Agrobacterium tumefaciens Chromosomal Virulence (chv) - NCBI

Plant Physiol. (1989) 91, 113-118

Received for publication March 16, 1989 and in revised form April 20, 1989

0032-0889/89/91/0113/06/$01 .00/0

Variation in Binding and Virulence of Agrobacterium tumefaciens Chromosomal Virulence (chv) Mutant Bacteria on Different Plant Species1 Martha C. Hawes* and Steven G. Pueppke2

Departments of Plant Pathology and Molecular and Cellular Biology, University of Arizona, Tucson Arizona 85721 culture cells. Mutations in either of two loci (chvA and chvB) also render the bacteria nonpathogenic on kalanchoe, tobacco, sunflower, tomato, and Jerusalem artichoke. Matthysse (14) and Thomashow et al. (22) also have obtained mutants that are deficient in binding and in virulence. Thus, there is evidence from several sources that one or more chromosomal genes are necessary for A. tumefaciens binding and transformation. Recently, Hookyaas and Schilperoort (13) reported that a few species, including potato, are susceptible to infection by chvB mutants. These results reopen the question of the role ofbacterial binding in transformation, i.e. if binding-deficient mutants can cause tumors, then binding may not be an important step in disease. We now report results from root cap cell binding assays to determine whether different plants vary in their abilities to bind chvB mutant cells, and whether such differences are correlated with susceptibility of plants to infection with the mutant strains. Root cap cells can be isolated nondestructively from many plants (10) and, therefore, can be used to evaluate binding and tumorigenesis in the same plant. Using this technique, we have observed significant variation in abilities of cells from different plant species to bind wild-type A. tumefaciens strains (11).

ABSTRACT Chromosomal virulence (chv) mutants of Agrobacterium tumefaciens have been reported to be deficient in binding to cells of zinnia, tobacco, and bamboo. The mutants are nonpathogenic on stems of Kalanchoe, sunflower, tomato, Jerusalem artichoke, and tobacco, but they cause tumors on tubers of Solanum tuberosum. We used a root cap cell binding assay to test ability of cells from individual plants of 13 different plant species to bind parent or chv mutant bacteria. The same plants were then inoculated to test for disease response. Cells from nine of the plant species were grossly deficient in their abilities to bind mutant bacteria, and the plants inoculated with mutant bacteria failed to form tumors. In contrast, root cap cells as well as root hairs and root surfaces of S. tuberosum, S. okadae, and S. hougasii bound chv mutant bacteria as well as wild type. Nevertheless, S. tuberosum roots inoculated with mutant bacteria did not develop tumors. Although S. okadae plants inoculated with mutant bacteria formed a few tumors, and S. hougasiH developed as many tumors in response to chv mutants as in response to the parent strain, the tumors induced by mutant bacteria were smaller.

Agrobacterium tumefaciens (Smith and Townsend) is a bacterial pathogen that causes the tumorigenic disease called crown gall on most dicotyledonous and a few monocotyledonous species (4, 5). Infection occurs when the bacteria enter susceptible plants through wounds in the roots and crowns, and Ti plasmid-borne bacterial genes are inserted and maintained in the plant's nuclear genome (reviewed in ref. 16). Stachel and Zambryski (21) have hypothesized that T-DNA transfer occurs via a mechanism that is similar to bacterial conjugation, a process that requires direct contact between donor and recipient cells. Genetic evidence corroborates the hypothesis that binding of bacteria to plant cells is a crucial step in the development of crown gall. Douglas and coworkers (6-8) found that A. tumefaciens strains with transposon insertions within a 1 1-kilobase pair segment of the bacterial chromosome, termed the chromosomal virulence (chv) region, are deficient in the ability to bind to isolated mesophyll cells from zinnia and to tobacco and bamboo suspension 'Supported by grant No. 86-CRCR- 1-2224 from the U.S. Depart-

