores and at the time before }X_ digitatum inoculation. If P _^ digitatum spores were not surrounded by enough P _^ cepacia cells they could still germinate and result in a successful infection. It appears that _____ digitatum spores are inhibited by antifungal substances produced by the bacterial cells. Antifungal antibiotics have been described as products of strains of |\_ cepacia

(5, 15) . The yield of these substances might be related to the number of bacterial cells. Increasing the number of bacterial cells significantly decreased green mold decay.

The effectiveness of _?___ cepacia in suppressing P. digitatum infection was high when applied just before or after the fungus but lost when applied a day later. This is typical of biocontrol of plant diseases which is used as a basic strategy in post-harvest biocontrol (14,20) .

"Britseal" wax was unfavourable to the _?___ cepacia strain when it was applied immediately onto fruit treated with the bacterial cells. A similar result was also found with strain B-3 of _______ subtilis which was used to control brown rot on peach by Pusey et al (1986) (12) . They found that although B-3 was generally compatible with commercial fruit wax, negative effects of wax on B-3 activity occurred in some instances. An alternative

application method could involve applying bacterial cells in a bulk-dip stage before fruit enter the packing line, leaving a longer interval time between _^. cepacia treatment and waxing, or using a more compatible form of wax.

Biological and chemical methods are by no means strict alternatives and enough evidence already exists to show that they can sometimes be combined with advantage (14) . Addition of 2,4-D (250 μg/ml) to the bacterial cell suspension seemed to overcome the disadvantage resulting from immediate wax application. The reason is unknown. 2,4-D is generally regarded as being effective in delaying citrus senescence and reducing susceptibility of citrus fruit to pathogen infection (3) . It has been substantially used in long term storage, particularly at ambient temperatures (6) . Other plant growth regulators are under investigation for their use on citrus fruit (1) . Although addition of 2,4-D to bacterial cells would negate the concept of total biological control without any added chemicals, compatibility of _______ cepacia cells with these plant growth regulators and their benefit in making the bacterial cells more compatible with commercial fruit waxing, is indicated.

As a response to wounds, living cells adjacent to damaged cells quickly synthesise lignin and phenolic substances associated with a significant increase of PAL activity (8) . The P _^ cepacia cells applied to the injured area suppressed wound healing with inhibition of PAL activity and lignin formation and resulted in brown pigmentation around some wound sites. This was possibly because substances produced by the bacterial cells were absorbed by damaged host cells, resulting in death of some plant cells adjacent to the injury. The brown pigmentation probably results from phenol oxidation in dead and neighbouring cells. The activity of PAL and lignin synthesising enzymes might have been prevented because of death of host cells or directly inhibited by the substances from P_ _^ cepacia. Another strain of _P____

cepacia which promoted P_ _s _ digitatum infection also suppressed PAL activity and lignin formation but it did not cause brown pigmentation around wounds (7) .

An hypothesis is that the protective strain of P_;_ cepacia produces both antibiotics and wound-healing inhibitive substances. It is therefore able to protect wounded fruit from E\_ digitatum infection through the antibiotics. On the other hand, the promotive strain (7) of _ _s _ cepacia may lack ability to produce antibiotics but may produce the substances which are responsible for suppression of PAL activity and lignin formation, thus favouring P _^ digitatum infection on undefended ^* wounded oranges. Under this hypothesis, the antifungal substances produced by the protective strain are different from the substances that suppress wound- healing.

Table 1-1. Effect of Pseudomonas cepacia (ID 2131) on germination of Penicillium digitatum spores.

One ml of P___ digitatum spore suspension (1.3 x 10 ⁵ spores/ml) was mixed with an equal amount of distilled water or £ _^ _ cepacia cell suspension (1.9 x 10 ⁸ cfu/ml) . These mixtures were supplemented with 1 ml fresh juice squeezed from orange fruit and diluted to 50% in distilled water. Germination was assessed on 100 spores in each of six replicate drops (20 μl) per treatment and germ tube length was measured from germinated spores among them.

² Measured as diameters around 1.5 cm wide colonies of P. cepacia or around agar pieces (0.5 x 0.8 cm) cut from inhibition zones and 0.5 cm away from 72-hr-old colonies on MEA, 48 hr after dusting with P _^ _ digitatum spores over the treated MEA surface.

³ Figures followed by the same letter in the Table are not significantly different by Waller-Duncan's Bayesian k ratio LSD at k = 100.

Table 1-2. Effect of Pseudomonas cepacia (ID 2131) on inhibition of Penicillium digitatum infection on Washington Navel oranges using different application methods under laboratory conditions.

^■ " ^■ Bacterial cells were suspended in distilled water at 1.9 x 10 ⁹ cfu/ml and applied to artificially injured

Washington Navel oranges. Cells were either applied as drops to wound sites (20 μl/wound) or sprayed onto the whole fruit with particular attention to injured areas.

In another method, the injured fruit were dipped in the cell suspension for a few seconds. Two hr after these treatments,a £_;_ digitatum spore suspension (2.6 x 10 ⁵ spores/ml) was pipetted into wound sites (20 μl/wound) .

The fruit were held at 25°C in plastic bags for 5 days.

² Figures followed by the same letter in the Table are not significantly different by Waller-Duncan's Bayesian k ratio LSD at k = 100.

Table 1-3. Effect of cell concentration of Pseudomonas cepacia (ID 2131) on Penicillium digitatum infection and on responses at inoculation sites in injured Valencia oranges.

^■ ""Valencia oranges were injured and treated by pipetting bacterial cells into wound sites at different concentrations followed by P _^ _ digitatum challenge. The fruit were held at 25°C for 5 days.

² Percentage of wounded sites that became brown after 3 days.

³ Diameter of brown tissue at wound sites after 3 days.

⁴ Figures followed by the same letter in the Table are not significantly different by Waller-Duncan's Bayesian k ratio LSD at k = 100.

Table 1-4. Effect of plant growth regulator (2/ 4-D) on biocontrol of Penicillium digitatum infection in

Washington Navel oranges under laboratory conditions.'

^• " ^■ Fruit were wounded and treated by applying drops of ID 2131 bacterial cell suspension with and without 2,4-D to each wound site. Two hr after treatment, P _^ digitatum spores at 6.5 x 10 ⁵ spores/ml were pipetted into wound site and the fruit were held at 25°C for 5 days. Figures followed by the same letter in the Table are not significantly different by Waller-Duncan's Bayesian k ratio LSD at k = 100.

Table 1-5. Influence of Pseudomonas cepacia (ID 2131) on the staining response of injured tissues of Valencia oranges at different stages during healing ¹ .

^■ " ^■ Valencia oranges were artificially injured by rubbing with 60-grit sandpaper and treated by pipetting a 40 μl cell suspension (1.2 x 10 ⁹ cfu/ml) of P _^ _ cepacia onto each wound. Control fruit received the same amount of distilled water. The treatment was replicated twice with ten fruit for each replication. Fruit were then held in plastic bags (RH>96%) at 25°C and 30°C, respectively. Segments were removed from injured sites and fixed in 95% ethanol for 16 hr. Ten handsections were made from each segment and stained with either toluidine blue O or Phloroglucinol-HCl. A total of 10 segments were examined per treatment.

