Journal of General Microbiology (1984), 130, 2447-2455. Printed in Great Britain 2447

Isolation of Mutants of Xanthomonas campestris pv. campestris Showing Altered Pathogenicity

By MICHAEL J . DANIELS* CHRISTINE E. BARBER, PETER C. TURNER, WENDY G. CLEARY A N D

MARIA K . SAWCZYC John Innes Institute, Colney Lane, Norwich NR4 7UH. UK

(Received 16 March 1984; revised 17 May 1984)

A procedure has been developed for rapidly inoculating large numbers of turnip seedlings with Xanthomonas campestris pv. campestris, the causal agent of black rot of crucifers. Following mutagenesis bacterial mutants were isolated which showed altered pathogenicity. The mutants gave a range of responses on plants and some were impaired in their ability to grow in seedlings.

INTRODUCTION

The mechanisms by which bacterial pathogens induce disease in plants are poorly understood, despite extensive physiological work on suspected pathogenicity determinants such as toxins, enzymes, growth factors and polysaccharides. While these factors may induce certain visible symptoms, the total disease phenomenon has many properties for which it is quite impossible at present even to guess at the nature of the mechanisms involved (Preece, 1982).

Experience with other complex biological stystems suggests that the isolation and analysis of bacterial mutants altered in their behaviour in plant hosts would shed some light on the process and help to define the general nature of the interactions. Hitherto, genetic studies of bacterial pathogenicity have been largely confined to the properties of mutants defective in the production of specific factors which were suspected to be involved in pathogenicity, such as toxins of Pseudomonas phaseolicola (Patil et al., 1974), pectolytic enzymes of Erwinia chrysanthemi (Chatterjee & Starr, 1977), extracellular polysaccharides and other factors of Erwinia amylovora (Bennett, 1980; Bennett & Billing, 1978), polysaccharides of Pseudomonas solanacearum (Staskawicz et al., 1983), and auxin produced by Pseudomonas savastanoi (Smidt & Kosuge, 1978). These experiments have yielded information about the role of the factors in disease. However, it is likely that other classes of pathogen products are also important in causing disease. In particular the determinants of host specificity are unknown. Suitable screening tests can often be devised to permit the isolation of bacterial mutants lacking specific products, but clearly this cannot be achieved if the general nature of the factors of interest is unknown. However, it is reasonable to suppose that mutations in many genes determining pathogenicity factors (sensu latu) will cause visible alterations in symptoms of infected plants. We concluded that to begin to build up a comprehensive picture of the disease process it would be helpful to isolate a collection of mutants altered in their behaviour in plants, without prejudice to the nature of the biochemical lesion, and then to use methods of molecular genetics to characterize the genes (defined by the mutations), and their products.

Many Gram-negative bacterial plant pathogens are potentially suitable for such experiments, but our choice of Xanthomonas campestris pv. campestris was based on our ability to inoculate individual plants on a large enough scale to permit screening for pathogenicity mutants. We considered it necessary to inoculate individual plants with each bacterial colony to avoid the risk of interference between separate inocula (Sequeira, 1984), and considerations of space therefore dictated the use of small seedlings. The number of colonies which would have to be tested to

0022-1287/84/OOO1-1850 $02.00 0 1984 SGM

2448 M . J . DANIELS AND OTHERS

isolate mutants further dictated the use of a very simple, rapid inoculation technique; procedures such as spraying plants with bacterial suspensions or infiltration with a syringe would be impracticable for use with thousands of isolates. A number of Pseudornonas spp. ( P . coronufaciens, P . lachrymans, P . mors-prunomm, P . phaseolicola, P . pisi, P . syringae, P . tabaci, and P. tomato) were tested in preliminary experiments but gave unreliable symptom production when inoculated by pricking with a needle. However, X . c. campestris gave satisfactory results and our expectation that this bacterium would also be amenable to genetic manipulation has been proved correct (Turner et al., 1984).