MATERIALS AND METHODS

Bacteria

Agrobacterium tumefaciens strains and mutants were gifts from E.W. Nester (University of Washington, Seattle, WA), and were described by Douglas et al. (6, 7). A723 contains pTiB6806 in the C58 chromosomal background. The chv mutants A1020, A1038, A1045, and A2501 are derivatives of A723, each of which has a Tn5 insertion in the chromosomal virulence B (chvB) region (7). Al 36 is a Ti plasmid-cured derivative of C58. Al 36 and A723 were maintained on gluconate-mannitol (G) medium (2) slants at 5°C; mutants were maintained on medium containing 100 jig of kanamycin per milliliter. Log phase cultures grown at 30°C in liquid G medium with constant shaking were used in binding assays and for inoculation studies.

ment of Agriculture. This is journal series No. 7051 of the Arizona Experiment Station. 2 Present address: Department of Plant Pathology, University of Missouri, Columbia, MO 6521 1.

Plant Material Seed sources were as follows: Lycopersicon esculentum L.

(tomato), Helianthus annuus (sunflower), Daucus carota L. 113

HAWES AND PUEPPKE

114

(carrot), Cucumis sativus L. (cucumber), and Zinnia elegans (zinnia) were from W. Atlee Burpee, Warminster, PA; Pisum sativum Little Marvel (pea) and Phaseolus vulgaris (pinto bean) were from Royal Seeds, Kansas City, MO; Glycine max L. Merr. McCall (soybean) was a gift from D.A. Whited, Department of Agronomy, North Dakota State University, Fargo; Zea mays L. W64A (corn) was from V.E. Gracen, formerly of Cornell University, Ithaca, NY: Avena sativa L. Victorgrain (oats) was a gift from H.H. Luke, Department of Plant Pathology, University of Florida, Gainesville. Solanum tuberosum seeds were a gift from D.F. Millikan, Department of Plant Pathology, University of Missouri-Columbia; S. hougasii and S. okadae seeds were from John Bamberg, Potato Introduction Station, Sturgeon Bay, WI. Binding

Root cap cells were isolated from aseptically germinated seedlings grown on water agar (0.8%) overlaid with Whatman filter paper, as described previously (l1). Root tips were immersed for 30 to 60 s in 50 uL of sterile distilled water in wells of a microtiter plate, and the water was agitated to release the cells. For binding assays, cells were adjusted by direct counts to a concentration of approximately 100 cells/ 100 uL of water. (Plant cell concentration does not influence binding at levels up to 105 cells/mL [M.C. Hawes, unpublished data].) Viability of root cap cells routinely was assessed by observation of cytoplasmic streaming. The concentration of bacteria in all binding studies was estimated turbidimetri-

Plant Physiol. Vol. 91, 1989

cally and adjusted to 2 x 107/mL, and 50 gL were added directly to root cap cell samples. (A concentration of 107/mL is saturating on pea root cap cells [M.C. Hawes, unpublished data].) After incubation for 2 h at 25°C, a 20-,uL sample of cells was washed over a 10-,um mesh screen with approximately 1 mL of water to remove unbound bacteria. Binding was evaluated by direct microscopic observation in one focal plane of the number of bacteria bound to the perimeter of each cell, as described previously (1 1). In the current study, binding was expressed as the percentage of wild type binding to isolated pea root cap cells, which bound a mean of 13 ± 2 bacteria per perimeter. Tumorigenesis Studies After root cap cells were harvested for the binding assay, the seedlings were stabbed with a dissecting needle at about 5-mm intervals from the crown to the root tip and were then immersed for 5 min in inoculum (1 07/mL). Plants were grown in plastic growth pouches (Northrup King, Minneapolis, MN) in growth chambers at 25°C during the day and 23°C at night (12) or in potting soil in clay pots in growth chambers and were watered daily. Inoculated plants were evaluated for tumor development after 2 to 3 weeks. At least four plants were included for each treatment, and experiments were replicated three to six times. In experiments with S. tuberosum, S. hougasii, and S. okadae, seedlings were grown in soil in 20-cm diameter clay pots for approximately 4 weeks. The plants were then inocu-