+ = complete staining ring around wound with a blue-green colour for Toluidine Blue 0 and red colour for phloroglucinol-HCl

± = incomplete staining ring around some wound sites - = no staining in each section.

Table 1-6. Comparison of effect of Pseudomonas cepacia (ID 2131) and a fungicide on infection by Penicillium digitatum in unwounded Washington Navel oranges under packing line conditions.

¹ Washington Navel oranges were sorted for uniform size and divided into groups for treatments. Each treatment consisted of sixty fruit and was replicated three times. The fungicide or bacterial suspension was mixed with Penicillium digitatum (5.6 x 10 ⁶ spores/ml) and applied onto fruit on a standard packing line. Fruit were stored at ambient temperature (12.5-17.5°C) for 40 days.

² Figures followed by the same letter in the Table are not significantly different by Waller-Duncan's Bayesian K ratio LSD at k = 100.

Table 1-7. Comparison of effect of Pseudomonas cepacia (ID 2131) and a fungicide on infection by Penicillium digitatum in unwounded Valencia oranges under packing line conditions ¹ .

¹ Valencia oranges were sorted for uniform size and divided into groups for treatment . Each treatment consisted of fifty fruit and was replicated three times. The control agents were applied to fruit on a standard packing line. Penicillium digitatum (5.6 x 10 spores/ml) was sprayed onto fruit after treatment. Fruit were stored at 12°C for 70 days.

² Figures followed by the same letter in the Table are not significantly different by Waller-Duncan's Bayesian K ratio LSD at k = 100.

Example 2 Studies on compounds produced by an isolate of Pseudomonas cepacia (ID 2131) and their antibiotic effect on Penicillium digitatum.

Antibiotics play an important role in the biocontrol of plant diseases (4) . A few antibiotics produced by P. cepacia have been characterised (5) and one of them was found to be effective in control of Botrvtis cinerea and Penicillium expansum on apple (15) . No antibiotics have been reported effective against citrus pathogens.

The effect on control of _____ digitatum infection by the isolate of P_ _^ cepacia shown 'in Example 1 appeared to be due to the antifungal substances produced by the bacterium. We have isolated the effective compound and tested its effect on inhibition of _9___ digitatum infection on citrus fruit .

Materials and Methods Equipment

Nuclear magnetic resonance (NMR) spectra were obtained on a Bruker AMX 400 spectrometer at 400 MHz (-""H) and at 100 MHz ( ¹³ C) . -""H-^-H correlated spectroscopy (H -

H COSY) spectrum was recorded on a Bruker AMX 600 spectrometer at 600 MHz.

Electron impact (El) mass spectra were determined at 70 eV on a spectrometer which was equivalent to a Kratoε MS 50 with a VG 3D8 control console at a source

temperature of 250°C. A high resolution mass spectrum was obtained on the same instrument. The El mass spectrum was further confirmed by a comparison with a chemical ionisation (Cl) spectrum obtained on a Hewlett Packard 5890 series II Gas Chromatograph with a Hewlett Packard 5989A mass spectrometer as detector. The gas chromatograph was equipped with an HP1 column (crosslinked methyl silicone gum, 12 m x 0.2 mm x 0.33 μm film thickness) . The initial temperature was set at 60°C and gradually increased to 200°C at a rate of 15°C per minute. The temperature of the injector was 200°C, while the detector was at 250°C.

High performance liquid chromatography (HPLC) was performed with a Waters 600E apparatus equipped with a Waters UV detector 490E. An infrared spectrum (IR) was measured on a FTS 20/80 (BIO-RAD, Digilab Division) .

Incubation of the bacterium

The strain of P _^ _ cepacia was incubated in broth consisting of malt extract 20 g, peptone 1 g, dextrose 20 g in 1 L water. Conical flasks (1 L ) were charged with 200 ml of the above mixture. After having been sterilised at 121°C for 15 min and cooled down, the flasks were inoculated with the bacterial cells collected from agar surface on MEA and incubated at 120 rpm on a Gallenkamp orbital shaker at 25°C for 3 days.

Isolation and purification of compounds

The fermentation broth (50L) was centrifuged at 10000 g for 10 min and the supernatant liquid was collected and partitioned three times with equal amounts of chloroform at room temperature. After vacuum drying at 40°C, the remaining syrup (4.75 g) was taken up successively with 50 ml benzene followed by 35 ml benzene-ethyl acetate mixture (80:20) . Samples (2 ml) were loaded onto each Sep-pak silica cartridge (Millipore, classic feature with 2 ml cartridge hold-up volumes) and eluted stepwise at a

flow rate of 2 to 4 ml/min with 2 ml of the following solvents for each cartridge : (a) benzene, (b) 80% benzene/20% ethyl acetate, (c) 50% benzene/50% ethyl acetate, (d) 20% benzene/80% ethyl acetate, and (e) ethyl acetate. Eluates were collected as 2 ml fractions and each fraction was dried by evaporation using a rotary film evaporator at 40°C. The residues were dissolved in acetone and checked for biological activity jin vitro. The effective fractions (3.7064 g) were dissolved in acetone and applied onto silica gel TLC plates (20 x 20 cm, gel layer 250 μm, with 254 nm fluorescent indicator on glass, SIGMA) which were developed by ascending chromatography in chloroform/petroleum spirit/methanol (60:30:10) . The plates were inspected under a UV lamp. The silica gel corresponding to each fluorescent or quenching band, and the gel between the UV-responding bands, were scraped into a centrifuge tube. The silica gel was then crushed and extracted three times with acetone and the extract checked for biological activity n vitro. Only one quenching band at Rf 0.43 inhibited the growth of P. digi tatum in vi tro . The effective fraction was then compared with pyrrolnitrin (obtained from Dr Janisiewicz, U.S. Department of Agriculture, Agricultural Research Service, Appalachian Fruit Research Station, WV 25430, U.S.A.) on silica gel TLC developed in CHCl ₃ /petroleum spirit/MeOH (60:30:10) and sprayed with diazotized sulfanilic acid (DSA) .

The effective fraction (17.5 mg) was dissolved in the mobile phase solvent (65% acetonitrile and 35% water) for HPLC purification. The HPLC was equipped with a reverse phase (C ₁₈ ) column (6μm, 25 x 100 mm, Nova-pak, Millipore water) . Isocratic elution was employed with 65% acetonitrile and 35% water at a flow rate of 6 ml/min. The UV detector was set at 230 nm. The fractions were detected by UV absorption and the eluates between each UV absorbing peak were collected separately and tested for bioactivity. The effective fraction appeared at 8 mm after sample injection.

The effective fraction (2.2 mg) was further purified on a HPLC semi-preparative column (lOμm, 9.4 x 500 mm, ODS-3 M9/50, Whatman) with acetonitrile/water (60:40) at a flow rate of 4 ml/min. The fractions were collected and tested as above. The effective fraction was obtained at 12 min after sample injection.

After the separation of the samples was completed the columns were eluted by using the same procedures for the sample separation except that the mobile phase solvent rather than the samples were injected. The eluates were collected as a column washing control shown in Fig. 7.

