Taxonomy and Pathogenicity of Erwinia Cacticida Sp. Nov. S

Home , Erwinia

INTERNATIONAL JOURNALOF SYSTEMATIC BACTERIOLOGY, Apr. 1991, p. 197-212 Vol. 41, No. 2 0020-7713/91/020197- 16$02.00/0 Copyright 0 1991, International Union of Microbiological Societies

Taxonomy and Pathogenicity of Erwinia cacticida sp. nov. S. M. ALCORN,l T. V. ORUM,’* ARNOLD G. STEIGERWALT,* JOAN L. M. FOSTER,3 JAMES C. FOGLEMAN,3 AND DON J. BRENNER2 Department of Plant Pathology, University of Arizona, Tucson, Arizona’; Centers for Disease Control, Atlanta, Georgia2; and Department of Biological Science, University of Denver, Denver, Colorado3

A total of 108 pectolytic, soft-rotting Erwinia strains were collected from 11 types of cacti growing in Arizona, Texas, northern Mexico, and Australia between 1958 and 1989. Four strains were collected from soils beneath or close to naturally rotting saguaro cacti. Collectively, these strains caused soft rots of saguaro, organ pipe, and senita cacti, Opuntia (cactus) fruits and pads, tomato fruits, and potato slices, but only occasionally caused soft rots of slices of carrot roots. A numerical cluster analysis showed that 98 of the 112 strains formed a uniform group (cluster 1A) that was distinguished from other pectolytic erwinias by an API 20E code of 1205131, by negative reactions in API 5OCHE tests for L-arabinose, myo-inositol, D-cellobiose, melibiose, and D-raffinose, and, in supplemental tests, by positive reactions for malonate and growth at 43°C. The average levels of DNA relatedness of 22 cluster 1A strains to the proposed type strain (strain 1-12) as determined by the hydroxyapatite method were 88% in 60°C reactions (with 1% divergence within related sequences) and 87% in 75°C reactions. The levels of relatedness to the type strains of other Erwinia spp. were 138% in 75°C reactions. Cluster 1A strains also had a characteristic cellular fatty acid profile containing cyclo-(11,12)- nonadecanoic acid (C19:ocycle cll-lz)and missing tridecanoic acid (CI3J, heptadecanoic acid (C1,:o), and cis-9-heptadecenoicacid (Cl,:l g), which separated them from other pectolytic erwinias. Collectively, these data indicate that the members of cluster 1A are members of a new species, which we name Erwinia cacticida. Three cactus strains in cluster 1B appear to represent a second new species that is closely related to E. cacticida; these strains are designated E. cacticida-like pending the availability of additional strains for testing. The remaining cactus strains (in cluster 4) have the physiological, DNA, and fatty acid profiles of Erwinia carotovora.

To our knowledge, Johnston and Hitchcock (22) were the Because of the anomalies in the original description of first workers to describe a bacterial soft-rot disease of cacti “Erwinia carnegieana,” the disparity between that descrip- in the United States. The cultures of these authors were tion and the very brief account by Johnston and Hitchcock isolated from prickly pear cacti (Opuntia tomentella Berger of their soft-rot pathogen of Opuntia spp., and the observa- and OpuntiaJicus-indica (L.) Mill.) that were originally from tions of Alcorn and Schuyler on their saguaro cactus strains, Guatemala and Columbia but were growing in the U.S. taxonomic tests of bacteria isolated from various naturally Department of Agriculture plant introduction garden in infected, soft-rotting cacti were initiated. During this work, Florida. The bacterium was briefly characterized as “an it was determined that apparently no culture with the char- actively motile, gram-negative, aerobic, and facultative an- acteristics of Erwinia carnegieana as described previously aerobic bacillus” which produced an acid reaction when it exists. (The type culture is a Klebsiella pneumoniae strain was grown in broth containing “glucose, saccharose, man- which is not pathogenic in cacti.) Furthermore, repeated nite, and salicin but none in maltose, lactose, dulcite, and attempts to isolate the previously described bacterium arabinose” (22). Subsequently, another new bacterial spe- failed. Therefore, a recommendation to reject the name cies, Erwinia carnegieana Standring 1942 (23), was de- Erwinia carnegieana was made (3). We report here other scribed (3, 6, 23, 28) as being the causal agent of bacterial results of our taxonomic studies, describe a new Erwinia necrosis (a soft-rot disease) of saguaro cacti (Carnegiea species, and indicate how this species can be identified by gigantea Britt. & Rose). Among the soft-rot erwinias, this using fatty acid analyses and API 20E and API SOCHE test species was unique in that it was gram positive, a character- strips (Analytab Products, Inc. , Plainview, N.Y.). istic also noted by Boyle (6),and had a host range limited to (A preliminary report of our early taxonomic studies has the saguaro cactus. Alcorn (1) isolated a number of soft-rot been presented previously [2].) erwinias from saguaro cacti and demonstrated that they had a broad host range. Schuyler (26) compared pectolytic MATERIALS AND METHODS strains from saguaro cactus postblooms (permanently closed flowers) with selected Alcorn strains and concluded that all Collection and isolation techniques. The data reported of these organisms were gram-negative bacteria with char- below are based on 88 cultures isolated from 11 varieties or acteristics similar to those of members of the Erwinia species of naturally infected cacti that generally were grow- carotovora group. More recently, Fucikovsky and Jaimes ing in native stands in Arizona (76 strains), Texas (5 strains), (17) isolated three biochemically different cultures of soft-rot Mexico (1 strain), and Australia (6 strains) (Table 1). Each erwinias from Opuntia spp. in Mexico; all were gram- strain listed in Table 1 was assigned a sequence number negative rods with peritrichous flagella. (S/N) and was placed in a cluster on the basis of the results of the numerical analyses described below. The Arizona strains included 1 strain recovered from Acanthocereus * Corresponding author. pentagonus (L.) Britt. & Rose growing in the cactus collec- 1- University of Arizona Agriculture Experiment Station publica- tion at the Boyce Thompson Arboretum, Globe, Ariz., and tion no. 7244. two strains from 0. ficus-indica specimens growing as land-

197 198 ALCORN ET AL. INT. J. SYST. BACTERIOL.

TABLE 1. Erwinia strains and other strains studied, listed in the sequence defined by a cluster analysis of 114 phenotypic characteristics - Year of isolation Site of S," Cluster Taxon StrainU Synonym(s)" Sourcea Host or year isolation received'

1 1A Erwinia cacticida 67-53 Carnegiea gig antea 1967(i) Arizona 2 1A E. cacticida 71-3-3 C. gigantea 1971(i) Arizona 3 1A E. cacticida 106 ICPB EC189 C. gigantea 1958(i) Arizona 4 1A E. cacticida 64-28 C. gigantea 1965(i) Arizona 5 1A E. cacticida 64-30 C. gigantea 1964(i) Arizona 6 1A E. cacticidu 66-186 C. gigantea postbloom 1966(i) Arizona 7 1A E. cacticida 66-50 ICPB EC290 C. gigantea 1966(i) Arizona 8 1A E. cacticida 67-106-6 C. gigantea 1967(i) Arizona 9 1A E. cacticida 66-78 ICMP 7453-81, Stenocereus thurberi 1966(i) Arizona ICPB EC292 10 1A E. cacticida 71-8-22 C. gigantea 1971(i) Arizona 11 1A E. cacticida 66-190 C. gigantea postbloom 1966(i) Arizona 12 1A E. cacticida 67-115-11 C. gigantea postbloom 1967(i) Arizona 13 1A E. cacticida 67-118-9 C. gigantea postbloom 1967(i) Arizona 14 1A E. cacticida 67-52 C. gigantea 1967(i) Arizona 15 1A E. cacticida ICMP 7449-81 ICPB EC297, 71- ICMP C. gigantea 1988(r) Arizona 8-3 (lost) 16 1A E. cacticida 89-6-1 Soil beneath "leaking" 1989(i) Arizona C. gigantea 17 1A E. cacticida ICMP 7451-81' ICPB 283, 71-4A ICMP Acanthocereus 1988(r) Arizona (lost) pentagonus I. 8 1A E. cacticida 83-41-2 C. gigantea 1983(i) Arizona 11 9 1A E. cacticida 84-19-el S. thurberi 1984(i) Arizona 20 1A E. cacticida 64-20 C. gigantea 1964(i) Arizona Z! 1 1A E. cacticida 66-42 ICMP 7448-81, C. gigantea 1966(i) Arizona ICPB EC282 22 1A E. cacticida 583-10-10 C. gigantea 1959(i) Arizona :!3 1A E. cacticida 65-145-2 C. gigantea postbloom 1965(i) Arizona 24 1A E. cacticida 78-25-2' C. gigantea 1978(i) Arizona 25 1A E. cacticida 78-29b S. thurberi 1978(i) Arizona 26 1A E. cacticida 67-116-4 C. gigantea postbloom 1967(i) Arizona 27 1A E. cacticida 78-28' ATCC 49482, Stenocrreus 1978(i) Mexico ICPB EC283 gummosus 2 8 1A E. cacticidu 66-93-3 C. gigantea 1967(i) Arizona 29 1A E. cacticida 82-la' Opuntia jicus-indica 1982(i) Arizona 30 1A E. cacticida DU 7 Fogleman S. gummosus 1986(r) Mexico 31 1A E. cacticidu 83-38b C. gigantea 1983(i) Arizona 32 1A E. cacticidu 87-5a C. gigantea 1987(i) Arizona 33 1A E. cacticida 66-187' C. gigantea postbloom 1966(i) Arizona 34 1A E. cacticida 66-188 C. gigantea postbloom 1966(i) Arizona 35 1A E. cacticida Texas 28' ICPB EC296 Richard son Opuntia phaeacantha 1971(r) Texas var. discata 36 1A E. cacticida 65-164a C. gigantea postbloom 1965(i) Arizona 37 1A E. cacticida Texas 27 Richardson 0.phaeucantha var. 1971(r) Texas discata 38 1A E. cacticida 66-31 0.phaeacantha var. 1966(i) Arizona discuta 39 1A E. cacticida 62-69-1" ATCC 49484, Ferocactus wislezenii 1962(i) Arizona ICPB EC286 (barrel cactus) 40 1A E. cacticida Texas 31 Richardson 0. phaeacantha var. 1971(r) Texas discuta 41 1A E. cacticida 62-55 0. phaeacantha var. 1962(i) Arizona major 42 1A E. cacticidu 62-59-5 ICPB EC221 0.phaeacantha var. 1962(i) Arizona major 43 1A E. cacticida 66-36 0. phaeacantha var. 1966(i) Arizona discata 44 1A E. cacticida 62-70-2' ATCC 49487 Opuntia fulgida (jump- 1962(i) Arizona ing cholla cactus) 45 1A E. cacticida 1-12= ATCC 49481T, C. gigantea 1958(i) Arizona ICMP 1551-66T, ICPB EC186T, Dye EH-3'' 46 1A E. cacticida 62-OP-41-2 0. phaeacantha var 1962(i) Arizona d isca ta Continued on following page VOL. 41, 1991 ERWZNZA CACTZCZDA SP. NOV. 199