Table I. Binding of A. tumefaciens by Root Cap Cells from Selected Plant Species Binding was measured by counting the number of bacteria observed in a single focal plane to be attached to the perimeter of each washed plant cell (1 1). The mean number of parent strain A723 cells bound to pea root cap cells (13 ± 2) was set at 100%, and other values are given in reference to pea. For all plants except Solanum species, values are based on at least 35 cells from each of two or three experiments, or a minimum of 70 cells. Standard deviations for these species were less than 10% of the mean. For Solanum species, which had very low root cap cell yields, values are based on a total of 16 to 24 cells, and standard deviations ranged from 19% to 63%. Data were analyzed by one-way analysis of variance. Wild type binding was significantly (0.05% level) lower than pea only for zinnia, soybean, corn, and oats. Binding of chv mutant bacteria to cells of Solanum species did not differ from wild type binding to pea at the 0.05% level of significance. Wild Type chv Mutant Bacteria Species Pea Tomato Sunflower Carrot Pinto bean Cucumber Zinnia

Soybean Corn Oats Solanum species S. tuberosum S. hougaii S. okadae a = nt not tested.

A723

A136 A1020 A1038 A1045 percentage of wild type binding to pea cells

A2501

100 100 123 95

4 15 8 4

3 15 4 4 0.9 0.3 1.4 2 0.2 0.2

3 38 3

115 100 53 38 8 2

100 98 107 94 100 108 46 46 8 2

91 104 122

85 107 92

70 62 76

70 81 91

73 nta nt

1 0.4 2 1

0.3 0.2

2 10 4 4 0.6 0.6 1.4 2 0.2 0.2

60 123 91

4 1 0.3 3 2 0.2 0.2

BINDING AND VIRULENCE OF chv MUTANTS

115

Mutant Binding

Figure 1. Binding of A. tumefaciens to isolated pea root cap cells. Cells were incubated for 2 h in a suspension of (A) parent strain A723 or (B) chvB mutant Al 020 at a concentration of 107 bacteria/mL and then washed and observed with interference contrast optics (x400).

lated at sites just above and below the soil line by injecting parent or mutant bacteria (10 pL at about 108/mL) with a hypodermic syringe (18 gauge needle) to a depth of 2 to 3 mm. The identity of chvB mutant bacteria in tumors on S. hougasii and S. okadae was confirmed as follows: tumors were ground in a mortar, and the supernatant solution was serially diluted and plated onto growth medium containing kanamycin. Single colony isolates were tested for motility (chvB mutants are nonmotile [3]) by visual observation and by measurement of spreading on swarm plates, and f-ketolactose production was assessed by standard procedures (1). The isolates were used to inoculate pea and S. hougasii as described above, and the presence of Tn5 in the isolates was confirmed by Southern hybridization of DNA with a probe containing part of the Tn5 kanamycin resistance gene. RESULTS Wild Type Binding

Binding levels of A723 and A 136 were very similar among cells of most dicot species, with approximately 11 to 15 bacteria bound per cell perimeter, or 85% to 115% of the mean level of binding to isolated pea root cap cells (Table I, Fig. 1 A). Only the percentage of binding to zinnia and soybean cells was substantially lower than the mean number ofbacteria bound to pea root cap cells. The lower number of bacteria bound to zinnia cells relates to the fact that zinnia root cap cells, with a mean perimeter of 149 ,um, are smaller than those of other plants, which have mean perimeters of 200 to 240 ,m. Wild type binding to soybean cells was less than half the level of binding to pea and other dicot cells, and binding to corn and oat cells was less than 10% the level of binding to pea cells. These values for the binding of A723 are consistent with those previously observed for binding of strain B6 to root cap cells of several species, including corn, oats, soybean, tomato, pea, and cucumber (1 1).