Xanthomoms c. campestris is the causal agent of black rot of crucifers (Smith, 1897), which is one of the most serious world-wide diseases of cruciferous crops (Williams, 1980). In addition to being an important pathogen, X . c. campestris is used industrially for the manufacture of xanthan gum (Jeanes et al., 1961). The only possible pathogenicity factors so far suggested for X . carnpestris pathovars are pectolytic enzymes (Starr & Nasuno, 1967), xanthan (Sutton & Williams, 1970) and low molecular weight acidic phytotoxic substances (Noda et al., 1980; Perreaux el al., 1982), but no evidence has been presented for a role of these latter substances in disease.

In this paper we describe the isolation of X . c. carnpestris mutants which are altered in their ability to cause disease in plants, by direct screening of large numbers of colonies from mutagenized cultures.

METHODS

Bacteria and culture media. All X . c. campestris strains were derivatives of NCPPB 1145. A spontaneous mutant resistant to rifampicin (100 pg ml-l) was used as the standard wild-type pathogen designated 8004. Complete medium (NYGB) contained the following ingredients (g 1- l ) : peptone (Oxoid), 5 ; yeast extract (Difco), 3; glycerol, 20; and solid medium (NYGA) contained in addition 10 g agar 1- (Lab My Salford, UK). Minimal medium was Minimal A (Miller, 1972) supplemented with glucose ( 5 g 1- l). Supplements of Casamino acids (Difco) at 1.5 g 1- or single amino acids at 75 mg 1- were added as appropriate. Cultures were grown at 25 "C; broth cultures were usually shaken. Strains were stored for short periods on agar at 4 "C or for longer periods as frozen suspensions in NYGB containing 20% (v/v) glycerol at - 20 "C. Turbidimetric measurements of bacterial growth were made with a Unicam SP1700 spectrophotometer at 600 nm, using 1 cm light path cells.

Other X . campestris pathovars used were: X . c. graminis (NCPPB 2700, 2701, 3040); X . c. holcicolu (NCPPB 1060); X . c. malvacearum (NCPPB 633); A'. c. oryzae (XO61, X071, X079, from J. A. Callow, University of Birmingham, UK); X. c.pisi(NCPPB 762); X . c . tramlucens (NCPPB 1943); X. c. vesicatoriu (NCPPB 422); X . c. cirians (NCPPB 1839) and X . c. zinniae (NCPPB 799). The Eschcrichia coli strain was ED 8767 (Murray et al., 1977). NYGB gave unsatisfactory growth of X . c. oryzae and for this pathovar Goto's medium was used (Parry et ul., 1982).

Mutagenesis. Portions (1 mi) of overnight cultures of X . c. campesrris 8004 were centrifuged for 1.5 min at 20 "C in a Micro Centaur centrifuge (MSE), and the cells were washed twice with 0.1 M-sodiUm citrate/citric acid buffer (pH 5 - 5 ) and resuspended in 1 ml of the same buffer. Then 10 pl of a freshly prepared solution of N-methyl-" nitro-N-nitrosoguanidine (NTG) (1 mgml-l in citrate buffer) was added and the tubes were left at room temperature for periods of 10 to 30 min. To remove the mutagen, cells were pelleted by centrifugation for 1 min, the supernatant fluid was discarded and the cells were resuspended in, and washed twice with 0.1 M-sodium phosphate buffer (pH 7.0). Each batch of mutagenized cells (corresponding to 1 ml original culture) was resuspended in 5 ml NYGB and incubated for 24 h at 25 "C to allow mutations to segregate. Only one mutant of any type was retained from each batch to avoid isolating siblings. Suitable dilutions of the overnight cultures were plated on NYGA containing rifampicin (50 pg ml- l) and single colonies were transferred to 'master' plates with sterile toothpicks to give an array of 50 colonies per plate, which were replica-plated with velvet to plates of (1) minimal medium, (2) minimal medium supplemented with Casamino acids and (3) NYGA containing 1 % (w/v) glucose. The use of these media permitted detection of (1) auxotrophs in general, (2) amino acid-requiring auxotrophs and (3) mutants altered in polysaccharide secretion. The strain X. c. campestris 8004 produces relatively little polysaccharide on NYGA, and supplements of glucose are necessary to study this character.