TABLE 1-Continued

~~ Year of isolation Site of S/N Cluster Taxon Strain" Synonym(s)O Source" Host or year isolation receivedb 47 1A E. cacticida 87-9-2~' C. gigantea 1987(i) Arizona 48 1A E. cacticida 88-6 C. gigantea 1988(i) Arizona 49 1A E. cacticida 84-19-e2 S. thurberi 1984(i) Arizona 50 1A E. cacticida Texas 29 Richardson 0. phaeacantha var. 1971(r) Texas discata 51 1A E. cacticida 82-lb 0.Jicus-indica 1982(i) Arizona 52 1A E. cacticida 66-24 0.phaeacantha var. 1966(i) Arizona discata 53 1A E. cacticida 67-58 C. gigantea 1967(i) Arizona 54 1A E. cacticida 65-279-8 C. gigantea 1965(i) Arizona 55 1A E. cacticida 64-OP-4 0. phaeacantha var. 1964(i) Arizona major 56 1A E. cacticida 64-OP-8 0. phaeacantha var. 1964(i) Arizona discata 57 1A E. cacticida NCPPB 671 Dye EH-4, 622 NCPPB C. gigantea 1980(r) Arizona (lost) 58 1A E. cacticida 88-3a C. gigantea 1988(i) Arizona 59 1A E. cacticida 88-21-1 C. gigantea 1988(i) Arizona 60 1A E. cacticida 83-42a Soil beneath "leaking" 1983(i) Arizona C. gigantea 61 1A E. cacticida 66-19-1 0. phaeacantha var. 1966(i) Arizona major 62 1A E. cacticida 83-42b Soil beneath "leaking" 1983(i) Arizona C. gigantea 63 1A E. cacticida DU 89-6.1' Fogleman Opuntia stricta 1989(r) Australia 64 1A E. cacticida DU 89-20.1 Fogleman 0. stricta 1989(r) Australia 65 1A E. cacticida 66-34 0. phaeucantha var. 1966(i) Arizona discata 66 1A E. cacticida 89-3-2' C. gigantea 1989(i) Arizona 67 1A E. cacticida DU 89-7.3 Fogleman 0. stricta 1989(r) Australia 68 1A E. cacticida DU 89-8.1' ATCC 49483 Fogleman 0. stricta 1989(r) Australia 69 1A E. cacticida DU 89-5.1 Fogleman 0. stricta 1989(r) Australia 70 1A E. cacticida 72-1 ICMP 7452-81, 0.fulgida 1958(i) Arizona ICPB EC188 71 1A E. cacticida 67-3 Dye EH-1, ICMP 0.phaeacantha var. 1958(i) Arizona 1381, ICPB major EC187 72 1A E. cacticida 66-12 C. gigantea 1966(i) Arizona 73 1A E. cacticida 64-OP-7 Opuntia violacea var. 1964(i) Arizona macrocentra 74 1A E. cacticida 67-54 C. gigantea 1967(i) Arizona 75 1A E. cacticida 556-1-2' 0. phaeacantha var. 1958(i) Arizona major 76 1A E. cacticida 87-10-1~ C. gigantea 1987(i) Arizona 77 1A E. cacticida 87-10-2~ C. gigantea 1987(i) Arizona 78 1A E. cacticida 87-9-1~ C. gigantea 1987(i) Arizona 79 1A E. cacticida DU 89-3.3 Fogleman 0. stricta 1989(r) Australia 80 1A E. cacticida 62-47-2' ICPB EC223 0. phaeacantha var. 1962(i) Arizona discata 81 1A E. cacticida Texas 30 Richardson 0. phaeacantha var. 1971(r) Texas discata 82 1A E. cacticida 623-2' C. gigantea 1959(i) Arizona 83 1A E. cacticida 89-5 C. gigantea 1989(i) Arizona 84 1A E. cacticida NCPPB 672' 623-2 NCPPB C. gigantea 1980(r) Arizona 85 1A E. cacticida 62-63' ATCC 49485, 0. phaeacantha var. 1962(i) Arizona ICPB EC293 major 86 1A E. cacticida 85-13b' C. gigantea 1985(i) Arizona 87 1A E. cacticida 71-8-23 ICPB EC294 C. gigantea 1971(i) Arizona 88 1A E. cacticida 84-19-e3 S. thurberi 1984(i) Arizona 89 1A E. cacticida 71-8-18 C. gigantea 1971( i) Arizona 90 1A E. cacticida 71-8-8' C. gigantea 1971 (i) Arizona 91 1A E. cacticida 75-14-2 C. gigantea 1975(i) Arizona 92 1A E. cacticida 65-95-5 C. gigantea postbloom 1965(i) Arizona 93 1A E. cacticida 78-29a' ATCC 49486 S. thurberi 1978(i) Arizona 94 1A E. cacticida 67-111-7' C. gigantea postbloom 1967(i) Arizona Continued on following page 200 ALCORN ET AL. INT.J. SYST.BACTERIOL.

TABLE 1-Continued - Year of isolation Site of S/N Cluster Taxon Straina Synonym(s)a Source" Host or year isolation receivedb

95 1A E. cacticida 89-9 C. gigantea 1989(i) Arizona 96 1A E. cacticida 66-185 C. gigantea postbloom 1966(i) Arizona 917 1A E. cacticida 88-10-b" C. gigantea 1988(i) Arizona 98 1A E. cacticida 89-10" C. gigantea 1989(i) Arizona 99 1B E. cacticida-like 88-2' C. gigantea 1988(i) Arizona 100 1B E. cacticida-like 89-4-1 C. gigantea 1989(i) Arizona 101 1B E. cacticida-like 88-4' C. gigantea 1988(i) Arizona 102 2 Erwinia chrysan- ATCC 11663T' ATCC Chrysanthemum sp. 1988(r) themi 103 3 Enterobacter agglo- ATCC 2715jT' ATCC Human 1988(r) merans 1.04 3 Erwinia herbicola ATCC 33243T' ATCC 1988(r) 105 3 Erwinia nigrifuens ATCC 1302tJT" ATCC Juglans regia (walnut) 1988(r) YO6 4 78-31-1 Beta vulgaris (sugar 1978(i) Arizona beet) YO7 4 Erwinia carotovora ICMP 4226-75T' Dye FD-lT, UR-7T, ICMP B. vulgaris 1988(r) California subsp. betavas- NCPPB 2795T, culorum UCPPB 193T, ATCC 43762T 'LO8 4 78-30 Stanghellini B. vulgaris 1978(i) Arizona 109 4 Erwinia carotovora E 173-78 ICPB 173 ICPB 1978(r) subsp. atroseptica 110 4 E. carotovora ATCC 33260T' ATCC Solanum tuberosum 1988(r) subsp. atroseptica (potato) 111 4 Erwinia atroseptica E 28-78 CUCPB E28 CUCPB 1978(r) 112 4 70-37-A1 S. tuberosum 1970(i) Arizona 113 4 70-37-B3 S. tuberosum 1970(i) Arizona 114 4 64-OP-5 0. phaeacantha var. 1964(i) Arizona major 115 4 64-0P-6" 0. violacea var. 1964(i) Arizona macrocentra 116 4 Erwinia carotovora EC 101 ICPB 1970(r) 117 4 E. carotovora EC 6-78 CUCPB EC6 CUCPB 1978(r ) 118 4 E. carotovora EC 14-78 CUCPB EC14 CUCPB 1978(r) 119 4 Erwinia aroideae EA 14-78 CUCPB EA14 CUCPB 1978(r) 120 4 Erwinia carotovora ATCC 15713=' ATCC S. tuberosum 1989(r) subsp. carotovora 121 4 Erwinia atroseptica E 179-78 ICPB EA179 ICPB 1978(r) 122 4 87-2-2 Stanghellini Brassica oleracea var. 1987(r) Arizona botrytis (cauliflower) 123 4 Erwinia carotovora KSB 10 Stanghellini B. vulgaris 1974(r) Arizona subsp. atroseptica 124 4 80-91 Stanghellini Cucurbita foetidissima 1980(r) Arizona (buffalo gourd) 125 4 71-33~ Capsicum annuum 1971(i) Arizona (chili fruit) 126 4 Erwinia sp. S-29 Stone Vegetable 1966(i) Arizona 127 4 Erwinia sp. s-10 Stone Vegetable 1966(i) Arizona 128 4 Erwinia sp. s-11 Stone Vegetable 1966(i) Arizona 129 4 Erwinia sp. S-8 Stone Vegetable 1966(i) Arizona 130 4 74-4 Stanghellini S. tuberosum 1974(r) Arizona 131 4 Erw inia atrosep tica E 153-78 ICPB EA153 ICPB 1978(r) 132 4 Erwinia aroideae EA 13-78 CUCPB EA13 CUCPB 1978(r) 133 4 Erwinia sp. S-17 Stone Vegetable 1966(ij Arizona 134 4 65-214' C. gigantea postbloom 1965(i) Arizona 135 4 67-111-3 C. gigantea postbloom 1967(i) Arizona 136 4 67-114-1 C. gigantea postbloom 1967(i) Arizona 137 4 67-110-8 C. gigantea postbloom 1967(i) Arizona 138 4 67-112-11 C. gigantea postbloom 1967(i) Arizona 139 4 67-117-14 C. gigantea postbloom 1967(i) Arizona 140 4 67-113-7 C. gigantea postbloom 1967(i) Arizona 141 4 65-82-6 Soil beneath decayed 1965(i) Arizona C. gigantea 142 4 65-46-1" C. gigantea 1965(i) Arizona 143 4 Erwinia carotovora EC 208-78 ATCC 495, Dye ICPB 1978(r) EG23, ICPB EC208 Continued on following page VOL.41, 1991 ERWZNZA CACTZCZDA SP. NOV. 201