With all plants tested except the Solanum species, chvB mutant binding was visibly less than that of the wild-type strain (Fig. 1 B). Mutant binding to most of the other species was no more than 4% of that of wild type binding to pea (Table I). However, with tomato cells, binding increased to as much as 38% of wild type binding. Binding to monocot cells was virtually nonexistent. In contrast, binding of chvB mutant bacteria to Solanum cells was substantially higher than with any other species and was within the range of wild type binding for all three solanum species. Unfortunately, the Solanum seedlings were very small, fragile, and difficult to handle, and yields of root cap cells ranged only from 0 to 10 cells per root (compared with a yield of several thousand cells per root for most of the other species [10]). Although mutant binding to the root cap cells of all three Solanum species appeared visibly to be much higher than mutant binding to cells of the other species, and values were not significantly different from wild type levels, the extremely low numbers of available cells and the high standard deviations made the results of the standard binding assay equivocal. Therefore, we carried out experiments to visually compare binding to the root surface and to root hairs of pea and the three Solanum species. Virtually no binding was visible on root surfaces or to root hairs of pea roots incubated with mutant bacteria and then washed (Fig. 2A). In contrast, chvB mutant bacteria bound in large numbers over the surface of the root, the root hairs (Fig. 2B), and the root cap (Fig. 2C) of S. hougasii, S. tuberosum, and S. okadae. Tumor Development

All tested plants except the three Solanum species grew well in growth pouches. Tumor development on pouch-grown plants in response to inoculation with A723 was efficient for pea, tomato, sunflower, carrot, pinto bean, cucumber, and zinnia, with 70% to 90% of inoculated plants forming tumors that were comparable to those shown in Figure 3. As others (17, 19) have observed with certain soybean genotypes, "McCall" soybean seedlings were relatively recalcitrant to tumor development: fewer than 10% of plants (3/35) inoculated with strain A723 developed tumors. No corn or oat plants developed tumors in response to A723, and none of the plants except Solanum species developed tumors in response to inoculation with chv mutant bacteria. Initially, attempts were made to grow S. tuberosum plants in growth pouches. Unfortunately, the seedlings developed very slowly, and contamination was a problem; few plants survived longer than 2 weeks. Although several of the surviving potato seedlings inoculated with strain A723 developed tumors when grown in pouches, none of the S. tuberosum plants inoculated with chv mutant bacteria formed tumors. In subsequent experiments with S. tuberosum, S. okadae, and S. hougasii, therefore, seedlings were planted in soil and grown for 3 to 4 weeks prior to inoculation. Efficiency of wild type tumorigenesis on these older plants was generally lower than on seedlings of the dicot species tested in growth pouches. In particular, S. okadae plants developed tough, woody stems and roots and only two of 12 plants inoculated with the parent

HAWES AND PUEPPKE

116

Plant Physiol. Vol. 91,1989

Figure 3. Tumors (arrows) on roots of zinnia seedlings inoculated with strain A723 and incubated in growth pouches for 2 weeks.

Figure 2. Binding of chv mutant Al 020 to root hairs of (A) pea or (B) Solanum hougasii, and (C) to a clump of root cap cells from S. hougasii. Seedling roots were incubated for 2 h in inoculum and then washed with water. Arrows indicate bacteria attached to surfaces of S. hougasii cells.

strain developed visible tumors (Table II). Tumorigenesis in response to mutant bacteria was generally lower: two of 27 S. okadae plants inoculated with chv mutant strains developed visible tumors. In contrast, more than half of the S. hougasii plants inoculated with four chv mutants developed obvious tumors (Table II). However, these tumors (Fig. 4A) were consistently smaller than those caused by A723 (Fig. 4B). In all cases, the identity of the bacteria isolated from galls matched that of the inoculum. Furthermore, single colony isolates of the recovered bacteria failed to cause tumors when used to inoculate pea but did cause tumors on S. hougasli plants.

DISCUSSION The pleiotropic nature of the chv mutants, which are nonmotile as well as binding-deficient, raises questions about whether binding is the only factor limiting virulence in these mutants. Our study confirms and extends previous observations (7, 8) that chvB mutant binding is reduced in a number of plant species, and that the reduced binding is correlated with a loss of virulence on those species. However, we also determined that, with three Solanum species, chvB mutant bacteria can exhibit wild type binding. Despite this finding,