Pathogenicity test. Seeds of hybrid turnip (Brussica campestris 'Just Right' from W. Atlee Burpee Co., Warminster, PA 18974, USA) were surface-sterilized by immersion in 10% (v/v) commercial NaOCl bleach (Payton Cleaning Supplies, Norwich, UK) for 30 min, washed several times with sterile distilled water and plated on water agar at 25 "C. After 20 h essentially all the seeds had germinated and they were transferred individually to wells of 'Replidishes' (catalogue no. 103, Sterilin) containing 2.5 ml of one-quarter-strength plant tissue culture medium (Murashige & Skoog, 1962) containing 1 % (wlv) agar and 5 pg kanamycin ml- (to inhibit any residual

Pathogenicity mutants of Xanthomonas campestris 2449

Fig. 1. Procedure for inoculating turnip seedlings from X . c. campestris colonies on a master plate. The seedlings are supported by the four-pronged device whilst being inoculated with the needle.

seedborne bacteria). The plates were incubated in a humid atmosphere, either individually in clear plastic sandwich boxes or in groups of six in standard plant propagators, in a growth room at 25 "C with illumination for lOh per day from sodium lamps and fluorescent tubes [photon flux density (400 to 700nm), 150pmol quanta m-2 s-']. After 3 d the seedlings were ready for use (3 to 5 cm high). Inoculation was effected by touching a sterile sharpened nichrome wire on a bacterial colony and then gently pricking through the stem of a seedling. To facilitate this manipulation the seedling was supported with a horizontal hand-held device consisting of four pieces of nichrome wire projecting 2.5 cm from an aluminium handle, the wires being inserted in holes drilled at the comers of a square with sides of 5 mm (Fig. 1). The time required to inoculate one dish (25 seedlings) was about 7 min. When more precise control of the inoculum was required the inoculating needle was dipped to a standard depth (3 mm) in a bacterial suspension of known density before being pricked through the stem of the seedling. After inoculation the seedlings were again incubated at 20 "C and observed daily.

Inoculation of more mature plants was achieved by infiltrating bacterial suspensions in water into the lamina of a fully expanded leaf of 6- to 8-weeksld greenhouse-grown turnip plants, using a 1 ml disposable syringe with a 26 G needle.

Bacterial content of infected plant tissue. Groups of five infected seedlings were placed in sterile hand-held glass homogenisers and vigorously disrupted for 5 min. The homogenates were made up to 10 ml with sterile water and a range of dilutions plated on NYGA containing rifampicin (50 yg ml- l). Reconstruction experiments showed that turnip seedling homogenates did not contain substances which significantly reduced the number of viable bacteria, and control experiments showed that the viability of X. c. campestris suspended in water did not decrease for at least 8 h at room temperature.

RESULTS

Pathogenicity test

Turnip seedlings inoculated with X. c. campestris showed a characteristic progression of symptoms over a period of 4 to 5 d. After 16 to 24 h, constriction of the stem was apparent around the point of inoculation. The affected area gradually spread up and down, becoming translucent and then turning brown and glistening. The cotyledons turned yellow and after 4 d the seedlings were completely collapsed and rotted. The progression of symptoms was identical whether the plants were incubated at 18,20 or 32 "C, although the rate of development increased as the temperature was raised, and incubation of seedlings in darkness after inoculation did not affect the results. The highest temperature at which our strain of X. c. campestris forms colonies on NYGA is about 34°C.

Several commercial varieties of other brassicas were tested (Brussels sprout, cabbage, calabrese, cauliflower, Chinese cabbage, mustard, radish and rape, including F hybrids where available), but none gave such reproducible, rapid germination, growth and disease expression as the turnip hybrid 'Just Right'.

The disease response was only obtained with living X. c. campestris cells. Inoculation with X. c. campestris killed by exposure to chloroform vapour or with E. coli gave no symptoms.

2450 M. J . DANIELS AND OTHERS

Table 1. Progression of disease in turnip seedlings inoculated with different concentrations of X. c. campestris

25 seedlings were inoculated with each bacterial suspension or with sterile water as a control and observed daily. The entries indicate the number of seedlings showing full symptoms or (in parentheses) the number with discoloration around the site of inoculation. After 9 d there was little further change in their appearance. Seedlings inoculated with water showed no signs of disease at any time after inoculation.