TABLE 1-Continued

Year of S/N Cluster Taxon Strain" Synonym(s)" Source" Host isolation Site of or year isolation receivedb 144 4 E. carotovora EC 218-78 ICPB 1978(r) 145 4 Erwinia atroseptica E 70-78 CUCPB E70 CUCPB 1978(r) 146 5 Erwinia quercina ATCC 29281TC Dye FB-lT, ICPB ATCC Quercus sp. (oak) 1988(r) EQ1OlT, NCPPB 1852T 147 5 Erwinia salicis ATCC 15712T' Dye EX2T, ICMP ATCC Salix alba (willow) 1988(r) 1587aT,NCPPB 447T 148 5 Erwinia tracheiphila ATCC 33245T' ATCC Cucumis melo 1988(r) 149 5 Erwinia mallotivora ATCC 29573=' ATCC Mallotus japonicus 1988(r) 150 5 Erwinia rubrifaciens ATCC 2929lTC ATCC Juglans regia 1988(r) 151 5 Erwinia psidii IBSBF 435' PDDC 8426 IBSBF 1988(r) 152 5 Erwinia amylovora ATCC 15580T' ATCC Pyrus communis (pear) 1988 (r) 153 5 Erwinia stewartii ATCC 8199T' ATCC 1988(r) 154 6 Erwinia ananas ATCC 33244T' ATCC Ananas comosus 1988(r) (pineapple) 155 6 Erwinia uredovora ATCC 19321T' ATCC Cereal rust uredia 1988(r) 156 6 Erwinia sp. ATCC 11773' ATCC C. gigantea 1988(r) Arizona 157 6 Erwinia rhapontici ATCC 29283Tc ATCC Rheum rhabarbarum 1988(r) (rhubarb) 158 7 Klebsiella oxytoca ICPB 3866 ICPB Cucumis sativus 1978(r) (cucumber) 159 7 Klebsiella ATCC 13883T' ATCC 1988(r) pneumoniae 160 7 Escherichia coli ATCC 1177ST' ATCC Urine 1988(r)