The effective fraction from the semi-preparative column was then further purified on a Nova-Pak silica

HPLC column (4 μm, 8 mm x 100 mm) with CHCl ₃ /hexane (1:1) at a flow rate of 1 ml/min. A single peak which appeared at 6 min after sample injection and the eluate before and after this peak, which did not give UV absorption, were collected separately (Fig. 6) . The material from the UV absorbing peak weighed 1.3 mg after vacuum drying. The dried pure active fraction (1.3 mg) was redissolved in carbon tetrachloride (CC1 ₄ ) and vacuum dried. The residue was taken up with deuterated dichloromethane (CD ₂ C1 ₂ ) for examination by NMR. The sample was then dried with nitrogen gas and redisolved in acetone for mass spectrometry, chloroform for IR and methanol for UV spectroscopy.

Test of fractions in vitro

The fractions obtained throughout the experiment were dried by rotary evaporation (40°C+_2) and assayed for antifungal activity in an agar diffusion test. Fractions were dissolved in acetone and placed in a well (0.5 cm in diameter) made in the centre of Petri plates containing 15 ml MEA. After having been left for 5 hr, plates were seeded by dusting J\_ digitatum spores and incubated at 25°C for 24 hr and inhibition zones around the wells were recorded.

In a separate experiment, a crude extract from 2000 ml fermentation broth was dissolved in methanol and

tested for its antifungal activity as above after being heated in a water bath at 20, 40, 60, 80 and 100°C for 30 min. respectively.

Test of the antibiotic in vivo A crude extract (215 mg) was obtained from 2000 ml fermentation broth and dissolved in a small amount of acetone and then diluted with water to 10 ml . Valencia oranges were artifically wounded and inoculated with P. digitatum spore suspension (4.8 x 10 ⁵ spores/ml) as described in Example 1. Fruit were treated by pipetting 20 μl of water, bacterial cells or crude extract into wounded sites at the different time intervals shown in Table 2-5.

In another experiment, Valencia oranges were inoculated with P. digi tatum at 5.9 x 10 ⁵ spores/ml 10 hr before being treated with the purified antibiotic (50 μg/ml) . The fruit were kept at 25°C and green mould decay was recorded as above. To investigate the effect of the antibiotic on host PAL activity, 40 μl purified antibiotic (50 μg/ml) was applied to each injured area on Valencia oranges and PAL was measured by the method described in Example 1.

Results Isolation and purification of antifungal metabolite

Isolation of the antifungal substance was monitored by bioassay with P. digi tatum . The unpurified extract was reasonably stable, its antifungal activity being maintained for up to two months at 5°C and still existing after heating up to 80°C in a water bath for 30 min

(Table 2-1) . The nearly pure antibiotic was unstable and lost its activity after being kept at 5°C for about two weeks. However, it could be stored in dichloromethane in the dark at -20°C to -70°C for an extended period up to one month without substantial loss of activity.

The column washing eluate (column bleed) did not give UV absorption on HPLC and could not be detected in mass spectrometry under the conditions used for the sample

measurement, but it could be measured on NMR (400 MHz) which indicated a -CH ₂ - chain-like structure (Fig. 7) . The NMR spectrum of the active sample showed three similar signals at δ 0.9, 1.35 and 1.5 ppm which appeared to represent contamination by the column bleed (Fig. 7) .

Characterisation of the antibiotic

Pyrrolnitrin had an Rf of 0.70 and gave a maroon colour with DSA. The active fraction extracted from ID 2131 culture had by comparison an Rf of 0.43 and gave a light yellow colour with DSA on a TLC plate developed in CHC1 ₃ /petroleum spirit/MeOH (60:30:10) .

Characterisation data of the isolated antibiotic are listed in Table 2-2. The molecular weight of this antibiotic was determined as 306 according to its Cl spectrum which gave an intense peak at 307 and a loss of H ₂ 0 (M-18) . The El spectrum had fragmentation ions at, 306(19), 288(20), 264(45), 246(20), 245(20) , 237(38), 227(5), 217(5) , 195(8), 175(9) , 165(10) , 149(17), 127(12), 109(19), 97(25) , 95(22), 81(20) , 69(27), 59(7), 55(43), 43(100) . The molecular formula was established as C ₁₉ H ₃₀ O ₃ (M ⁺ = m/z 306.2229, C ₁₉ H ₃₀ O ₃ calculated as 306.2195) as determined by high resolution (HR) mass spectrometry. The HR mass spectrum provided the following fragmentations at, 306.2229 (C ₁₉ H ₃₀ O ₃ ) , 288.2071 (C ₁₉ H ₂₈ 0 ₂ ) , 264.2067 (C ₁₇ H ₂₈ 0 ₂ ), 246.1661 (C ₁₆ H ₂₂ 0 ₂ ) , and 237.1780 (C ₁₅ H ₂₅ 0 ₂ ) . Assuming a norditerpenoid structure, a likely scheme for El and HR fragmentations is given in Fig. 10. The ^-H NMR and ¹³ C NMR spectra indicated that the compound had a terpene-like structure (Table 2-3, Fig. 9) . Further support for this type of structure came from the ""-H - ¹ H correlation spectrum (COSY) (CD ₂ C1 ₂ ) indicating a complicated coupling pattern typical of cyclic terpenes (Table 2-4) .

The existence of the -C=0 was indicated by

the IR spectra with an IR absorption at -1709 cm ^-1 . The presence of another functional group -C-OH was indicated by the IR spectrum (-2931 cm ^"1 ) and supported by a loss of one molecule of H ₂ 0 in the mass spectrum. Resonance at δ 67 ppm in the ¹³ C NMR spectrum indicated the existence of one alcohol group. Additional ¹³ C NMR resonances at δ 195, 130 and 137.5 were assigned to the presence of -C=0 and -C=C- groups indicated by the IR data. The chemical shift (δ 5.6) and the multiplicity of a single -""H signal of the -"Ή NMR indicated the presence of the system:

C-C-C=C-C

II I I

0 C H This was supported by the evidence that the protein resonating at δ 5.6 appeared as a multiplet coupled to the proton resonating at δ 2.00 as indicated by COSY experiments.

The ¹ H - -"-H correlation spectrum (COSY) also indicated the presence of protons at δ 1.5 ppm and δ 1.34 ppm overlapped by the contaminants from the C ₁₈ column (Table 2-4) . It was likely that the proton at δ 0.9 in the compound was from the column bleed since it also appeared in the column eluate as seen in its -""H NMR spectrum. The proton at δ 1.35 coupled to that at δ 0.9 indicated that this proton might also be from the column bleed, as was the proton at δ 1.5 which was coupled to that at δ 1.35. However, the other proton at δ 1.5 was also coupled to those at δ 2.5 and 4.05, indicating its relationship with these protons in the active compound. Similarly, the proton at δ 1.35 was coupled to that at δ 2.0, being another proton in the antibiotic. The protons at δ 3.45 and 6.3 were coupled to that at δ 0.9, indicating that they were also impurities as was the hydrogen at δ 9.5 which was coupled to that at δ 3.45 ppm.

Biological properties of the antibiotic The crude extract from P. cepacia culture gave excellent control of P . digi tatum infection (Table 2-5) .