no visible tumors formed on S. tuberosum plants inoculated with chvB mutant bacteria in this study, although potato tubers are apparently susceptible to infection (13). The proportion of tumor formation induced by chvB mutant bacteria on S. okadae plants was even lower than tumorigenesis by the parent strain. Only in the case of S. hougasii did response to inoculation with chvB mutant bacteria approximate that of the wild type. However, the tumors that formed were substantially smaller than those induced by A723. Data from some systems have suggested that more than one type of Agrobacterium binding to plant cells can occur, and that such binding may or may not be of the type that leads to transformation (15). One possible reason that high binding levels do not always lead to virulence is that some or all of the observed binding is of an ineffective nature. Alternatively, the findings suggest that, at least in the tested species, reduced binding ability is not the only deleterious effect of the chvB mutation on virulence of A. tumefaciens. The results confirm and extend previous observations that different plants vary in their abilities to bind A. tumefaciens cells, and that the differences are correlated with disease reaction (1 1). However, it is clear that, even if differences in binding abilities among plants have an impact on susceptibility, such properties are not the only factors that determine host range in plants. Just as binding ability in bacteria is not the only requirement for pathogenesis, ability of plant cells to bind bacteria, if it is a requirement at all, is clearly not the only factor required for susceptibility. For example, vir geneinducing chemicals appear to be one limiting factor in susceptibility in several species, including maize. Arabidopsis, and soybean (18, 20, 23). In addition, T-DNA transfer has been

BINDING AND VIRULENCE OF chv MUTANTS

117

Table II. Tumor Formation on Selected Plant Species Except for the Solanum species, seedling roots were wounded at three sites, then immersed 5 min in inoculum. Subsequently, plants were maintained in growth pouches for 2 weeks and then evaluated for the presence of tumors. Seedlings of Solanum species were grown in soil for several weeks before inoculation just above and below the soil line. Tumor development was evaluated 3 to 6 weeks later. Wild Type chv Mutant Bacteria Species Al 020 Al 045 Al 038 A723 A2501 A2503 plants with tumors/plants inoculated

Cucumber Zinnia

7/10 5/9 7/10 7/11

0/12 0/10 0/10 0/10 0/10 0/2 0/3

0/10 0/9 0/10 0/9 0/9 0/9 0/10

0/10 0/9 0/10 0/9 0/8 0/9 0/8

0/9 0/8 0/9 0/11 0/6 0/9 0/8

Soybean Coin Oats

3/35 0/10 0/12

0/4 0/10 0/11

0/5 0/8 0/7

0/8

0/8

0/6

nta

0/10

nt nt

nt nt

8/14 2/12 6/10

0/10 1/10 3/5

0/10 0/10 4/6

0/10

0/10

0/11

1/7 3/7

nt

nt nt

Pea Tomato Sunflower

Carrot Pinto Bean

Solanum species S. tuberosum S.okadae S. hougasii a = nt not tested.

9/10 10/10 10/10

3.

4. 5.

6. 7. Figure 4. Tumors (arrows) that developed on S. hougasii plants 3 weeks after inoculation with (A) chv mutant Al 020 or (B) strain A723.

detected by virus amplification in maize, even though the cells of this species exhibit limited binding capacities (9). Nevertheless, the striking quantitative correlation between binding and susceptibility among a wide range of plant species and bacterial strains is consistent with the hypothesis that the inability to form stable surface contact with transforming bacteria is at least one of the possible factors that can prevent or reduce the efficiency of tumor development. Analysis of such host factors among different species is made difficult by the fact that some may be affected in several steps. Thus, maize appears to have limiting levels of vir gene-inducing compounds as well as reduced binding levels. We have, therefore, initiated studies to utilize genetic dissection of host response in Pisum sativum, rather than interspecies comparisons, to address questions about the potential significance of binding and other proposed steps in crown gall pathogenesis. LITERATURE CITED 1. Bernaerts MJ, DeLey J (1963) A biochemical test for crown gall bacteria. Nature 197: 406-407 2. Bhuvaneswari TV, Pueppke SG, Bauer WD (1977) Role of lectins

8. 9.

10. 11.

12.

13.

14.

15.