a Abbreviations: ATCC, American Type Culture Collection, Rockville, Md. ; Dye, D. W. Dye, International Collection of Micro-Organisms from Plants, Plant Disease Division, New Zealand Department of Scientific and Industrial Research, Auckland, New Zealand; CUCPB, Cornell University Collection of Phytopathogenic Bacteria, Ithaca, N.Y. ; Fogleman, J. C. Fogleman, Department of Biological Sciences, University of Denver, Denver, Colo.; IBSBF, Instituto Biol6gico de Sao Paulo, Sao Paulo, Brazil; ICMP and PDDC, International Collection of Micro-Organisms from Plants, Plant Disease Division, New Zealand Department of Scientific and Industrial Research, Auckland, New Zealand; ICPB, International Collection of Phytopathogenic Bacteria, University of California, Davis; NCPPB, National Collection of Plant Pathogenic Bacteria, Harpenden, United Kingdom; Richardson, R. H. Richardson, Department of Zoology, University of Texas, Austin; Stone, W. T. Stone, formerly of the Department of Plant Pathology, University of Arizona, Mesa; Stanghellini, M. E. Stanghellini, Department of Plant Pathology, University of Arizona, Tucson; UCPPB, Department of Plant Pathology, University of California, Berkeley. i, Year isolated; r, year received. Isolate used for DNA studies. scape plants in Tucson, Ariz. In addition to the 88 strains agar [Difco Laboratories, Detroit, Mich.], 5 g of agar, and 20 isolated from naturally infected cacti, 1 strain came from a g of peptone in 1liter of distilled water; sufficient 5 N NaOH rot following artificial injury to an agria cactus (Stenocereus was added to adjust the pH to ca. 7.2 after autoclaving) or gummosus (Engelm.) Gibson & Horak) in Mexico (16), 19 glucose yeast extract agar (20 g of glucose, 10 g of yeast strains came from postbloom saguaro cactus flowers that did extract, and 10 g of peptone in 1 liter of distilled water). not show obvious symptoms (26), 4 strains came from soils Single colonies were subjected to at least three cycles of beneath or close to rotting saguaro cacti, 1 strain came from streaking and subculturing to confirm their purity and were a saguaro cactus and was deposited with the American Type retested for pectolytic activity before being stored in nutrient Culture Collection by Bryan and May (strain ATCC 11773 broth (Difco) and on glucose yeast extract carbonate agar [= SIN 156]), and 14 strains came from cultivated crops in slants (glucose yeast extract agar containing 1% CaCO,) at 4 Arizona. In addition, 20 type strains and 13 other reference to 6°C. Stored cultures were transferred every 6 to 9 months. strains were included in the study. Strains S/N 103 and S/N Taxonomic studies have been conducted intermittently 104 were received as type strains of Enterobacter agglom- over the last 30 years and have included not only many erans and Erwinia herbicola, respectively, and are identified strains maintained at the University of Arizona but also as such throughout this paper even though both of these subcultures from strains sent previously to other laborato- organisms have recently been classified as Puntoea agglom- ries. In 1988 and 1989 the largest number of strains (160 erans strains (18). The 14 strains isolated from cultivated strains) and control cultures were tested concurrently (Table crops in Arizona included five subcultures (S/N 126 through 1). Included in the 160 strains was one duplicate (S/N 82 and S/N 129 and S/N 133) of strains used in studies by Stone (31) S/N 84), which varied consistently in fatty acid characteris- and one subculture (S/N 123) tested by de MendonCa and tics. In addition, data from tests performed with 13 duplicate Stanghellini (11). Except as noted below, the data given are cultures were used to estimate the variability of responses to from tests conducted in 1988 and 1989 on strains collected tests; these duplicate cultures were not included in further between 1958 and 1989 (Table 1). analyses. Bacteria that were pectolytic under aerobic conditions Morphological and physiological tests. KOH tests (32) and when they were streaked onto pectate medium (13) (modified Gram staining (9), in which we used alkaline gentian violet, by deleting yeast extract) were subcultured onto either Kopeloff-Beerman iodine, and 0.25% aqueous safranin 0 potato dextrose peptone agar (PDP) (40 g of potato dextrose and decolorized with 30% acetone in 95% ethanol, were 202 ALCORN ET AL. INT. J. SYST.BACTERIOL. conducted by using ca. 13- to 24-h cultures (grown on PDP or cactus was injected with ca. 0.25 ml of a suspension of a ca. glucose yeast extract carbonate agar) of all strains. Smears 24-h culture in SDW with a turbidity between ca. no. 1 and of known gram-positive and gram-negative bacteria also 4 on the McFarland scale (ca. 2 X lo8 to 8 X lo8 CFUlml). were included on each slide as controls. Motility was de- Control plants were similarly treated with SDW alone. All of tected by observing suspensions of cultures of similar ages in the cacti were incubated in a greenhouse at ca. 34°C. nutrient broth under a compound microscope (magnification, Because of quarantine restrictions, plants inoculated with x 600 with phase-contrast or dark-field illumination). The Erwinia ananas, Erwinia mallotivora, Erwinia psidii, Er- flagellation of representative strains was observed by using winia sakis, and strains isolated from Australian material an electron microscope or a compound microscope or both. were incubated under 12 h of fluorescent and incandescent For compound microscopy the bacteria were first stained by light at 34°C in a locked growth chamber. The cacti used for using the rapid technique of Mayfield and Inniss (24). Cata- all inoculations were deprived of water for several weeks lase activity was determined by adding 3% H,02 to bacteria prior to inoculation. Following injection, the plants were growing on PDP or in an API mannitol cupule. The Hugh- watered daily and assessed after 4 days. To do this, each Leifson (21) method was used for oxidation-fermentation plant was severed at the groundline, and the aerial portion tests. was weighed. Subsequently, any rotted material was exca- Most of the media used for physiological tests were vated and weighed. The test was repeated three times for inoculated with sterile distilled water (SDW) suspensions of each strain with each of the two cactus species. Strains were 20- to 28-h cultures grown on PDP. The exceptions were the considered to be pathogenic if there was at least 8% rot in at tests involving API 20E and API 5OCHE strips; these test least two of the three plants inoculated. This criterion was preparations were inoculated with bacteria grown on nutri- chosen so that only those strains that unequivocally caused ent agar (Difco). Inoculated substrates were incubated at soft rot were scored as positive. 30"C, and tests were read by using the standard protocol (9) In various years, suspensions of selected strains were or the manufacturer's instructions. The API procedures which we used were essentially those of Mergaert et al. (25) injected into unripened fruits of the Indian fig cactus (0. and Verdonck et al. (33). Over the years, we have found that $ficus-indica) and ripened tomato fruits (Lycopersicon escu- the API 20E Voges-Proskauer, citrate, and nitrate tests lentum (L.) Mill.) that had been surface sterilized with 2% occasionally give false-negative results. Therefore, every Amphyl (Lehn and Fink Industrial Products Div. , Sterling isolate which gave a negative result for these characteristics Drugs, Inc., Montvale, N.J.). Suspensions (ca. 1 ml) of in 1988 and 1989 was retested and scored negative in our data selected strains also were injected into separate pads of 0. set only if the repeat test result was also negative. ficus-indica (infiltrated tissues temporarily appeared to be In supplemental tests, acid production from carbohydrates water soaked) and into 2.5- to 18-cm-tall saguaro and senita was detected in purple broth base (Difco) in 1978, in 1% (Lophocereus schottii (Engelm.) Britt. & Rose) cacti. At peptone (Difco) broth base containing bromthymol blue least two separate inoculations were made per strain. Refer- indicator in 1971, and in phenol red broth base (Difco) in ence strains were inoculated similarly. The inoculated ma- 1964 and 1962. The production of reducing substances from terial was incubated appropriately and observed for 7 to 10 sucrose was detected by using the procedures of Dye (13) days. A strain was considered to be pathogenic only when a with (1989 tests) and without beef extract. Brown, orange, or major portion of the seedling or fruit rotted or when disease olive brown reactions after 24 h were recorded as positive; in the pads progressed beyond the area of initial water green, marine, and blue reactions were recorded as negative. soaking. Malonate and tartrate media were made in 1989 as described Carrot (Daucus carota L.) roots and potato (Solanum by Dye (13) or by substituting malonate and tartrate for tuberosum L.) tubers were first thoroughly scrubbed in citrate in Simmons citrate agar (Difco). All other tests were running warm tap water (detergent was added in 1989) and performed by using the methods of Edwards and Ewing (15) then either surface sterilized with 2% amphyl or (in 1989) or the methods described in the Manual of Microbiological soaked in a 0.5% sodium hypochlorite solution containing Methods (9). ca. 1 ml of Tween 80 per 4 liters for 5 min. The treated carrot Growth and acid production at 30, 36, and 43°C also were roots and potato tubers were sliced aseptically into ca. determined. For these tests, dilute suspensions (ca. no. 1 on 2-mm-thick sections and either placed directly on two sheets the McFarland scale [turbidity standards were obtained from of new, 9.0-cm-diameter filter paper in sterile, plastic petri Remel Microbiology Products, Lenexa, Kans.]) of 18- to plates (in 1989) or placed on sterile, glass V-shaped rods that 24-h cultures were made in SDW. Single loopfuls (loop were resting on similar filter paper in petri dishes (previous inside diameter, 3 mm) of the suspensions were transferred years). The filter papers were moistened to saturation with to 10-ml portions of Difco purple broth base (containing 1% SDW. Two drops of a bacterial suspension (turbidity in the dextrose) and to Dye YS broth (13). Tests were carried out range of McFarland standards no. 1to 4) were placed on the in a circulating water bath at 43"C, in a circulating air center of each slice (three slices per strain in 1973 and one incubator at 36"C, and in a standard radiant heat incubator at slice per strain in 1978 and 1989), after which the plates were 30°C. The tubes were preincubated for 5 days at the desired sealed with either Parafilm or masking tape (Shurtape, temperatures to check for contamination, inoculated, and Hickory, N.C.) and incubated at 30°C. Reactions were then immediately returned to the appropriate incubation considered positive if tissue softening or maceration oc- units. Readings were made after 5 days. To prevent contam- curred within 2 days as determined by gentle probing of the ination, the caps of all of the tubes in the water bath were slices with a sterile glass rod. In 1989 the controls consisted sealed with parafilm (American Can Co., Greenwich, of end and middle slices of potato tubers and carrot roots. Conn.). Each tuber or root slice was placed in a separate plate and Soft-rot tests. The soft-rot capabilities of the strains tested inoculated with SDW. If the controls rotted, all bacterial were determined in 1989 by using single, greenhouse-grown inoculations of slices from the same potato tuber or carrot seedlings of saguaro and organ pipe (Stenocereus thurberi root were repeated. For inoculations in previous years, the (Engelm.) S. Buxb.) cacti that were ca. 3 to 12 cm tall. Each control slices were placed on V rods in the same plates as VOL.41. 1991 ERWINIA CACTICIDA SP. NOV. 203 inoculated slices. If the controls rotted, the test was repeated. Fatty acid analyses. Quantitative analyses of cellular fatty acid compositions were performed by using a gas-liquid chromatographic procedure similar to that described by De Boer and Sasser (10). Strains were grown on tryptic soy broth agar (Difco) for 24 h at 28°C. Several loopfuls of cells were added to 1.0 ml of 3.75 M NaOH in 50% aqueous methanol and heated for 30 min in a sealed tube in a boiling water bath. The saponified material was then acidified by adding 1.08 ml of 6 M hydrochloric acid, the fatty acids were methylated by adding 0.92 ml of methanol, and the resulting mixture was heated in a sealed tube for 10 min at 80°C. After cooling to room temperature, the fatty acid methyl esters were extracted with 1.2 ml of hexane-diethyl ether (1:l)and washed with 3.0 ml of 0.3 M NaOH. Fatty acid determinations were made by using a microbial identification system (MIS) (Hewlett-Packard, Palo Alto, Calif.) which included a Hewlett-Packard model 5890A gas FIG. 1. Peritrichous flagella on a cell of Erwiniu cctcticidu 1-12= chromatograph equipped with a flame ionization detector. (A) and fimbriae on a cell of the same strain (B). Bars = 1.0 pm. With this system, we used a phenylmethyl silicone fused Cells were negatively stained in 1% phosphotungstic acid for 2 rnin silica capillary column (25 m by 0.2 mm) and hydrogen as the and photographed by using an Hitachi model H500 electron micro- carrier gas. The gas chromatograph was temperature pro- scope. grammed from 170 to 270°C at a rate of 5"Urnin. Identifica- tion of cellular fatty acids (via their methyl esters) was an automated function of the MIS in which a Hewlett-Packard gram and agglomeration schedule are the same as those model 9000 series 300 computer was used along with the MIS based on the modified Gower similarity matrix, after trans- software program (version 3.0). The Hewlett-Packard model formation by the following formula: SG = 1 - (Dcb/114), 3392A integrator component of the MIS provided the data where D,, is the distance between strains or clusters in the used by the system to estimate the relative percentages of city block distance matrix, S, is the corresponding element the fatty acids that were present in each sample. in the similarity matrix, and 114 is the number of character- Numerical analyses. Test error and reproducibility were istics per strain. analyzed by using data from 13 code-numbered, duplicate DNA hybridization. DNA hybridization studies were done strains and formulas recommended by Austin and Priest (5). by using the hydroxyapatite method and previously de- The following three categories of duplicates were analyzed: scribed techniques (7). Strains 1-12T (= S/N 4ST) (T = type recent (1988) subcultures (five pairs), pre-1977 subcultures strain), 88-2 (= S/N 99), and DU 89-8.1 (= S/N 68) were and subcultures of our strains returned to us from other labeled enzymatically in vitro by using [32P]dCTP provided collections (six pairs), and the same reference strain from in a nick translation reagent kit (Bethesda Research Labo- two different collections (two pairs). Individual test vari- ratories, Inc., Gaithersburg, Md.) as directed by the manu- ances, probability of error for an individual test, pooled facturer. Labeled DNAs were reacted with DNAs from variance, and average probability of an erroneous test result selected cultures (Table 1) that represented the range of were calculated. strains isolated from cacti, as identified by dendrographic Each of the 160 bacterial strains was evaluated for 130 analyses, and with DNAs from the various type cultures. A characteristics, including soft-rot capability and the occur- total of 49 of the DNA source strains were tested as rence of 39 separate fatty acids (the total number of fatty numbered strains; the identities of these strains were not acids identified from all strains). Thirteen of these charac- revealed until the conclusion of the study. Two strains (S/N teristics were either positive or negative for all 160 strains 63 and S/N 68) were identified as cactus strains from and were not used to generate dendrograms. Three (phos- Australia and were tested separately from the others. Three strains, strain ICPB EC186T, Erwinia cypripedii EC155T, phatase and fatty acids and C15:1 11) of the 130 characteristics were not used because analysis of duplicate and Erwinia milletiae 9572-82T, were tested as known strains revealed an error rate that was too high. Most of the strains; these three strains were used only in DNA tests and characteristics analyzed were coded 0 for negative or 1 for were not included in the tests used for numerical analysis. positive; the exceptions were the API SOCHE tests, in which positive reactions after 3, 6, 24, and 48 h of incubation at RESULTS 30°C were coded 5,4, 3, and 2, respectively, as described by Mergaert et al. (25). A modified Gower similarity index (S,) The typical pectolytic cactus strain (for convenience, this (27) was calculated, with the API 5OCHE codes weighted by term includes the pectolytic strains isolated from soils be- dividing by 5 (the possible range of values) and with all neath or close to rotting saguaro cacti but not the nonpec- characteristics (negative, partial, and positive matches) tolytic strain isolated from a saguaro cactus, strain ATCC counted in the denominator of the index (the S, was 11773) was a capsule-producing, gram-negative, nonspore- modified by including negative matches). A dendrogram was forming, rod-shaped organism (0.5 to 1.2 by 0.5 to 1.7 pm) generated by using the Cluster program of SPSS/PC+ that usually occurred singly (Fig. l), but occasionally oc- (SPSS, Inc., Chicago, Ill.) and the unweighted pair group curred in chains of several cells. The rapid, 3% KOH Gram average method of clustering, Weighted coded data for each test (32) for all cactus strains was negative (stringy); this strain were used by the SPSS program to calculate a distance result is in agreement with traditional Gram stain results. matrix, using the city block measure. The resulting dendro- Certain strains (e.g., strains 64-30 [= S/N 51, 65-164a [= S/N 204 ALCORN ET AL. INT. J. SYST.BACTERIOL.

O/o SG 70 80 90 I00 1 Type Strains in Cluster I 1 I I I 1 I Cluster ldentif icotion

IA Eiwinia cacticida

16 2 € chrysanthemi 3 Enterobacter agglomeruns, Erwinia herbicolu, € nigrifluens

E carotovo ra a frosept icu, 4 E carot o vofa & e tovus cu loru m, E.corotoYoro curofovoru

€. quercina, E solicis, € tracheiphilu, 5 E mallotivoro, E rubrifaciens, E psidih E amylo voru, € stewartii 6 €. ononus, E.uredovora, E rhopontici 7 Klebsie No pneumoniae, Escherichio GO/;

I I I I I I I I 70 80 90 I00 '10 SG FIG. 2. Simplified dendrogram showing major clusters and the type strains located in them. SG is the Gower similarity index modified to count negative matches. The unweighted pair group average method of clustering was used. The length of the base of each triangle is proportional to the number of strains in the cluster. A total of 160 strains are represented in the dendrogram, which is based on 114 characteristics per strain.