No. of seedlings diseased following inoculation with suspensions of given concn

r 7

Time after (c.f.u. ml-I) inoculation A

( 4 109 1 08 107 1 06 1 0 5

1 2 (4) 1 (23) 0 (0) 0 (0) 0 (0) 2 25 4 (21) 0 (0) 0 (0) 0 (0) 3 25 22 (3) 4 (21) 0 (0) 0 (3) 4 25 22 (3) 5 (20) 2 (8) 1 (2) 7 25 25 6 (19) 4 (7) 1 (2) 9 25 25 7 (18) 5 (10) 1 (4)

Moreover, other X . c . campestris pathovars (specific for other plant taxa) gave, with two exceptions, either no symptoms or a faint brown ring around the inoculation point (about 1 mm diameter) which remained localized and had no effect on the growth of the seedlings. The pathovars tested were X . c. graminis, holcicola, malvacearum, oryzae, pisi, translucens, vesicatoria, r;itians and zinniae. Of these xanthomonads, only pv. uesicatoria and pv. zinniae gave symptoms resembling the black rot incited by pv. campestris. However, pv. uesicatoria is known to be capable of infecting brassicas (Shackleton, 1966), and while little is known about the relationship between pv. zinniae and pv. campestris, we have observed that both cause black rot- like symptoms in seedlings of Zinnia elegans.

When seedling homogenates were diluted and plated immediately after inoculation it was found that the standard inoculation procedure introduced about 5 x lo6 c.f.u. into each seedling. However, with experience it was possible to exert some control over this factor by varying the 'lightness of touch' of the needle on the bacterial colonies. More precise control was achieved by dipping needles to a standard depth in a bacterial suspension of known titre. Subsequent assay of seedlings inoculated in this fashion showed that the bacterial content of each seedling was equivalent to 0.1 p1 of the suspension used for inoculation, i.e. a seedling inoculated with a suspension of X . c. campestris at lo8 c.f.u. ml- would acquire lo4 c.f.u. Table 1 shows the progression of disease among groups of seedlings inoculated with suspensions of different bacterial concentrations. One may conclude that there is a threshold inoculum concentration required for successful infection which appears to be about lo2 to lo3 c.f.u. per seedling (i.e. inoculation with 0-1 p1 of suspensions at lo6 to lo7 c.f.u. ml-l).

The growth of X . c. campestris in turnips was studied by inoculating seedlings with a suspension of bacteria and at daily intervals homogenizing five seedlings individually for determination of viable count. Fig. 2 shows the results obtained in a representative experiment. The bacteria grew and reached a final level of lo8 to lo9 c.f.u. per seedling and during the period when the growth rate of the population was greatest the doubling time was about 4 h. For comparison, X . c. campestris shaken cultures have a doubling time of about 3 h in NYGB at 25 "C. The final titre reached in infected seedlings was independent of the inoculum level (provided that this was greater than the threshold value : data not shown). It is difficult to assess accurately the efficiency of the homogenization procedure in releasing bacteria quantitatively from plant tissue. However, an experiment was performed in which groups of seedlings infected with the same bacterial suspension were compared following disruption by the standard method, or by vigorous grinding with sand in a mortar. The recovery of bacteria was the same in both cases, within experimental error. In view of the data for reproducibility shown in Fig. 2 we consider that the procedure gives sufficiently accurate estimates of the bacterial content of plant tissues for the screening of mutants for inability to grow in plants.

Pathogenicity mutants of Xanthornonas carnpestris 245 1

t 1 1 I I

1 2 3 4 3 1

Time after inoculation (d)

Fig. 2. Growth of X . c. campestris in turnip seedlings. Twenty-five seedlings were inoculated from a suspension of X. c. campestris (1 x lo9 c.f.u. ml- l). Groups of five seedlings were taken immediately and at daily intervals thereafter and homogenized individually for determination of viable bacterial count. Each point represents one seedling.