Its protective ability was even stronger than the bacterial cells themselves, particularly in being able to control the disease when applied 24 hr after inoculation with the pathogen. The purified antibiotic at 50 μg/ml significantly inhibited P. digi tatum infection but did not suppress host PAL activity (Table 2-6; Fig. 11) .

Antibiosis is hypothesised to be a major mechanism of biocontrol according to the observation of inhibition zones on agar plates in many interactions between plant pathogens and biocontrol agents (32) . The antifungal compound extracted from the fermentation broth of the strain ID 2131 of P. cepacia had a different structure to the previously described antibiotics from this species, pyrroles, pseudanes and cepacins. The data resulting from the current experiments suggests that it is a cyclic nor-diterpene (C19) .

Cyclic terpenes as a class of natural products display a bewildering array of structural types. Over a hundred different carbon frameworks have been detected so far (31) . Among the well known substances are the gibberellins, plant growth hormones, which were first isolated from a fungus Gibberella fujikuroi , the cause of

'foolish disease' of rice plants (30) . A few terpenoid antibiotics have also been obtained from micro-organisms and plant (22; also see Fig. 12) . Pleuromutilin is an antibacterial substance which was isolated from a Basidiomycete , Pleurotus mutilus (33) . Fusidic acid, another antibiotic, was extracted from Fusidium coccineum

(23) . The work by Seubert (29) investigating the pathway of forming cyclic terpenes in the utilisation of geraniol and related compounds in Pseudomonas ci tronelloi indicated that these compounds could be produced in this genus of bacteria.

The biological formation of cyclic terpenes has been traced back to the operation of a unifying biosynthetic scheme (27) . Based on the pattern of proton coupling in

COSY, -""H and ¹³ C NMR data and comparing these data with those of typical terpenes in the literature (24 and 26) ,

the antibiotic produced by ID 2131 of P. cepacia appears to be a nor-diterpene. A proposed structure for this antibiotic is shown in Fig. 10.

The antibiotic extracted from the growth culture of the strain ID 2131 was strongly inhibitive to infection by P. digi tatum on oranges. This was particularly distinctive when the pathogen had been applied to wounded sites 24 hr earlier, under which conditions the antagonistic cells were not able to provide their protection. This indicated that the antibiotic effect observed in vi tro could also be seen in the inhibition of P. digi tatum infection in vivo .

The purified antibiotic did not suppress host PAL activity, indicating that the inhibition of P. digi tatum spore germination and suppression of host wound healing by the bacterial cells of the strain ID 2131 and its growth culture came from different compounds, which must be controlled by different genes. Therefore, the genes regulating the production of the deleterious compound responsible for suppressing host wound healing should be capable of being deleted and the biocontrol effect by this strain thereby improved. Such genetic manipulations have been done on Agrobacterium radiobacter K84 that possesses a transferable region on its plasmid that often results in decrease or loss of biocontrol effect (25) . Biocontrol effect has been significantly improved after a deletion of this region, with the production of a new strain K1026 (28) .

Table 2-1. Effect of heating the extracted compound at different temperatures on growth from Penicillium digitatum spores ¹ .

^■ " ^■ The crude extract was heated in a water bath for 30 min at the temperatures shown above. After heating the extracts were tested by placing 20 μl of the solution into wells made on MEA. After being left at 7°C for 5 hr the plates were dusted with P _^ _ digitatum spores and incubated at 25°C for 24 hr before measuring inhibition zones as diameters.

Table 2-2. Physical properties of the antibiotic.

Test Properties

Appearance Amorphous

Empirical formula (MW) C _I9 H ₃₀ O ₃ (306.2229 ¹ )

IR υ^ cm ^"1 2931, 1709, 1362, 1231 Rf on silica gel ³ 0.43

C _!9 H ₃₀ O ₃ requires 306.2195 (+ 0.0034).

Solvent: chloroform/petroleum spirit/methanol (60:30:10).

Table 2-3. ¹ H NMR data for the antibiotic obtained dichloromethane - d, (CD,C1,) at 400 MHz.

¹ Chemical shifts (δ) are in parts per million downfield from TMS.

² The same signal appeared in the C ₁₈ column bleed.

³ The signal of protons in the compound was overlapped by the contaminants from the C, ₈ column bleed, but the protons at δl.35 and 1.5 ppm could be seen in COSY which were coupled to other protons in the active compound.

3065

- 32 -

Table 2 -4 Correlation spectroscopy of H - H Cosy at

600 MH 1

The spectra were run in dichloromethane-d ₂ .

Table 2-5. Effect of cells of Pseudomonas cepacia (ID 2131) and extract from their culture on decay in wounded oranges inoculated with Penicillium digitatum ¹ .

Fruit were wounded to a depth of 3 mm at 4 equidistant sites around the equator and then inoculated at these sites with either 20 μl of water, _______ cepacia cells (2.4 x 10 ⁹ cfu/ml) or crude extract (215 mg dissolved in 10 ml acetone-water solvent) derived from

2000 ml fermentation broth of _9___ cepacia culture. Drops (20 μl) containing _______ digitatum spores (5.2 x 10 ⁵ ) were applied to wounded sites at time intervals before or after relative to the time of applying the treatments as shown above. Fruit were held in plastic bags at 25°C for 5 days. ² Figures in the Table followed by the same letter are not significantly different by Waller-Duncan's Bayesian k ratio LSD at k=100. Data are the percentage of decayed fruit per treatment that developed green mould decay.

Table 2-6. Effect of the purified antibiotic from the growth culture of the strain ID 2131 of Pseudomonas cepacia on Penicillium digitatum infection of Valencia oranges ¹ .

Treatment Percentage of P. digitatum infection

Water 95.2 a ²

Antibiotic 0.0 b

¹ The antibiotic 20 μl/wound (50 μg/ml) was applied to the wounded sites on Valencia oranges 10 hr after P. digitatum (5.9 x 10 ⁵ spores/ml) inoculation. The fruit were incubated at 25°C for 5 days after P. digitatum inoculation.

² Figures in the table followed by the same letter are not significantly different by Waller-Duncan's Bayesian k ratio LSD at k=100. Data are the percentage of decayed fruit per treatment that developed green mould decay.

Example 3

Biocontrol of Penicillium digitatum infection on postharvest oranges by a strain of Pseudomonas glathei (ID 2859)

A number of antagonists have been reported to be inhibitive to postharvest fruit pathogens through antifungal substances (10,15) . Two yeasts that do not produce antibiotics are reported to be antagonistic to

Penicillium digitatum and V____ italicum infection on citrus

fruit (21) . However, their effect is- poor and less than that of antagonistic bacteria.

We have obtained an isolate of Pseudomonas glathei, which does not produce antibiotics but is able to protect wounded oranges from ________ digitatum infection.

Materials and methods Inocula

A bacterium was isolated from Valencia oranges and identified (ID 2859) as a strain related to Pseudomonas glathei by the MicroStation™ System, Biolog method (Anon, 1992 Release 3.01; Biolog Inc Hayward CA) . Isolate ID 2859 was also analysed for fatty acid profile on an HP 5898A Microbial Identification System at the NSW Agricultural and Fisheries, Biological and Chemical Research Institute.