361, and 87-10-2x [ = S/N 771) appeared to form more chains test (negative), and fatty acids C12:o (positive), C14:o (posi- than others. The cells were motile by means of peritrichous tive), c16:o (positive), and C16:1 (positive). The follow- flagella; fimbriae could occur (Fig. 1). ing API SOCHE characteristics were positive for all strains Colonies that were 18 to 24 h old on PDP were translucent but were included in cluster analysis (Fig. 2) because strains ivory and convex with entire margins. Such colonies fre- turned positive at different rates: D-glucose, D-fructose, quently had striations of dense material, giving a barred D-mannitol, and N-acetylglucosamine tests. The analysis of appearance. Older colonies darkened slightly, becoming duplicate strains indicated that the average probability of an opaque and losing the striations. erroneous test result for the entire data set was less than 2%. Numerical analyses. The following characteristics were However, three characteristics (phosphatase and fatty acids positive or negative for all 160 strains tested and thus were c,,:, not included when we generated the dendrogram: oxidation and 11) showed individual error rates of more than in the oxidation-fermentation test (positive), fermentation in 15% and therefore were deleted from the database. This left the oxidation-fermentation test (positive), catalase (posi- a total of 114 observations per strain that were used in tive), Gram staining (negative), oxidase (negative), API 20E developing a dendrogram (Fig. 2). The data utilized were arginine dihydrolase test (negative) and H,S test (negative), from the 1988 and 1989 physiological tests, from the 1989 API SOCHE L-xylose test (negative) and P-methylxyloside soft-rot tests, and from motility, temperature, and fatty acid VOL. 41, 1991 ERWZNZA CACTZCZDA SP. NOV. 205

TABLE 2. Soft-rot capabilities of cactus and reference strains categorized into clusters by numerical analysis"

~~ Cluster 1A Cluster 1B Clusters 2 and 4 Clusters 3, 5, and 6 Cluster 7 Host ~ No. of % of strains No. of % of strains No. of % of strains No. of % of strains No. of % of strains strains positive strains positive strains positive strains positive strains positive Cacti Carnegiea gigantea 98 83 3 100 41 44 15 0 3 0 Lophocereus schottii 14 71 6 100 8 0 Opuntia ficus-indica Pads 9 78 1 100 8 0 Fruits 68 91 41 66 15 0 Stenocereus thurberi 98 93 3 100 41 37 15 0 3 0 Noncac ti Daucus carota root slices 98 9 3 0 41 76 15 0 3 0 Lycopersicon esculentum fruits 58 100 23 70 5 0 Solanurn tuberosum tuber slices 98 97 3 100 41 100 15 6 3 0

~ ~~ a Data from 1989 inoculations of saguaro and organ pipe cacti, carrots, and potatoes, 1978 inoculations of senita and Indian fig cacti, and 1973 inoculations of tomatoes. The clusters correspond to the following taxa: cluster lA, Erwinia cacticidcr; cluster lB, Erwinia cacticida-like; cluster 2, Erwinia chrysantherni; cluster 4, Erwinia carotovora; clusters 3, 5, and 6, other Erwinia spp.; cluster 7, Eschericia coli and Klebsiella spp.

studies (see below). The cluster analysis indicated that most both types of strips were included in the numerical analysis cactus strains fell into one major group, cluster 1A (Fig. 2 to generate Fig. 2. When the test results for the nine API 20E and Table 1). This cluster contained 83 strains isolated from carbohydrates, as well as acid production from glucose at 36 soft rots of 11 types of cacti from Arizona, Texas, northern and 43°C (somewhat duplicated by growth in YS broth at 36 Mexico, and Australia, 12 strains isolated from the corolla and 43"C), were omitted from the data set, the resulting areas of apparently normal saguaro cactus postblooms, and dendrogram (not shown) had the same major groupings as 3 strains from soils associated with rotting cacti. Three those shown in Fig. 2. The only difference was that cluster 5 cactus strains made up cluster lB, while 11 strains (mostly (in Fig. 2) was inserted between clusters 3 and 4;clusters 1A isolated from postblooms) were members of cluster 4. Clus- and 1B were unchanged. Also, in identifying cluster 1A ter 1 was distinct at the 83% level from cluster 4, which isolates, the API SOCHE L-arabinose test (no strain positive) contained not only several cactus strains but also a number was somewhat more reliable than the API 20E arabinose test of agricultural strains and the type cultures for the subspe- (6% of strains positive) (Table 3). cies of Erwinia carotovora. In turn, the cluster 4 strains were To determine the stability of test reactions, data from distinct from the remaining Erwinia and non-Erwinia refer- previous determinations were reviewed. Nineteen physio- ence strains. Strain ATCC 11773 of Bryan and May, which logical tests have been repeated with available cactus strains was isolated from a saguaro cactus, was in cluster 6 along at intervals ranging from 10 to 26 years. Except for the with the Erwinia species that were least similar (71%) to the variability in the API 20E test results noted below, results majority of the cactus strains. for the following tests were consistent for tests conducted in Soft-rot tests. Cactus strains, as well as Erwinia caroto- 1962 to 1964, 1970 and 1971, 1977 and 1978, and 1988 and vora and Erwinia chrysanthemi reference strains, caused 1989: pectate (positive), oxidase (negative), nitrate (posi- soft decays of a number of the cacti tested and of tomato tive), Voges-Proskauer (positive), indole (negative), glucose fruits and potato slices (Table 2). However, only a few of the (positive), and sucrose (positive). Results for the following cactus strains rotted carrot slices. (Inoculations of saguaro tests were consistent for tests conducted in 1962 to 1964, cacti in 1962, 1964, 1969, and 1978 and of carrots and 1970 and 1971, and 1988 and 1989: salicin (positive), fructose potatoes in 1973 and 1978 gave similar results.) With one (positive), galactose (positive), dulcitol (negative), and mal- exception, Erwinia, Klebsiella, and Escherichia coli refer- onate (positive). Results for the following tests were consis- ence strains did not cause tissue softening in any test; the tent for tests conducted in 1962 to 1964, 1977 and 1978, and only exception was Erwinia rhapontici, which softened 1988 and 1989: mannitol (positive) and rhamnose (positive). potato slice tissue. Strain ATCC 11773 of Bryan and May, Results for the following tests were consistent for tests which differed so greatly from other cactus strains in phys- conducted in 1970 and 1971, 1977 and 1978, and 1988 and iological tests, also did not cause soft rot in any test. 1989: melibiose (negative) and a-methylglucoside (negative). Physiological studies. The reactions of all of the strains Results for the following test were consistent for tests tested in API 20E and API SOCHE tests and in supplemental conducted in 1962 to 1964,1970 and 1971, and 1977 and 1978: physiological tests are shown in Table 3. Strains were 5% NaCl (positive). Results for the following tests were grouped according to the clusters defined by the dendrogram consistent for tests conducted in 1962 to 1964 and 1970 and (Fig. 2). A taxonomically and ecologically interesting obser- 1971: formate (positive) and KCN (negative). With the vation is that only the cluster 1A strains grew at 43°C. exception of the NaC1, formate, and KCN reactions, the It should be noted that the API 20E and API SOCHE strips results of all of these tests were included in the dendro- contained nine carbohydrates in common. However, the test graphic analysis. The 1978 API 20E test results agreed with conditions differed (e.g., API SOCHE cupules were covered the 1988 results with some variability in the citrate, Voges- with sterile mineral oil). Thus, the results from API SOCHE Proskauer, and nitrate tests (the API 20E media in 1978 were cupules were often different, with the amygdalin and myo- inoculated with suspensions of bacteria grown on PDP [as inositol tests showing the most pronounced differences (Ta- opposed to nutrient agar in 19881 and were more frequently ble 3). As recommended by Verdonck et al. (33), data from negative for citrate in 1978). 206 ALCORN ET AL. INT. J. SYST.BACTERIOL.

TABLE 3. Percentage of positive reactions in physiological and other tests for strains in clusters determined by numerical analysis"

% of positive strains Test Cluster 1A Cluster 1B Cluster 2 Cluster 3 Cluster 4 Cluster 5 Cluster 6 Cluster 7 (n = 98)b (n = 3) (n = 1) In = 3) (n = 40) (n = 8) (n = 4) (n = 3) API 20E tests ONPG hydrolysis" 100 100 100 67 100 50 100 100 Lysine decarboxylase 0 0 0 0 0 0 0 100 Ornithine decarboxylase 0 0 0 0 0 0 0 33 Citrate 95 100 100 0 90 38 75 67 Urease 0 0 0 0 0 0 0 33 Tryptophan deaminase 0 0 0 0 0 0 50 0 Indole 0 0 100 0 20 0 50 67 Voges-Proskauer (acetoin) 96 33 100 100 98 88 75 33 Gelatin liquefaction 0 0 0 0 8 0 0 0 Glucose 100 100 100 100 98 100 100 100 Mannitol 100 100 100 100 100 88 100 100 Inositol 0 0 100 33 0 0 50 67 Sorbitol 0 0 0 33 8 0 75 100 Rhamnose 99 33 100 100 95 0 100 100 Saccharose 100 100 100 100 100 100 100 67 Melibiose 0 0 100 33 65 13 100 100 Am ygdalin 100 100 100 100 100 25 100 67 L- Arabinose 6 100 100 100 100 38 100 100 Nitrate reduction 98 100 100 67 100 25 75 100 API 5OCHE tests Glycerol ( l)d 83 67 100 67 75 75 75 100 Erythritol (2) 0 0 0 0 0 0 25 0 D-Arabinose (3) 0 0 0 0 3 0 0 67 L-Arabinose (4) 0 100 100 100 100 50 100 100 D-Ribose (5) 100 100 100 100 100 88 100 100 D-Xylose (6) 29 0 100 100 98 25 75 67 Adonitol (8) 0 0 0 0 0 0 25 67 D-Galactose (10) 100 67 100 100 100 75 100 100 D-Mannose (13) 100 67 100 100 100 75 100 100 L-Sorbose (14) 0 0 0 0 0 0 25 67 L-Rhamnose (15) 100 100 100 100 95 25 100 100 Dulcitol (16) 0 0 0 0 3 0 50 67 myo-Inositol (17) 0 0 100 67 93 25 100 67 D-Sorbitol (19) 1 0 0 33 8 13 75 100 a-Methyl-D-mannoside (20) 0 0 0 0 0 0 50 0 a-Methyl-D-glucoside (21) 0 0 0 0 35 25 0 67 Amygdalin (23) 7 0 0 0 68 0 75 0 Arbutin (24) 99 100 100 100 100 38 100 67 Esculin (25) 100 100 100 100 100 38 75 100 Salicin (26) 100 100 100 100 100 38 100 67 D-Cellobiose (27) 0 0 100 0 100 0 75 67 D-Maltose (28) 5 0 0 67 33 13 100 100 Lactose (29) 20 0 0 0 93 13 75 100 Melibiose (30) 0 0 100 33 85 25 100 100 Saccharose (31) 100 100 100 100 100 100 100 67 Trehalose (32) 77 0 0 100 93 50 100 100 Inulin (33) 0 0 100 0 3 0 0 0 Melezitose (34) 1 0 0 0 0 0 50 33 D-Raffinose (35) 0 0 100 33 90 25 100 67 Starch (36) 0 0 0 0 0 0 25 67 Glycogen (37) 0 0 0 0 0 0 0 33 Xylitol (38) 0 0 0 0 0 0 50 33 P-Gentiobiose (39) 91 0 0 33 100 25 75 100 D-Turanose (40) 1 0 0 0 8 13 50 33 D-LyXOSe (41) 0 0 0 0 0 0 50 0 D-Tagatose (42) 0 0 0 0 0 0 0 100 D-Fucose (43) 0 0 0 33 0 0 0 0 L-Fucose (44) 0 0 0 0 3 0 0 100 D-Arabitol (45) 0 0 0 100 8 0 50 67 L-Arabitol (46) 0 0 0 0 0 0 25 33 Gluconate (47) 4 0 0 67 23 38 100 100 2-Keto-gluconate (48) 0 0 0 67 0 0 75 67 5-Keto-gluconate (49) 0 33 0 0 0 0 100 33 Pectate 100 100 100 0 100 0 0 33 Continued on following page VOL. 41, 1991 ERWINIA CACTICIDA SP. NOV. 207