The reliability of the pathogenicity test was assessed by inoculating groups of 100 seedlings each with a single colony from a culture of X . c. carnpestris. About 95 % of the seedlings showed the characteristic disease symptoms, while the remainder suffered only local discoloration. However, when the colonies giving the apparent negative results were retested, all were found to be pathogenic, i.e. the negative results were caused by some artefact of the first inoculation or by variation in the host seed. In repeated tests of some 1500 colonies of unmutagenized X . c. carnpestris no spontaneous non-pathogenic variants were found.

Isolation and properties of mutants

The efficiency of the mutagenesis procedure was assessed from the proportion of auxotrophic mutants recovered. Exposure of X . c. campestris to NTG for 20 min yielded about 3% auxotrophs among the survivors and this level of mutagenesis was chosen for subsequent experiments. A limited number of amino acid-requiring mutants was studied and examples were found requiring the following supplements : arginine, cysteine or methionine, histidine, leucine, methionine, proline, serine or threonine, and threonine. In addition, mutants were found which lacked the yellow pigments characteristic of Xanthomonas (xanthomonadins; Starr et al., 1977), were temperature-sensitive (able to grow at 20 but not at 32 “C), were sensitive to UV light, or which showed altered, more slimy, colony morphology, possibly because of mutational alterations in polysaccharide synthesis (Whitfield et al., 1981). No mutants were found which failed to produce polysaccharide.

Colonies of mutagenized bacteria were tested initially on single turnip seedlings. Any which showed a response which was considered abnormal were subcultured and retested on groups of not less than five seedlings per colony. The use of groups of seedlings eliminated variation due to heterogeneity of seeds in the initial test. At this stage the majority of colonies were found to give wild-type symptoms but those which were consistently abnormal were tentatively regarded as ‘pathogenicity mutants’ and were studied further. All were found to be prototrophic (able to grow on minimal medium), motile (by microscopic examination of exponentially growing broth cultures) and able to grow at a rate comparable to the wild-type. The latter property was tested roughly by the rate of increase of colony diameter and more accurately by turbidimetric measurements of growth in shaken broth cultures. Such pathogenicity mutants arose at a frequency of approximately 1/ 10 the frequency of auxotrophs.

2452 M. J . DANIELS AND OTHERS

Fig. 3. Typical responses evoked by wild-type A’. c. campestris and by four mutants following infiltration into fully expanded turnip leaves. The photographs were taken 4 d after inoculation.

It was possible to recognize four types of mutant from seedling inoculations. One gave no visible symptoms, seedlings resembling those ‘inoculated’ with a sterile needle. The second gave a localized response consisting of a dark opaque ring around the inoculation point, about 2 mm diameter, which failed to enlarge and had no effect on seedling growth. The third type gave partial symptoms consisting of a dark translucent zone spreading from the inoculation point for about 5mm, but showing no further spread and no general rotting. Seedlings frequently succumbed to these infections, probably because of impaired conduction of water and nutrients through the infected tissue. The fourth type of mutant showed a reduced probability of infection, i. e. inoculation of a group of seedlings gave 10 to 20% showing symptoms similar to wild-type, whereas the remainder showed either no symptoms or a localized response. As noted above, comparable experiments with the wild-type strain gave 90 to 95 % successful infection. Representative mutants gave a range of responses when infiltrated into mature turnip leaves. Suspensions in water or broth at about 5 x lo7 c.f.u. ml- were infiltrated, about 20 p1 of fluid

Pathogenicity rn u tan ts of Xan thomonas campes t r is

Table 2. Behaviour of X . c. campestris mutants in turnip seedlings

2453

Dose responsef No. of seedlings diseased following inoculation

with suspensions of given concn (c.f.u. ml-l)

Strain

8004 8220 8255 8284 8224 8225 8237 8288 8221 8226 8253

Reaction type*

Wild-type A A A B B B B C C C

I 3

5 x 109 5 x 108 5 x 1 0 7 5 x 106 5 x 105

5 4 4 5 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 4 3 0 0 0 3 4 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 5 1 0 0 0 2 2 0 0 0

Growth$ Growth factor of

bacteria in seedlings inoculated from suspensions of given concn (c.f.u. ml-1) & I x 109 1 x 107

46 283 28 35 4

17 62 169 6

28 38

- -

-

- 11 10 50 116

- -

* Reaction type following inoculation of seedlings from bacterial colonies: A, no visible symptoms; B, local

t Dose response: five seedlings were inoculated from suspensions of given density; entries in the table indicate