1/5

0/10 0/8 0/10 0/10 0/8 0/8 0/4

in plant-microorganism interactions. I. Binding of soybean lectin to rhizobia. Plant Physiol 60: 486-491 Bradley DE, Douglas CJ, Peschon J (1984) Flagella-specific bacteriophages of Agrobacterium tumefaciens: Demonstration of virulence of nonmotile mutants. Can J Microbiol 30: 676681 DeCleene M (1985) The susceptibility of monocotyledons to Agrobacterium tumefaciens. Phytopathol Z 113: 81-89 DeCleene M, DeLey J (1976) The host range of crown gall. Bot Rev 42: 389-466 Douglas CJ, Halperin W, Nester EW (1982) Agrobacterium tumefaciens mutants affected in attachment to plant cells. J Bacteriol 152: 1265-1275 Douglas CJ, Staneloni RJ, Rubin RA, Nester EW (1985) Identification and genetic analysis of an Agrobacterium tumefaciens chromosomal virulence region. J Bacteriol 161: 850-860 Douglas CJ, Halperin W, Gordon M, Nester E (1985) Specific attachment of Agrobacterium tumefaciens to bamboo cells in suspension cultures. J Bacteriol 161: 764-766 Grimsley N, Hohn T, Davies JW, Hohn B (1987) Agrobacterium mediated delivery of infectious maize streak virus into maize plants. Nature 325: 177-179 Hawes MC, Pueppke SG (1986) Sloughed peripheral root cap cells: Yield from different species and callus formation from single cells. Am J Bot 73: 1466-1473 Hawes MC, Pueppke SG (1987) Correlation between binding of Agrobacterium tumefaciens by root cap cells and susceptibility of plants to crown gall. Plant Cell Rep 6: 287-290 Hawes MC, Robbs SL, Pueppke SG 1989) Use of a root tumorigenesis assay to detect genotypic variation in susceptibility of 34 cultivars of Pisum sativum to crown gall. Plant Physiol 90: 180-184 Hooykaas PJJ, Schilperoort RA (1986) The molecular basis of the Agrobacterium-plant interaction: Characteristics of Agrobacterium virulence genes and their possible occurrence in other plant-associated bacteria. In B Lugtenberg, ed, Recognition in Microbe-Plant Symbiotic and Pathogenic Interactions. Springer-Verlag, Berlin, pp 189-202 Matthysse AG (1987) Characterization of nonattaching mutants of Agrobacterium tumefaciens. J Bacteriol 169: 313-323 Neff NT, Binns AN (1985) Agrobacterium tumefaciens interaction with suspension-cultured tomato cells. Plant Physiol 77: 35-42

118

HAWES AND PUEPPKE

16. Nester EW, Gordon MP, Amasino RM, Yanofsky MF (1984) Crown gall: A molecular and physiological analysis. Annu Rev Plant Physiol 35: 387-413 17. Owens LD, Cress DE (1985) Genotypic variability of soybean response to Agrobacterium strains harboring Ti or Ti plasmids. Plant Physiol 77: 87-94 18. Owens LD, Smigocki AG (1988) Transformation of soybean cells using mixed strains ofAgrobacterium tumefaciens and phenolic compounds. Plant Physiol 88: 570-573 19. Pedersen HC, Christiansen J, Wyndaele R (1983) Induction and in vitro culture of soybean crown gall tumors. Plant Cell Rep 2: 201-204 20. Sheikholeslam SN, Weeks DP (1987) Acetosyringone promotes

Plant Physiol. Vol. 91,1989

high efficiency transformation of Arabidopsis thaliana. Plant Mol Biol 8: 291-298 21. Stachel S, Zambryski P (1986) Agrobacterium tumefaciens and the susceptible plant cell: A novel adaptation of extracellular recognition and DNA conjugation. Cell 47: 155-157 22. Thomashow MF, Karlinsey JE, Marks JR, Hurlbert RE (1987) Identification of a new virulence locus in Agrobacterium tumefaciens that affects polysaccharide composition and plant cell attachment. J Bacteriol 169: 3209-3216 23. Usami S, Morikawa S, Takebe I, Machida Y (1987) Absence in monocotyledonous plants of diffusible factors inducing TDNA circularization and vir gene expression in Agrobacterium. Mol Gen Genet 209: 221-226