TABLE 3-Continued

% of positive strains Test Cluster 1A Cluster 1B Cluster 2 Cluster 3 Cluster 4 Cluster 5 Cluster 6 Cluster 7 (n = 98)b (n = 3) (n = 1) (n = 3) (n = 40) (n = 8) (n = 4) (n = 3) Phosphatase 47 67 100 67 68 38 75 100 Reducing substances 0 0 0 0 30 25 25 0 produced from sucrose Motility 95 100 100 67 80 63 50 33 Erythromycin 1 0 100 33 5 88 75 0 Malonate 100 100 100 67 0 0 25 33 Tartrate 0 0 0 33 0 13 50 33 Rots saguaro cactus 83 67 100 0 43 0 0 0 Rots organ pipe cactus 93 100 100 0 35 0 0 0 Rots carrot slice 9 0 100 0 75 0 0 0 Rots potato slice 97 100 100 0 100 0 25 0 Acid produced from glucose 100 100 100 100 75 25 25 100 at 36°C Growth at 36°C in YS broth 100 100 100 100 75 25 25 100 Acid produced from glucose 99 0 0 0 0 0 0 100 at 43°C Growth at 43°C in YS broth 99 67 0 0 0 0 0 100 Yellow pigment 0 0 0 67 0 13 50 0 See text for the numerical analysis procedures which we used. For API SOCHE tests the results given are the percentages of strains that were positive within 48 h. See text for tests in which reactions were positive or negative for all strains. n is the number of strains tested. Clusters were based on the dendrogram shown in Fig. 2. ONPG, o-Nitrophenyl-P-D-galactopyranoside. The numbers in parentheses are the numbers of the cupules on API SOCHE strips; a missing number indicates that the test was positive or negative for all strains.

Fatty acid analyses. The following fatty acids were present for relatedness to strain 1-12T (= S/N 45T) in 75°C reactions, in all 160 strains tested: dodecanoic acid (C12J, tetrade- in which only closely related sequences can reassociate. All canoic acid (C14:o), hexadecanoic acid (C16:J, and cis-9- of the strains that exhibited more than 70% relatedness at hexadecenoic acid (C16:l 9). In addition, 31 identified and 75°C were subsequently tested for relatedness and diver- 2 unidentified fatty acids and 2 long-chain alcohols were gence in 60°C reactions. Two subcultures of strain 1-12T recovered from the 160 bacterial strains tested (Table 4). The (strains ICPB EC186T and ICMP 1551-66T)were included in chain lengths of the known acids and alcohols ranged from 10 the DNA tests and exhibited 100 and 93% relatedness to to 21 carbon atoms; 26 (66.7%) of these compounds were labeled strain 1-12T at 60"C, respectively. A total of 22 other saturated (Table 4). The following fatty acids were particularly cluster 1A strains were highly related to cluster 1A strain useful for distinguishing between the cactus strains in clusters 1-12T. The average levels of relatedness were 88% (range, 82 1A and 1B and the soft-rot strains in cluster 4: hendecanoic to 99%) in 60°C reactions (with 1.0% divergence in related acid (Cl1:J, 3-hydroxydodecanoic acid (Clzz0 tridecanoic sequences) and 87% (range, 76 to 100%) in 75°C reactions acid (C13:J, 3-hydroxypentadecanoic acid 30H), cis-7- (Table 5). Strains 88-2 (= S/N 99) and 88-4 (= S/N lOl), pentadecenoic acid (C15:l ,), heptadecanoic acid (C1,:J, members of cluster lB, were 62 and 69% related to labeled cis-Pheptadecenoic acid (Cl,:l J, cyclo-(11,12)-nonade- strain 1-12T in 60°C reactions. When 88-2 was labeled, 88-4 canoic acid (C19:O cycle cll-12), and unknown acid 13.961. was closely related at 60°C (85%), whereas 1-12T was not Decanoic acid (Clo:o) was present in most of the cluster 4 strains and in many of the cluster 1 strains but was absent in all closely related (53%). When labeled Australian strain DU (= S/N other strains except the Erwinia stewartii strain (S/N 153). 89-8.1 68) was compared with Australian strain DU The standard MIS library (version 3 .O) identified 37 cluster 89-6.1 (= S/N 63), cluster 1A strain 1-12T, and cluster 1B 1A strains as Enterobacter (Pantoea) agglomerans, 17 strain 88-2, the levels of DNA relatedness at 60°C were 97, strains as Erwinia sakis, 14 strains as Escherichia vulneris, 84, and 60%, respectively. Conversely, cluster 4 cactus 1 strain as Hafnia alvei, and 1 strain as Serratia plymuthica strains 64-OP-6 (= S/N 115), 65-215 (= S/N 134), and 65-46-1 at an acceptable confidence level. The remaining cluster 1A (= S/N 142) exhibited 24, 31, and 44% relatedness, respec- strains were not identified to species. In dendrographic tively, to labeled 1-12Tat 75°C. All type strains exhibited low analyses (data not shown) based only on fatty acid charac- levels of relatedness to 1-12T; the type strains of the Erwinia teristics, the strains in cluster 1 and Erwinia chrysanthemi carotovora group (25 to 38% relatedness at 75°C) were the segregated together and were separate from all other soft-rot closest. Strain ATCC 11773 (= S/N 156) (Erwinia sp.), which strains (cluster 4). However, also included with these organ- was isolated by Bryan and May from a saguaro cactus (4), isms were the following typically non-soft-rot species: Er- was only 11% related to strain 1-12T at 75°C. Strains ATCC winia herbicola, Erwinia mallotivora, Erwinia nigrifuens, 27155T (= S/N 103) (Enterobacter agglomerans), ATCC Erwinia quercina, Erwinia rubrifaciens, Erwinia sakis, and 33243T (= S/N 104) (Erwinia herbicola), and ATCC 15712T Erwinia tracheiphila. We also studied the ratios of fatty (= S/N 147) (Erwinia sakis), type strains of species that acids as suggested previously (10) for subspecies of Erwinia were closely identified with cluster 1A strains in the fatty carotovora, but variability in the ratios precluded their use. acid analysis, exhibited only 9, 22, and 12% relatedness, DNA relatedness. All of the strains tested were screened respectively, to 1-12~at 75°C. 208 ALCORN ET AL. INT.J. SYST.BACTERIOL.

TABLE 4. Percentages of strains in each cluster having detectable fatty acids and long-chain alcohols" - % of strains in: Fatty acid methyl ester structure Cluster 1A Cluster 1B Cluster 2 Cluster 3 Cluster 4 Cluster 5 Cluster 6 Cluster 7 (n = 98)b (n = 3) (n = 1) (n = 3) (n = 40) (n = 8) (n = 4) (n = 3) lo:o 73 33 0 0 98 13 0 0 10:0 3 OH 19 0 0 0 0 0 0 0 1l:O 0 0 0 0 55 0 0 0 12:O 2 OH 0 0 0 0 0 0 25 33 12:O 3 OH 31 0 0 0 58 50 25 0 13:O 1 33 0 0 100 13 100 67 13:O IS0 0 0 0 0 0 0 25 0 14:O IS0 3 0 0 0 0 3 0 0 0 14:O 2 OH 0 0 0 0 0 0 100 33 15:O 35 100 100 67 100 38 100 100 15:O IS0 1 0 0 0 0 0 0 0 15:O ANTEISO 1 0 0 0 3 0 0 0 15:O 3 OH 0 0 0 0 63 0 0 0 15:l CIS 7 0 0 0 0 55 0 0 0 16:O (alcohol) 0 0 0 0 0 0 25 0 16:O IS0 0 0 0 0 3 0 0 0 16:O 3 OH 0 0 0 0 18 13 0 0 16:l CIS 7 0 0 0 0 3 0 0 0 16:l CIS 11 16 0 0 0 53 25 25 33 17:O 8 100 100 33 100 25 100 67 17:O ANTEISO 0 0 0 0 3 0 0 0 17:O 3 OH 0 0 0 0 3 0 0 0 17:O Cyclo 99 100 100 100 45 63 100 100 17:l CIS 9 0 0 100 0 100 13 25 33 17:l CIS 10 1 0 0 0 0 0 0 0 17:l CIS 11 0 0 0 0 45 0 0 0 17:l ISO, A5 1 0 0 0 3 0 0 0 18:O 81 67 100 67 90 63 100 67 18:l TRANS 9 (alcohol) 0 0 0 0 0 0 25 0 18:l CIS 9 0 0 0 33 0 0 0 0 19:O IS0 10 0 0 0 0 0 0 0 19:O CyClO Cll-12 90 67 100 33 0 25 75 67 21:l TRANS 11 1 0 0 0 3 0 0 0 Unknownsc 13.961 28 0 0 0 73 50 100 33 14.503 99 100 100 100 95 100 100 100

a See text for fatty acids which were detected in all isolates. n is the number of strains tested. Designations are the retention times relative to the retention times of known, straight-chain index peaks.