1 Growth: seedlings were inoculated from suspensions of given density; entries indicate the factor by which the

brown lesion; C, limited spread of lesion (see text).

the number of seedlings showing full disease response.

number of viable bacteria per group of seedlings increased in two days following inoculation (-, not done).

being needed to water-soak I cm2 of leaf area. The mutants and the wild-type strain were each infiltrated into four leaves ranging from just fully expanded to near-senescent on each of two plants. Plants were kept in sealed plastic bags to maintain high humidity. Some mutants (e.g. 8225 and 8253) gave no visible symptoms after 1 week. Mutant 8220 produced some small holes in the infiltrated area but no further visible changes, whereas mutant 8284 gave similar holes but which were accompanied by slowly spreading distal chlorosis. On the other hand, a number of mutants gave hypersensitive collapse of the infiltrated area. With mutant 8255 the collapsed tissue desiccated to give a papery lesion, whereas 8237 showed collapse but with less marked desiccation. In mutant 8221 collapse was accompanied by distally spreading chlorosis, and with 8224 and 8226 the collapsed tissue turned black (with distal chlorosis). Mutant 8288 showed symptoms resembling the wild-type in the infiltrated area, but the damage failed to spread. For comparison, wild-type symptoms included chlorosis, blackening and rotting of infected tissue, and spreading distally and proximally so that the whole plant eventually became destroyed. Fig. 3 shows examples of some responses. The responses did not depend on the age of the leaf inoculated.

The behaviour of mutants in seedlings was examined further in experiments summarized in Table 2. Groups of five seedlings were inoculated with a range of dilutions of suspensions. While some mutants (e.g. 8220,8237,8255,8284 and 8288) gave no symptoms, the remainder were able to give some infection if inoculated at high concentrations. In all cases the threshold level was considerably higher than the value for the wild-type.

The ability of the same mutants to grow in turnip seedlings was measured by inoculating ten seedlings with a suspension of each strain. Five seedlings were pooled and homogenized immediately and the second group was similarly processed 2 d later. The results are expressed in Table 2 as the factor by which the number of viable bacteria per seedling group increased in 2 d. Most mutants showed reduced growth compared with the wild-type, but four which appeared to grow well (8220,8224,8237 and 8253) were retested using a smaller inoculum and their ability to undergo considerable multiplication was confirmed. All the experiments comparing the behaviour of mutants in seedlings were performed at least twice, with essentially identical results.

2454 M . J . DANIELS A N D OTHERS

DISCUSSION

The pathogenicity test which we have used for X . c. cumpestris makes it possible for one person to screen many hundreds of mutagenized colonies in a day, so that mutants of altered pathogenicity can be isolated without the expenditure of unreasonable time and effort, and without recourse to the costly facilities required for maintaining large stocks of mature plants. It must be stressed that the test is regarded as a ‘screen’ for pathogenicity, not as a definitive model of the whole disease process. In the field X. c. cumpestris is dispersed by factors such as rain and wind and enters plants through natural openings, principally hydathodes, as well as wounds (Smith 1897; Cook et ul., 1952). Some varieties of cabbage owe their resistance to black rot to a property which renders them less susceptible to invasion through the hydathodes (Staub & Williams, 1972). This aspect of the infection process is by-passed by our needle inoculation technique, so it follows that any bacterial genes specifically involved in the early stages of natural infections would not be detected in our experiments.

It is also important to understand that our operational definition of ‘pathogenicity genes’ is somewhat unconventional. We have arbitrarily chosen to include all mutants which show unimpaired growth in our standard laboratory media, but which produce a response on plants that differs from that incited by the wild-type. Thus all auxotrophic mutants identified by prior screening on minimal medium were not initially tested for pathogenicity because it was considered that any failure to incite disease might be ascribable to the trivial, but difficult to prove, cause that plant juices contain insufficient of the required supplement to support adequate growth of the pathogen (Coplin et ul., 1974). In retrospect this caution may not have been necessary, for most auxotrophs subsequently tested retained full pathogenicity, an exception being a proline-requiring mutant. All pigmentdeficient and ‘slimy’ mutants were pathogenic. It has sometimes been suggested that our approach may result in the isolation of mutants in pathways of intermediary metabolism which may be dispensable for growth explunta in our simple laboratory media, but are necessary for growth in pluntu. Our view is that even if such genes would not normally be considered to encode pathogenicity factors they must nonetheless be taken into account if one wishes to understand fully the disease process. Moreover, since it is not at present possible to predict the nature of these putative essential pathways, there is no way of excluding mutants of this type from the collection, in the way that auxotrophs can be excluded.