It produced the following profile.

This profile identified the strain as being most closely related to Pseudomonas cepacia but the correlation was not particularly high. While the Biolog identification does not ^" definitely include the strain within Pseudomonas glathei we believe that this is the appropriate species designation for this strain.

The fatty acid identification results were as follows :

RT AREA Ar/Ht Respon ECL NAME COMMENT 1 COMMENT 2

1 . 579 28055000 0 . 056 7 . 027 Solvent Peak (Min rt

4.574 1809 0.038 1.033 12.000 12:0. 3.63 ECL deviates 0.000 Ref -0.002

9.319 2402 0.047 0.945 15.487 Sum In Feature 3 4.41 ECL deviates-0.003 14:0 30H/16:1 ISO I

9.859 5229 0.051 0.942 15.817 16:1 CIS 9 9.57 ECL deviates 0.000

10.158 12990 0.048 0.941 16.000 16:0 23.74 ECL deviates 0.000 Ref. -0.001

11.671 7103 0.051 0.938 16.888 17:0 CYCLO 12.94 ECL deviates-0.000 Ref. -0.001

12.762 3850 0.062 0.939 17.517 16:0 30H 7.02 ECL deviates-0.003

13.292 12837 0.052 0.940 17.822 Sum In Feature 7 23.45 ECL deviates-0.000 18:1 CIS 11/t 9/t 6

15.164 6290 0.054 0.949 18.900 19:0 CYCLO Cll-12 11.60 ECL deviates 0.000 Ref. 0.001

15.486 1960 0.052 0.952 19.087 18:1 20H 3.62 ECL deviates-0.001

16.540 963 0.486 19.702 Max ar/ht

17.373 2636 0.081 20.187 Max rt

19.689 20800 0.553 21.540 Max rt

2402 SUMMED FEATURE 3 4.41 12:0 ALOE? Unknown ID:928

16:1 ISO 1/14:0 14:0 30H/16:1 3OH ISO I

12837 SUMMED FEATURE 7 23.45 18:1 CIS 11/t 9/t 18:1 TRANS 6 9/6t/cll

18:1 TRANS 6/t9/cll

Solvent Ar Total Area Named Area % Named Total Amount Nbr Ref ECL Deviation Ref ECL Shift 28055000 55433 54470 98.26 51464 4 0.001 0.001

TSBA [REV 3.30] * NO MATCH*

The bacterial cells were collected from MEA and incubated in nutrient broth (nutrient broth 8 g, yeast extract 5 g, dextrose 10 g in 1 L water) on a shaker (120 rpm) at 25°C for 72 hr. The fermentation broth was centrifuged at 10000 g for 15 min and the bacterial cells resuspended in distilled water.

P. digitatum spore suspension was prepared as in Example 1.

In vitro tests on fungus Effect of the bacteria on germination of P _^ _ digitatum spores was tested on agar plates. One ml of P _^ . glathei cell suspension (3.9 x 10 ¹⁰ cfu/ml) was placed on the centres of MEA plates and incubated at 25°C or 30°C for 48 hr. The plates were then dusted with P _^. digitatum spores and incubated at 25°C for 48 hr before measuring diameters of inhibition zone around the bacteria. Fruit and their inoculation

Valencia oranges were artificially injured (3 mm deep) and inoculated with a P^ digitatum spore suspension (3.1 x 10 ⁵ spore/ml) at a rate of 20 μl per wounded site. Two hr later, the fruit were placed in plastic bags and divided into four groups. One group was kept at 25°C for 4 days, the other three were incubated at 30°C for a period of 24 hr, 48 hr or 72 hr, respectively. The fruit were then kept at 25°C for 6 days after being taken out from the 30°C environment. Water treatment was used as a control by the same method.

Experiments were arranged in a completely randomised design. Each treatment was replicated twice with 5 fruit for each replication. Results were measured as percentage of inoculation sites showing symptoms and data were transformed as arcsines. The transformed data were analysed for variance and the means were compared using a Waller-Duncan's Bayesian K ratio LSD at k=100.

Results ID 2859 did not produce any surrounding inhibition zone in the fungus at either 25 or 30°C conditions.

(Table 3-1) .

Application of the bacterium 2 hr after inoculation with the fungus was highly effective in controlling fungal decay in fruit healed at 30°C, but not in fruit at 25°C (Table 3-2) . It is interesting to note also that healing at 30°C without addition of the bacterium decreased the amounts of fungal decay. The beneficial effects of healing at 30°C increased with time from 24 to 72 hr at this temperature.

The strain of _X_ glathei did not produce visible inhibition zones in the fungus on agar plates. It gave an inhibition of P _^ digitatum infection on oranges at 25°C if the fruit were inoculated with P _^ digitatum 2 hr later than its application, but not if inoculated 2 hr earlier. It strongly protected the wounded fruit from P. digitatum infection if fruit were healed at 30°C for a period after inoculation with P___ digitatum and treated with the bacterium.

Table 3-1. Inhibition of Penicillium digitatum spore germination by Pseudomonas glathei ¹ (ID 2859) .

¹ The bacteria were incubated on MEA at 25°C or 30°C for 48 hr. After dusting with _____ digitatum spores, the plates were placed in an incubator at 25°C for 48 hr and the surrounding inhibition zones were measured.

Table 3-2. Effect of the strain of Pseudomonas glathei on control of Penicillium digitatum infection to wounded oranges.

¹ Valencia oranges were wounded and inoculated with P_ _s _ digitatum (3.1 x 10 ⁵ spore/ml) 2 hr before or after the time of applying the treatments as shown above. Then the fruit were kept in plastic bags at 25°C for 4 days.

Valencia oranges were inoculated with P _^ digitatum (3.1 x 10 ⁵ spore/ml) and two hr later treated with the bacterial cells (3.9 x 10 ¹⁰ cfu/ml) . After being left for another hr at 25°C, the fruit were incubated at 30°C for 24 hr, 48 hr or 72 hr, respectively. Fruit were then kept at 25°C for 6 days after being taken out from the 30°C environment.

³ Figures in the Table followed by the same letter are not significantly different by Waller-Duncan's Bayesian K ratio LSD at k=100.

Example 4

Biocontrol of Penicillium digitatum infection on post harvest oranges by a strain of Pseudomonas glathei (ID 2859)

Materials and Methods Inocula

The bacterium isolated from Valencia oranges and identified (ID 2859) as a strain related to Pseudomonas

glathei by the method of MicroStation™ System, BioLog (Anon, 1992) was used. Bacterium (ID 2131) was used for comparison in some experiments. The bacterial suspensions were prepared by culture in nutrient both at 25°C for 48 hr. The bacterial cells were collected by centrifugation at 10,000 g and suspended in distilled water. The concentration of antagonists was determined as colony forming units per ml.