DISCUSSION genus Erwinia, exhibited levels of relatedness of 38% or less to strain 1-12Tat 75°C; the most closely related type culture By using numerical taxonomy, tissue-macerating bacterial was strain ATCC 33260 (Erwinia carotovora subsp. atrosep- strains isolated from cacti and soils beneath or adjacent to tica). rotting cacti (i.e., cactus strains) were compared with type Some strains in cluster 1A have been in international Erwinia strains and reference strains belonging to the genus collections for more than 20 years and have been included in and related genera. Analyses of 114 characteristics (includ- Erwinia ing API, fatty acid, soft-rot, and temperature sensitivity a number of important studies on taxonomy. Gra- tests) indicated that 98 (87.5%) of the cactus strains fell into ham identified strains NCPPB 671 (= S/N 57) and NCPPB cluster 1A (Fig. 2), which had an internal similarity level of 672 (= S/N 84) as Erwinia carotovora strains (19); however, 93%. This cluster joined cluster 1B (three cactus strains) at in a subsequent unpublished numerical taxonomic analysis, the 89% level. No reference strains segregated in either he decided that these two strains are clearly separated from cluster 1A or cluster 1B; the closest reference organisms other soft-rot coliform bacteria (20). Starr and Mandel (30) were Erwinia chrysanthemi (cluster 2) (similarity level, 83%) reported the guanine-plus-cytosine contents of eight strains and Enterobacter agglomerans, Erwinia herbicola, and Er- isolated from saguaro cactus and three types of Opuntia spp. winia nigrijluens (cluster 3) (similarity level, 82%). The (strains S/N 3, S/N 42, S/N 45T, S/N 46, S/N 70, S/N 71, S/N members of all other clusters exhibited less than 80% simi- 75, and S/N 80), including strain 1-12T, which they referred larity to clusters 1A and 1B. to as Erwinia carnegieana; the guanine-plus-cytosine values DNA studies confirmed the uniqueness of cluster 1A; the ranged from 50.8 to 51.7 mol% for the cactus strains, levels of relatedness to strain 1-12Tranged from 82 to 99% at compared with values of 51.0 to 52.6 mol% for Erwinia 60°C for 22 selected cactus strains representing the spectrum carotovora strains. Since the differences were significant of cluster 1A. Conversely, DNAs from 22 reference type statistically, Starr and Mandel concluded that Erwinia carn- cultures, including 20 cultures of organisms belonging to the egieana might be viewed as a separate taxon. Brenner et al. VOL.41, 1991 ERWINIA CACTICIDA SP. NOV. 209

TABLE 5. Levels of DNA relatedness between cactus strain 1-12Tand other cactus strains, Erwinia species, and nonerwinias

Source of unlabeled DNA 32P04-labeledDNA from E. cacticida 1-12Tn % Cluster Species Strain Relatedness % Relatedness at 60"Cb % Divergence' at 75"Cb 1A Erwinia cacticida 1-12T 100 0.0 100 1A E. cacticida ICPB EC186T 100 0.0 96 1A E. cacticida NCPPB 672 99 1.0 94 1A E. cacticida EC 223 98 0.0 96 1A E. cacticida 556-1-2 94 0.0 93 1A E. cacticida ICMP 1551-66T 93 3.0 90 1A E. cacticida 62-63 91 1.0 89 1A E. cacticida 623-2 91 0.5 85 1A E. cacticida 62-70-2 90 1.0 91 1A E. cacticida 78-28 90 0,o 82 1A E. cacticida 78-25-2 89 1.0 88 1A E. cacticida 67-111-7 88 1.5 83 1A E. cacticida Texas 28 87 0.5 100 1A E. cacticida 89-3-2 87 0.5 89 1A E. cacticida 66-187 87 1.0 88 1A E. cacticida 87-9-2~ 87 0.5 79 1A E. cacticida 62-69-1 86 1.5 92 1A E. cacticida 82-la 86 1.0 88 1A E. cacticida 85-13b 86 0.5 86 1A E. cacticida 71-8-8 86 1.o 80 1A E. cacticida ICMP 7451-81 85 2.0 83 1A E. cacticida 88-10-b 84 1.0 84 1A E. cacticida DU 89-8.1d 84d 0.6d 91d 1A E. cacticida 78-29a 83 1.5 76 1A E. cacticida 89-10 82 1.o 83 1B E. cacticida-like 88-4 69 5.5 62 1B E. cacticida-like 88-2 62 7.5 48 2 Erwinia chrysanthemi ATCC 11663T 13 3 Erwinia herbicola ATCC 33243= 22 3 Enterobacter aggtomerans ATCC 27MT 9 3 Erwinia nigrifuens ATCC 1302gT 14 4 Erwinia carotovora subsp. ICMP 4226-75T 25 betavasculorum 4 Erwinia carotovora subsp. ATCC 33260T 53 10.0 38 atroseptica 4 Erwinia carotovora 64-OP-6 24 4 E. carotovora 65-46-1 31 4 E. carotovora 65-214 44 4 Erwinia carotovora subsp. ATCC 15713T 35 carotovora 5 Erwinia quercina ATCC 29281T 14 5 Erwinia salicis ATCC 15712T 12 5 Erwinia tracheiphila ATCC 33245T 3 5 Erwinia mallotivora ATCC 29573T 13 5 Erwinia rubrifaciens ATCC 29291T 14 5 Erwinia amylovora ATCC 15580T 11 5 Erwinia psidii IBSBF 435T 19 5 Erwinia stewartii ATCC 8199T 6 6 Erwinia uredovora ATCC 19321T 8 6 Erwinia ananas ATCC 33244T 16 6 Erwinia sp. ATCC 11773 11 6 Erwinia rhapontici ATCC 29283T 16 7 Escherichia coli ATCC 11775T 7 7 Klebsiella pneumoniae ATCC 13883T 13 Erwinia cypripedii EC15ST (= ATCC 292679' 4 Erwinia milletiae 9572-82T (= ATCC 33261T>e 6 Except as noted below, 32P04-labeledDNA from strain 1-12= was reacted with unlabeled DNAs from the same strain (homologous reaction) and all other strains in duplicate. The reassociation values in homologous reactions ranged from 55 to 81% before normalization. Control reactions, in which labeled DNA was incubated in the absence of unlabeled DNA, showed 0.4 to 2.5% binding to hydroxyapatite; these control values were subtracted before normalization. Percent relatedness values were calculated as follows: percent relatedness = [(percentage of DNA bound to hydroxyapatite in heterologous reaction)/ (percentage of DNA bound to hydroxyapatite in homologous reaction)J x 100. Percent divergence values were calculated on the assumption that each 1% decrease in the thermal stability of a heterologous DNA duplex compared with the thermal stability of the homologous DNA duplex was caused by 1% unpaired bases within the duplex; they were calculated to the nearest 0.5%. In this case, "P04-labeled DNA from strain DU 89-8.1 was hybridized to unlabeled DNA from strain 1-12T rather than the reverse. In the same test, strain DU 89-8.1 exhibited 97% relatedness to strain DU 89-6.1 at 60°C (100% at 75°C) and Worelatedness to cluster 1B strain 88-2. This strain was used only in DNA homology studies. 210 ALCORN ET AL. INT. J. SYST.BACTERIOL.

(8) also concluded that cactus strains belonging to cluster 1A nov., with strain 1-12T as the type strain. We believe that the form a separate taxon, which they named “Pectobacterium use of this name is particularly appropriate considering the carnegieana.” The decision of these authors was based on extended host range of this group. For the present, the tests conducted with strains ICPB EC186T (= 1-12T = SIN distinctive cluster 1B strains are referred to as Erwinia 45T) and ICPB EC223 (= 62-47-2 = SIN SO), which were cacticida-like. approximately 60% related to Erwinia carotovora ATCC 495 Not all pectolytic cactus strains are Erwinia cacticida (= S/N 143). Also, when used as a labeled DNA reference strains. While 86 of 89 strains isolated from cactus rots strain, ICPB EC186T was more closely related to Erwinia (including 46 of 47 strains isolated from saguaro cactus rots) carotovora strains (57 to 71% at 60°C) than to the other are Erwinia cacticida or Erwinia cacticida-like strains, 3 are strains tested belonging to the Enterobacteriaceae (8). The Erwinia carotovora strains. Furthermore, only 12 of 19 results of the DNA studies reported here are consistent with saguaro cactus postbloom strains are Erwinia cacticida the results of the previous work. strains; the remaining 7 strains are also Erwinia carotovora Strains belonging to cluster 1A were included in three strains. This difference in diversity between strains isolated large numerical taxonomic studies of the genus Erwinia. Dye from stem rots and postblooms parallels the differences in (14) concluded from the results of four different methods of yeast communities isolated from decaying cladodes (pads) cluster analysis that the three cactus strains which he studied and fruits of Opuntia stricta observed by Starmer et al. (29). formed a discrete phenon, which he called Erwinia car- The fatty acid data alone, without other characteristics, negieana. Mergaert et al. (25) included strain NCPPB 671 were sufficient to clearly separate Erwinia cacticida strains (received as Erwinia carotovora) among 123 strains in stud- (cluster 1A) from Erwinia carotovora strains, with one ies of the genus Erwinia in which API procedures were used. exception (Erwinia cacticida S/N 82 clustered with Erwinia On the basis of the API profile of 1205131 (with which we carotovora strains). However, analyses in which we used agree) and on the basis of the results of 107 additional SPSS or MIS software (in which the relative abundance of enzymatic tests, strain NCPPB 671 (= S/N 57) (unfortunate- fatty acids was considered) also grouped many strains be- ly, our original culture has been lost) formed a single longing to nonpectolytic Erwinia species plus Enterobacter member cluster that joined the Erwinia carotovora cluster (Pantoea) agglomerans with the Erwinia cacticida strains. between Erwinia carotovora subsp. betavasculorum and Furthermore, the standard MIS library identified many Er- Erwinia chrysanthemi (25). Verdonck et al. (33) included winia cacticida strains as Enterobacter (Pantoea)agglomer- strains NCPPB 671 (= S/N 57), NCPPB 672 (= S/N 84), and ans, Erwinia salicis, or Escherichia vulneris. Therefore, to PDDCC 1382 (not used in this study) in determinations of the be effective in identifying Erwinia cacticida strains, fatty reactions of a large number of members of the Enterobac- acid determinations must be supplemented with pectolytic teriaceae in API 20E and API 5OCHE tests. These three tests in order to differentiate the nonpectolytic species. strains formed phenon 33 of Verdonck et al., which was The strains from Australia merit special mention. During distinct from the other strains tested. (Although the focus of this study a number of bacterial strains isolated from rotting our work was not the taxonomy of the genus Erwinia as a 0. stricta in Australia were obtained by one of us (J.C.F.) for group, within the limits of the number of reference strains other research purposes. Six of these (strains S/N 63, S/N tested, we noted the taxonomic disparity between the type 64, S/N 67, S/N 68, S/N 69, and S/N 79) proved to have the species of this genus, Erwinia amylovora, and the soft-rot fatty acid profiles of Erwinia cacticida and to be pectolytic. erwinias and also greater diversity among the Erwinia Consequently, they were subjected to our battery of tests. carotovora strains [in our cluster 41 than is recognized by the An analysis of the test results (Fig. 2) placed all of these three named subspecies of Erwinia carotovora; certainly this strains in cluster 1A. Furthermore, two of these strains (S/N genus still is in need of a critical evaluation.) 63 and SIN 68) were used in DNA hybridization studies and The results of our studies and the results of the other found to exhibit 84% relatedness to Erwinia cacticida type investigations mentioned above are consistent and collec- strain 1-12Tat 60°C. Therefore, these strains also are Erwinia tively indicate that cluster 1A strains should be recognized cacticida strains. as a separate species. (The two strains belonging to cluster We can find no record that workers deliberately intro- 1B tested exhibited 62 and 69% DNA relatedness to strain duced this or any other bacterium into Australia as part of 1-12T at 60°C and therefore may represent a new species; the campaign to control Opuntia cacti biologically. Since the however, the small number of strains suggests that such a bacterium appears to be uniquely associated with cacti, we demarkation should be delayed pending the availability of postulate that it was possibly accidentally introduced into additional specimens.) Several authors (8, 14, 33) have Australia with prickly pear cacti (Opuntia spp.) or, more identified strains in cluster 1A as Erwinia carnegieana (or likely, was associated with cactus-feeding insects imported Pectobacterium carnegieana), but clearly the strains of in tests to control the prickly pears biologically (12). The cluster 1A differ significantly from the previously described most successful of these insects, Cactoblastis cactorum organism Erwinia carnegieana (3, 23). What name then is Berg, was received from Argentina, where it had been appropriate for this group? associated with bacterial soft rots (see p. 48 in reference 12); Although the taxonomic determinations made by Johnston a similar association has been observed in Australia (12). If and Hitchcock (22) to identify the bacterium associated with our premise is correct, the host and geographic ranges of their Opuntia soft rot are few, the characteristics are in Erwinia cacticida extend in the Western Hemisphere much accord with our observations for cluster 1A (except for their beyond the cacti that we sampled in Arizona, Texas, and description of colonies as “dirty white, inclining to yel- northern Mexico. low”). Furthermore, their description of symptoms in Description of Erwinia cucticida sp. nov. Alcorn, Orum, prickly pears is similar to descriptions of symptoms reported Steigerwalt, Foster, Fogleman, and Brenner. Erwinia cacti- for saguaro cactus strains (1).Johnston and Hitchcock called cida (cac.ti.ci.da. Gr. n. kaktos, prickly plant; Gr. v. cid, to their strain “Bacillus cacticidus,” a binomial that was never kill; M. L. adj. cacticida, cactus killing). Cells are motile, accepted. In deference to their observations, we propose gram-negative, nonsporeforming rods (ca. 0.5 to 1.2 by 0.5 to that the cluster 1A strains be named Erwinia cacticida sp. 1.7 pm) with peritrichous flagella and fimbriae. Capsules VOL.41, 199 ERWNU CACTICIDA SP. NOV. 211