The frequency of occurrence of pathogenicity mutants compared with auxotrophs suggests that a considerable number of genes may be involved, and this is reflected in the range of phenotypic behaviour shown by the mutants.

Some mutants are able to grow in seedlings without causing disease, showing that ability to grow, while undoubtedly necessary, is not a sufficient criterion for pathogenicity. A number of mutants provoke a hypersensitive response leading to localization of the infection, suggesting that the mutations have removed factors which in the wild-type suppress the response. There is considerable evidence for the existence of protective factors of this type, but little is known of their nature (Sequeira, 1976).

Some mutants are ‘leaky’ in the sense that a proportion of inoculated seedlings show normal symptoms, particularly when inoculated with large numbers of bacteria, It may be useful to consider that there is a finite probability that a single bacterial cell may initiate a successful infection (Ercolani, 1973), and that this probability factor is influenced by a number of gene products so that mutations may reduce the probability.

In common usage ‘pathogenicity’ is regarded as an intrinsic property of a bacterial species, in the sense that X . c. campestris is pathogenic to plants whereas E. coli is not: on the other hand ‘virulence’ is a quantitative property of individual strains of a pathogenic species. The leaky mutants may then be regarded as ‘virulence mutants’ whereas those giving no symptoms under any circumstances would be ‘pathogenicity mutants’.

The range of phenotypes shown by the mutants gives grounds for optimism that our approach may lead to the identification of a range of pathogenicity genes. Work is in progress to clone the genes defined by the mutations as the first step in analysing their functions. Allied to physiological studies of the mutants and of plants infected with them, this approach should gradually enable us to build up a picture of the disease process.

Pathogenicity mutants of Xanthomonas campestris

We are grateful to Giles Reed and Mark Dingle for help with certain experiments.

2455

REFERENCES

BENNETT, R. A. (1980). Evidence for two virulence determinants in the fireblight pathogen Erwinia amylovora. Journal of General Microbiology 116,351- 356.

BENNETT, R. A. & BILLING, E. (1978). Capsulation and virulence in Erwinia amylmora. Annals of Applied Biology W, 41-45.

CHAITERJEE, A. K. & STARR, M. P. (1977). Donor strains of the soft-rot bacterium Erwinia chrysanthemi and conjugational transfer of the pectolytic capacity. Journal of Bacteriology 132, 862-869.

COOK, A. A., WALKER, J. C. & LARSON, R. H. (1952). Studies on the disease cycle of black rot of crucifers. Phytopathology 42, 162-167.

COPLIN, D. L., SEQUEIRA, L. & HANSON, R. S . (1974). Pseudomonas solanacearum : virulence of biochemi- cal mutants. Canadian Journal of Microbiology u),

ERCOLANI, G. L. (1973). Two hypotheses on the aetiology of response of plants to phytopathogenic bacteria. Journal of General Microbiology 75, 83-95.

JEANES, A., PITISLEY, J. E. & SENTI, F. R. (1961). Polysaccharide B-1459 : a new hydrocolloid polyelec- trolyte produced from glucose by bacterial fermenta- tion. Journal of Applied Polymer Science 5, 519-526.

MILLER, J . H. (1972). Experiments in Molecular Genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press.

MURASHIGE, T. & SKOOG, F. (1 962). A revised medium for rapid growth and bioassays with tobacco tissue cultures. Physiologia plantarum 15, 473-497.

MURRAY, N. E., BRAMMAR, W. J. & MURRAY, K. (1977). Lambdoid phages that simplify the recovery of in vitro recombinants. Molecular and General Genetics 150, 53-61.