Penicillium digitatum was isolated from decayed Washington navel oranges and then cultured routinely on malt extract agar (MEA, malt extract 20g, dextrose 20g, peptone lg in IL distilled water) at 25°C. Inocula for experiments were produced by wound-inoculating orange fruit, incubating the fruit at 25°C for 6 days and then harvesting the conidia by collecting them on sterilised cotton swabs for dusting, or by brushing them into sterile distilled water containing 0.05% wetting agent Lissapol (ICI, Sydney, Australia) for spraying. The suspension of Pj. digitatum spores was filtered through four layers of muslin and adjusted to the required concentration.

In vitro tests on the fungus

Effect of the bacteria on growth of P_ _^ digitatum was tested on agar plates. Cell suspensions of Pj. glathei (3.9 x 10 ¹⁰ cfu/ml) were placed in the centres of MEA plates at 100 μl/plate and incubated at 25°C or 30°C, respectively, for 48 hr. The plates were then dusted with £,,_ digitatum spores and incubated at 25°C for 48 hr before measuring diameters of inhibition zones around the bacteria. In another experiment, the same amounts of bacterial cells were loaded on to MEA. The plates were immediately dusted with P _^ _ digitatum spores and incubated at 30°C for 72 hr before being moved to 25°C for a further 48 hr. The growth of the fungus around the bacterial colonies was investigated under both temperature conditions.

Fruit and their inoculation

Washington navel oranges were obtained and

artificially injured to a depth of 3 mm using a sterile nail through a cork. The fruit were inoculated with P. digitatum (6.7 x 10 ⁵ spores/ml) and, 2 hr later, treated with the bacterial strain ID 2859 of }X_ glathei (1.7 x 10 ⁹ cfu/ml) by pipetting 20 μl of the bacterial suspension into the wounded site. They were either directly incubated at 25°C or held at 30°C for 24 hr before being transferred to 25°C for 5 days.

Valencia oranges were handled similarly. In one experiment as shown in Table 4-1, wounds on fruit were treated by pipetting 20 μl of the suspension of P. digitatum spores (3.1 x 10 ⁵ spores/ml) at intervals of 2 hr before or 2, 10 or 24 hr after pipetting 20 μl of the bacterial strains (ID 2859 at 1.9 x 10 ⁹ cfu/ml or ID 2131 at 1.2 x 10 ⁹ cfu/ml) or water onto each site. The fruit were held at 25°C for 4 days after P _^ _ digitatum inoculation.

In another experiment as shown in Table 4-2, 20 μl P. digitatum (3.1 x 10 ⁵ spores/ml) were inoculated onto each wound on fruit 4 hr before or 2 hr after 20 μl of the bacterium ID 2859 (1.9 x 10 ⁹ cfu/ml) or ID 2131 (1.2 x 10 ⁹ cfu/ml) or water. The fruit were held at 30°C for 24 hr, 48 hr or 72 hr respectively, and then incubated at 25°C for a further 5 days after having been moved out from the 30°C.

Recovery of the bacteria

Valencia oranges were washed and wounded to a depth of 3 mm at four equidistant points around their equators and each point was treated by pipetting 20 μl of the bacterial suspension (ID 2859) at 1.6 x 10 ⁹ cfu/ml or water as a control. The same volume of _?___ digitatum spore suspension (2.3 x 10 ⁵ spores/ml) was applied to each point 2 hr later. After being held at 30°C for 24 hr the fruit were moved to 25°C for 8 days. The incidence of P _^ digitatum infection was recorded and then the tissues at the inoculated sites were removed on the 5th and 9th days after _____ digitatum inoculation and incubated on MEA.

In an additional treatment, the fruit were treated as above with the bacterium (ID 2859) and J _ digitatum, but after 4 days incubation at 25°C the inoculated points on the fruit were washed using a syringe with 70% ethanol and then sterilised distilled water and immediately reinoculated with P___ digitatum (2.1 x 10 ⁵ spores/ml) . The fruit were held at 25°C for a further 4 days. Water treatment was used as a control in the same method. Observations under a scanning electron microscope Washington navel oranges were wounded to a depth of 3 mm and treated with bacterial cells of the strain ID 2859 at 2.3 x 10 ⁹ cfu/ml. Two hr later, the wound site was challenged with P___ digitatum at 6.5 x 10 ⁵ spores/ml. After the fruit were incubated at either 25°C for 48 hr or 30°C for 24 hr and then at 25°C for a further 24 hr, the tissues at wound sites were removed (0.5 x 0.2 x 0.3 cm) and examined using the method described by Huang et al (34) . Eight specimens from four fruit at each temperature were observed. Statistics

The treatments were arranged in a completely randomised design. Each treatment was replicated three times with 10 fruit for each replication. Amount of infection was measured by assessing the presence or absence of soft rot symptoms at each of the four inoculation sites on a fruit . Results were expressed as the percentage of inoculation sites showing symptoms, and data were transformed to arcsines. The transformed data were analysed as an analysis of variance, and means were compared with Waller-Duncan's Bayesian k-ratio LSD rule (k = 100) (35) . Results

The bacteria (ID 2859) grew normally on MEA at both 25°C and 30°C and did not inhibit surrounding fungal growth under these conditions. £_;_ digitatum did not grow on MEA at 30°C until the plates were removed to 25°C.

Figure 13a shows that bacterial cells of _?___ glathei (ID 2859) were sparse and fungal spores had germinated

after inoculated fruit were directly incubated at 25°C for 48 hr. However, when the fruit were incubated at 30°C for 24 hr before being moved to 25°C the bacterial cells became abundant and the fungal spores did not germinte (Figure 13b) .

At 25°C the bacterial strain ID 2859 was ineffective in biocontrol of fruit decay if it was applied 2 hr after the fungus (Figure 14; Table 4-1) . The biocontrol effect of ID 2859 significantly increased when intervals between applying the bacterium and then the fungus were increased from 2 to 24 hr (Table 4-1) . The biocontrol effect brought about by ID 2859 was inferior to that by ID 2131 unless the bacteria were applied 10 hr before the f ngus.

Strain ID 2859 was very effective in biocontrol even when applied 4 hr after the fungus, where fruit had been held at 30°C for 24 hr before being returned to 25°C. It was totally effective where fruit had been held at 30°C for 48 hr. It was superior to strain ID 2131, under these conditions (Figure 14; Table 4-2) . Both strains prevented infection when applied 2 hr before the fungus where fruit were held at 30°C for periods of 24 to 72 hr.

The bacterium ID 2859 could be readily recovered from all sites where it and the fungus had been applied (Table 4-3) . Interestingly, _P___ digitatum could also be recovered from most sites where it had been applied alone or with the bacterium. _____ digitatum did not cause any symptoms, however, at wounded sites where the bacterium was present. When the sites inoculated with both organisms were sampled for isolates of fungi and bacteria, then washed with 70% ethanol and water and reinoculated with P_;_ digitatum, no bacterial cells were recovered and 77.5% of the sites developed green mould decay another 4 days later. Discussion The bacterial isolate ID 2859 of P___ glathei gave a moderate biocontrol of P____ digitatum infection on postharvest oranges at normal conditions (25°C) . Its protective ability was significantly improved if the

inoculated fruits were incubated at 30°C for 24 hr. It has already been established that wounded oranges can heal faster at 30°C and become less susceptible to P. digitatum infection, however, a satisfactory reduction of Penicillium decay requires at least 72 hr healing at 30°C and >96% RH (36) . This was confirmed in this experiment. Exposing the fruit to such a high temperature for 72 hr can result in a reduced storage life due to high respiration rate (37) . Satisfactory reduction of Penicillium decay was achieved in this experiment when the wounded fruit were inoculated with the antagonist ID 2859 and healed at 30°C for 24 hr, even where the pathogen occupied the infection court before application of the antagonist. This technique of application overcomes a common weakness of biocontrol agents namely slower protection against plant pathogens compared with chemicals (38,39) .