NODA, T., SATO, Z., KOBAYASHI, H., IWASAKI, S. & OKUDA, S. (1980). Isolation and structural elucida- tion of phytotoxic substances produced by Xantho- monas campestris pv. oryzae (Ishiyama) Dye. Annals of the Phytopathological Society of Japan 46,663466.

PARRY, R. W. H., CALLOW, J. A. & MORRIS, E. R. (1982). Improved method for obtaining viable counts of Xanthomonas oryzae. International /Rice Research Newsletter 7 , 14.

PATIL, S. S., HAYWARD, A. C. & EMMONS, R. (1974). An ultraviolet induced non-toxigenic mutant of Pseudomonas phaseolicola of altered pathogenicity. Phytopathology 64, 590-595.

PERREAUX, D., MARAITE, H. & MEYER, J. A. (1982). Identification of 3 (methylthio) propionic acid as a blight-inducing toxin produced by Xanthomonas campestris pv. manihotis in vitro. Physiological Plant Pathology 20, 3 1 3-3 19.

PREECE, T. F. (1982). The progression of bacterial disease within plants. In Bacteria and Plants, pp. 71- 83. Edited by M. Rhodes-Roberts & F. A. Skinner. London : Academic Press.

5 19-529.

SEQUEIRA, L. (1976). Induction and suppression of the hypersensitive reaction caused by phytopathogenic bacteria: specific and non-specific components. In Specificity in Plant Diseases, pp. 289-309. Edited by R. K. S. Wood & A. Graniti. New York: Plenum Press.

SEQUEIRA, L. (1984). Cross protection and induced resistance : their potential for plant disease control. Trends in Biotechnology 2, 25-29.

SHACKLETON, D. A. (1966). A study of the relationship between Xanthomonas vesicatoria (Doidge, 1920) Dowson, 1939, and X. campestris (Pammel, 1895) Dowson, 1939. New Zealand Journal of Science 9,

SMIDT, M. & KOSUGE, T. (1978). The role of indole-3- acetic acid accumulation by alpha methyl-trypto- phan-resistant mutants of Pseudomonas savastanoi in gall formation on oleanders. Physiological Plant Pathology 13, 203-2 14.

SMITH, E. F. (1897). Pseudomonas campestris (Pammel). The cause of brown rot in cruciferous plants. Zentralblatt fur Bakteriologie (Abteilung 11) 3, 284- 291.

STARR, M. P. & NASUNO, S. (1967). Pectolytic activity of phytopathogenic xanthomonads. Journal of General Microbiology 46, 425-433.

STARR, M. P., JENKINS, C. L., BUSSEY, L. B. & ANDREWES, A. G. (1977). Chemotaxonomic signifi- cance of the xanthomonadins, novel brominated aryl-polyene pigments produced by bacteria of the genus Xanthomonas. Archives of Microbiology 113, 1- 9.

STASKAWICZ, B. J., DAHLBECK, D., MILLER, J. & DAMM, D. (1983). Molecular analysis of virulence genes in Pseudomonas solanacearum. In Molecular Genetics of the Bacteria - Plant Interaction, pp. 345- 352. Edited by A. Piihler. Berlin & Heidelberg: Springer-Verlag.

STAUB, T. & WILLIAMS, P. H. (1972). Factors influenc- ing black rot lesion development in resistant and susceptible cabbage. Phytopathology 62, 722- 728.

SUTTON, J. C. & WILLIAMS, P. H. (1970). Comparison of extracellular polysaccharide of Xanthomonas cam- pestris from culture and from infected cabbage leaves. Canadian Journal of Botany 48, 645-65 1.

TURNER, P., BARBER, C. & DANIELS, M. (1984). Behaviour of the transposons Tn5 and Tn7 in Xanthomonas campestris pv. campestris. Molecular and General Genetics 195, 101-107.

WHITFIELD, C., SUTHERLAND, I. W. & CRIPPS, R. E. (198 1). Surface polysaccharides in mutants of Xantho- monas campestris. Journal of General Microbiology 124, 385-392.

WILLIAMS, P. H. (1980). Black rot: a continuing threat

507-522.

to world crucifers. Plant Disease 64. 736-742.