The bacterium ID 2859 was readily recovered from the inoculated sites as was _P____ digitatum. This indicates that the bacterium did not eradicate _____ digitatum from the fruit but rather inhibited its infection of the fruit. Washing the wound sites inoculated with the bacterium before re-inoculating with the pathogen resulted in no later recovery of the bacterial cells and a significant loss of biocontrol at these sites. This suggests that the mode of action by this antagonist requires the presence of the bacterial cells and may involve a nutrient/space competition. Such a mode of action has also been suggested for biocontrol by yeast antagonists on apples (40) .

P. digitatum spores germinate well in aqueous extracts of orange rind or juice, or in sugars in the presence of phosphate buffer; but very poorly in water with or without the buffer or in sugars without the buffer. It has been concluded, therefore, that P. digitatum spores do not germinate in nature until they reach unhealed injured areas on the fruit. In order to infect the fruit, exopolygalacturonase (exo-PG) needs to

be induced from P____ digitatum spores by the substrate pectic acid which is formed through the activity of endogenous pectinmethylesterase (PME) on pectin in the damaged fruit cells. One, or more than one, of the following modes of action, therefore, may be involved in inhibition of P_j _. digitatum infection by the bacterium ID 2859; 1) competing for nutrients or for substrates like pectic acid, or for oxygen with __χ digitatum spores; 2) physically separating P_j_ digitatum spores from the injured tissue on fruit; 3) deactivating enzymes with exo-PG or PME activity.

It would seem the greater effectiveness of the bacterial antagonist when inoculated first into wounded fruit was due to it sufficiently establishing itself on the infection court. In order to be effective in biocontrol at 25°C, the non-antibiotic producing antagonist ID 2859 requires a much longer time than the antibiotic producing antagonist ID 2131 (at least 10 hr) to establish at the infection court before inoculation of the pathogen. This observation suggests that ID 2131 established biocontrol more quickly due to antibiotic production than does ID 2859 which does not produce an antibiotic.

The biocontrol by ID 2859 could be significantly improved if the inoculated were held at 30°C for a period. At 30°C for 24 hr, not only do the fungal spores not germinate but the bacterial cells have become much more numerous, indicating that they had multiplied quickly and established themselves during this time. As a consequence of establishment, the bacteria may be able to compete for nutrients with the pathogens and/or separate the pathogen spores from the injured tissues, from which the fungal spores presumably must obtain necessary nutrients for their germination.

Table 4-1. Effect of the strains ID 2859 of Pseudomonas glathei and ID 2131 of Pseudomonas cepacia on control of Penicillium digitatum infection on wounded oranges at 25°C .

¹ Valencia oranges were wounded and inoculated with P_ ₅ _ digitatum (3.1 x 10 ⁵ spore/ml) at the time intervals shown above relative to treatments with the bacteria (ID 2859 at 1.9 x 10 ⁹ cfu/ml and ID 2131 at 1.2 x 10 ⁹ cfu/ml) or water. The fruit were then kept in plastic bags at 25°C for 4 days after _______ digitatum inoculation.

² Figures in the table followed by the same letter are not significantly different by Waller-Duncan's

Bayesian K ratio LSD at k=100.

Table 4-2. Effect of the strains ID 2859 of Pseudomonas glathei and ID 2131 of Pseudomonas cepacia on control of Penicillium digitatum infection on wounded Valencia oranges which had been held at 30°C for a period before being transferred to 25°C.

Percentage infection at 25°C after fruit had been inoculated with the fungus at the times shown relative to application of the treatments and had been held at 30°C for the period shown in the second row below ¹ "

Treatment

4 hr earlier 2 hr later

24 hr 48 hr 72 hr 24 hr 48 hr 72 hr

Water 75.0 a ² 55.5 b 55 ά ID 2859 10.0 d 0.0 e 0.0 e ID 2131 28.3 c 122 d 5 J d

¹ Valencia oranges were inoculated with P. digitatum (3.1 x 10 ⁵ spore/ml) 4 hr before or 2

hr after treatment with the bacteria ID 2859 1.9 x 10 ⁹ cfu/ml, ID 2131 1.2 x 10 ⁹ cfu ml or

water. After incubation at 30°C for 24 hr, 48 hr or 72 hr, fruit were then kept at 25°C for

5 days and the infections by P. digitatum were recorded.

² Figures in the table followed by the same letter are not significantly different by Waller-

Duncan's Bayesian K ratio LSD at k=100.

Ν.T. not tested.

Table 4-3. Occurrence of fungal infection and recovery of fungal colonies and viable bacterial cells at sites inoculated with the strain ID 2859 of Pseudomonas glathei and with Penicillium digitatum, and in control

5 sites with P. digitatum alone ¹ .

Number of sites giving fungal and bacterial colonies and showing green mould decay ²

Treatment

5 days 9 days

ID 2859 P. df green mould ID 2859 P. d. green mould

¹ Valencia oranges were washed and wound-inoculated with the bacterium ID 2859 (1.6 x 10 ,9

cfu ml) and challenged with P. digitatum at 2.3 x 10 ⁵ spores/ml 2 hr later. After being held at 30°C for 24 hr the firuit were moved to 25°C for 8 days. The recovery of fungal and bacterial isolate and the occurrence of green mould decay were recorded at the 5th and 9th day after P. digitatum inoculation.

² Mean values are from a total of 40 examined sites. Each value is followed by ± sd indicating the standard deviation calculated from three replications of each treatment.

³ P. d. = P. digitatum.

⁴ No further examination was done since the whole fruits had become covered with green mould at 9 days.

Industrial Application The present invention is of use in agriculture such as in the fruit industry for protection of fruit from fungal decay.

References

1) Anon, 1990. Growth regulators: 1990 progress report, Citrograph. October: 278-279. 2) Bradford, .M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein-dye binding. Anal. Biochem. 12:248-2 ^{^} 54.

3) Eckert, J. . , 1983. Control of postharvest diseases with antimicrobial agents pp 265- 285. In: Postharvest physiology and crop preservation. eds Lieberman, . , Plenum press, NY. 572 pp.

4) Fravel, D.R. 1988. Role of antibiosis in the biocontrol of plant diseases. Ann. Rev. Phytopathol. 26:75-91

5) Homma, Y. , Sato, Z., Hirayama, F., Konno, K. , Shirahama, H. and Suzui, T., 1988. Production of antibiotics by Pseudomonas cepacia as an agent for biological control of soilborne plant pathogens. Soil Biol . Biochem. 21 (5) :723-728.

6) Huang, Y. , 1987. The fresh-keeping and preservation effects of agent AB on the storage of citrus fruit. Sichuan J. Agric. Sci. 3:30-35. 7) Huang, Y. , Deverall, B. J. , and Morris, S.C., 1991.

Promotion of Penicillium digitatum infection of orange fruit by a strain of Pseudomonas cepacia.

Phytopathology 81(6) 615-618.

8) Ismail, M. A., and Brown, G. E., 1979. Postharvest wound healing in citrus fruit: Induction of phenylalanine ammonia-lyase in injured 'Valencia' orange flavedo. J. Amer. Soc. Hort . Sci. 104 (1) : 126-129.

9) Janisiewicz, . J. 1987. Biological control of diseases of fruits, pp 153-165. In: Biocontrol of plant diseases. Vol. II. R. J. Mukerji and K. L. Gary, eds. CRC press, Boca Raton, FL, 198 pp. 10) Janisiewicz, . J. , and Roitman, J. , 1988. Biological control of blue mold and gray mold on apple and pear with Pseudomonas cepacia Phtopathology. 78:1691-1700.

11) Nyfeler, R. , Egli, T.H. , and Nevill, D. 1990. Phenylpyrroles, a new class of agricultural fungicides related to the natural antibiotic pyrronitrin. In Ciba-Geigy's Contribution. ^" Ciba- Geigy, Switzerland.

12) Pusey, P.L., Wilson, C. L., Hotchkiss, M. W. , and Franklin, J. D., 1986. Compatibility of

Bacillus subtilis for postharvest control of peach brown rot with commercial fruit waxes, dicloran, and cold-storage conditions. Plant Dis. 70: 587-590. 13) Rahe, J.E., Kuc, J. , and Chuang, CM. 1970. Cinnamic acid production as a method of assay for phenylalanine ammonialyase in acetone powders of Phaseolus vulgari s . Photochemistry. 9: 1009-1015. 14) Rishbeth, J. 1988. Biological control of air-borne pathogens. Phil. Trans. R. Soc. Lond. B 318:265- 281.

15) Roitman, J. N. , Mahoney, N, E., Janisiewicz, W. J. and Benson, M. 1990. A new chlorinated phenylpyrrole antibiotic produced by the antifungal bacterium Pseudomonas cepacia. J. Agric. Food Chem. 38:538-541.

16) Stell, R. G. D., and Torrie, J. H. , 1980. Multiple comparisons. Pages 172-191. In: Principles and procedures of statistics. McGraw-Hill Book Co., NY.

17) Singh, V., and Deverall . , B. J. , 1984. Bacillus subtilis as a control against fungal pathogens of

citrus fruit. Trans. Br. Mycol. Soc, 83:487-490.

18) Utkhede, R. S., and Sholberg, P. L., 1986. In vitro inhibition of plant pathogens by Bacillus subtilis and Enterobacter aerogenes and jLn vivo control of two postharvest cherry diseases. Can. J. Microbiol., 32:963-967.

19) Walker, J. C. 1952. Diseases of vegetables crops. McGraw-Hill Book Co., NY. 529 pp.

20) Wilson, C. L., and Pusey, P. L. , 1985. Potential for biological control of postharvest plant diseases. Plant Dis. , 69:375-378.

21) Wilson, C. L., and Chalutz. E., 1989. Postharvest biological control of Penicillium rots of citrus with antagonistic yeasts and bacteria. Scientia Hortic, 40:105-112.

22) Glasby, J. S. (ed) 1982. Encyclopaedia of the terpenoids . John Wiely and Sons, New York 1025 pp.

23) Godtfresden W.O., Daehne W., Vangedal S., Marquet A., Arigoni D, and Melera A., 1965. The stereochemistry of fusidic aacid. Tetrahedron . 21:3505-3530.

24) Hall L.D. and Sanders J.K.M. 1981. Analysis of the proton nuclear magnetic resonance spectrum of llβ- hydroxyprogesterone by one and two dimensional methods. Some implications for steroid and terpenoid chemistry. Journal of Organic Chemistry. 46:1132-1138.

25) Jones D.A., Ryder M.H. , Clare B.G., Farrand S.K. and Kerr A. 1988 Construction of a Tra deletion mutant of pAgk84 to safeguard the biological control of crown gall. Molecular and General Genetics . 212:207-214.

26) McCrindle R. and Nakamura E. 1974. Constituents of Solidago Species, Part VI. The constitution of diterpenoids from a chemically distinct variety of Solidago serotina . Canadian Journal of Chemistry. 52:2029-2036.

27) Ruzicka L., 1963. Pwespektiven der biogenese und der chemie der terpene. Pure and Applied Chemistry. 6:493-523.

28) Ryder M.H. , and Jones D.A. 1990. Biological control of crown gall. In: H. David and R.J. Cook

(eds) , Biological Control of Soil -borne Plant Pathogens . CAB International 45-63 pp.

29) Seubert W. and Fass E., 1964. Untersuchungen uber den bakteriellen abbau von isoprenoiden. Biochemische Zei tschrift . 341:35

30) Stowe B.B., Stodola F.H., Hayashi T. and Brian P.W., 1961. The early history of gibberellin research, pp 465-471. In Plant Growth Regula tor. Iowa State University Press. 850 pp. 31) Wolff G., Polonsky J. , Scartazzini R., Settimj G and Arigoni D. 1965. Quoted in G. Harris, Dictionary of Organic Compounds . 4th ed. , 2nd suppl . p.94. Oxford University Press. N.Y.

32) Wood R.K.S. and Tveit M. 1955. Control of plant diseases by use of antagonistic microorganisms.

Botany Review.21 :441-492.

33) Arigoni D. 1968. Some studies in the biosynthesis of terpenes and related compounds. Pure and Applied Chemistry. 17:331-347. 34) Huang Y. , Deverall B.J. and Morris S.C. 1993. Effect of Pseudomonas cepacia on postharvest biocontrol of Penicillium digi tatum infection and wound responses of citrus fruit. Australasian Plant Pa thology: 22: (in press) . 35) Steel, R.G.D., and Torrie J.H. (eds) 1980. Multiple comparisons. pages 172-191. In: Principl es and procedures of s ta tisti cs . McGraw- Hill Book Co. , NY. 36) Barmore CR. and Brown G.E. 1979. Role of pectolytic enzymes and galacturonic acid in citrus fruit decay caused by Peni cillium digi ta tum . Phytopa thology. 69:675-678.

37) Vines H.M. , Grierson W. and Edwards G.J. 1968. Respiration, internal atmosphere and ethylene evolution of citrus fruit. Proceeding of American Society for Horticul tural Science . 92:227-234. 38) Cook R.J. and Baker J.F. (eds) . 1983. The Nature and Practice of Biological Control of Plant Pathogens . American Phytopathological Society, St. Paul, Minnesota, 593 pp.

39) Bay-Petersen J. (ed) . 1991. The Biological Control of Plant Diseases . Based on proceeding of the international seminar 'Biological control of plant diseases and virus vectors' Japan, Sep. ^" 17-21, 1990. 215 pp.

40) Wilson CL. and Chalutz E., 1989. Postharvest biological control of Penicillium rots of citrus with antagonistic yeasts and bacteria. Scientia Horticul tura . 40:105-112.