may be produced. The cells of some strains are almost round (e.g., 0.8 by 0.9 km); short chains of cells are frequent in some strains. Colonies grown on PDP for 24 h at 30°C are small, smooth, glistening, circular, entire, slightly convex, ivory, and translucent and frequently have striations. They produce a distinct odor and become opaque, off-white, and butyrous with age. aJ Cultures are catalase positive, oxidase negative, nitrate reducing, facultatively anaerobic, and pectolytic. In addition, they produce acid from a number of carbohydrates, including D-glucose, D-ribose, D-galactose, D-fructose, D- mannose, D-mannitol, L-rhamnose, P-gentiobiose, esculin, Pa mCI NCI h, NI-000 salicin, n-acetylglucosamine, and usually trehalose, but not from L-arabinose, myo-inositol, cellobiose, melibiose, a-methylglucoside, maltose, or dulcitol. Cultures usually do II II I ++ I +, not produce acid from lactose, Strains utilize formate and malonate and grow in the presence of 5% NaCl but not in KCN broth. Strains are usually not sensitive to erythromycin and do not produce reducing substances from sucrose. II II I I I I+ They are most easily recognized by their positive reaction on pectate, growth at 43"C, and the seven-digit API 20E profile of 1205131. Complete biochemical characteristics are shown a+ I+ + +++ I in Table 3. Tests useful in differentiating Erwinia cacticida from other Erwinia species are shown in Table 6. Fatty acid characteristics alone distinguish Erwinia cacticida from Erwinia I+ II + I ++ I carotovora but not from some other Erwinia species or Enterobacter (Pantoea) agglornerans. However, the presence of ClgC0cycle cll-12and the absence of C13:o, C17:o,and C17:1 9, together with a positive test result for pectate a+ a+ + I +a1 hydrolysis, can be used to identify Erwinia cacticida strains. The guanine-plus-cytosine content of the DNA is 50.8 to 51.7 mol% (30). Natural habitats are the soft-rotting tissues of cacti growing in the wild. a+ a+ + I ++ I Strain 1-12 (= ATCC 49481), which was isolated from a saguaro cactus in 1958, is the type strain. Strains ICPB EC186T and ICMP 1551-66Tare subcultures of strain 1-12T. II I1 I I +a1 Strain ICPB EC186T was sent to M. Starr at the International Collection of Phyto athogenic Bacteria in 1959; Dye obtained ICPB EC186-? , labeled it strain EH-3T in his studies, a+ I+ Q I I++ and deposited it with the Plant Disease Division, New Zealand Department of Scientific and Industrial Research, Auckland, New Zealand, in 1966, where it is now labeled strain ICMP 1551-66T. This strain is troublesome. Although a+ +I + a+l I ostensibly a subculture of strain 1-12T, it differs from 1-12Tin its reactions in a number of tests, including tests for trehalose, lactose, reducing substances produced from sucrose, and fatty acids C13:o, C17:o, and C17:1 9. On the II II + I +a1 other hand, strain ICMP 1551-66T exhibits 93% DNA relatedness to strain 1-12T at 60°C.

a1 a1 I I +++ ACKNOWLEDGMENTS We thank Mary Ann Repman (formerly of the Department of Plant Pathology, University of Arizona) for the electron micro- I+ II I I +++ graphs and J. D. Mihail (Department of Plant Pathology, University of Missouri, Columbia) for help in certain pathogenicity tests and in reviewing the manuscript. II I+ + +I I I The work by J.C.F. was supported by NIH grant GM-34820. REFERENCES 1. Alcorn, S. M. 1961. Some hosts of Erwinia carnegieana. Plant Q+ I+ + ++++ Dis. Rep. 455877590. 2. Alcorn, S. M., and T. V. Orum. 1978. Erwinia-caused soft-rot of cacti. Phytopathol. News 12:135. 3. Alcorn, S. M., and T. V. Orum. 1988. Request for an Opinion: %+ ++ + ++++ rejection of the names Erwinia carnegieana Standring 1942 and Pectobacteriurn carnegieana (Standring 1942) Brenner, Steiger- 212 ALCORN ET AL. INT. J. SYST.BACTERIOL.

Walt, Miklos and Fanning 1973. Int. J. Syst. Bacteriol. 38:132- erans (Beijerinck 1888) Ewing and Fife 1972 to Pantoea gen. 134. nov. as Pantoea agglomerans comb. nov. and description of 4. American Type Culture Collection. 1989. Catalog of bacteria and Pantoea dispersa sp. nov. Int. J. Syst. Bacteriol. 39:337-345. bacteriophages, 17th ed. American Type Culture Collection, 19. Graham, D. C. 1972. Identification of soft rot coliform bacteria, Roc kville , Md . p. 273-279. In H. P. Maas Geesteranus (ed.), Proceedings of the 5. Austin, B., and F. Priest. 1986. Modern bacterial taxonomy. Van 3rd International Conference on Plant Pathogenic Bacteria. Nostrand Reinhold (UK) Co. , Ltd., Wokingham, Berkshire, Centre for Agricultural Publishing and Documentation, Wa- United Kingdom. geningen, The Netherlands. 6. Boyle, A. M. 1949. Further studies of the bacterial necrosis of 20. Graham, D. C. 1986. Personal communication. the giant cactus. Phytopathology 39:1029-1052. 21. Hugh, R., and E. Leifson. 1953. The taxonomic significance of 7. Brenner, D. J., A. C. McWhorter, J. K. Leete Knutson, and fermentative versus oxidative metabolism of carbohydrates by A. G. Steigerwalt. 1982. Escherichia vulneris: a new species of various gram-negative bacteria. J. Bacteriol. 66:24-26. Enterobacteriaceae associated with human wounds. J. Clin. 22. Johnston, T. H., and L. Hitchcock. 1923. A bacteriosis of prickly Microbiol. 15: 1133-1140. pear plants (Opuntia spp.). Trans. Proc. R. SOC. S. Aust. 8. Brenner, D. J., A. G. Steigerwalt, G. V. Miklos, and G. R. 47: 162-164. Fanning. 1973. Deoxyribonucleic acid relatedness among erwin- 23. Lightle, P. C., E. T. Standring, and J. G. Brown. 1942. A iae and other Enterobacteriaceae: the soft-rot organisms (genus bacterial necrosis of the giant cactus. Phytopathology 32:303- Pectobacterium Waldee). Int. J. Syst. Bacteriol. 23:205-216. 313. 9. Conn, H. J. (ed.). 1957. Manual of microbiological methods. 24. Mayfield, C. I., and W. E. Inniss. 1977. A rapid, simple method McGraw-Hill Book Co., Inc., New York. for staining bacterial flagella. Can. J. Microbiol. 23:1311-1313. 10. De Boer, S. H., and M. Sasser. 1986. Differentiation of Erwinia 25. Mergaert, J., L. Verdonck, K. Kersters, J. Swings, J. M. carotovora ssp. carotovora and E. carotovora ssp. atroseptica Boeufgras, and J. De Ley. 1984. Numerical taxonomy of Erwinia on the basis of cellular fatty acid composition. Can. J. Micro- species using API systems. J. Gen. Microbiol. 130:1893-1910. biol. 32:796-800. 26. Schuyler, M. E. 1968. M.S. thesis. University of Arizona, 11. de Mendonqa, M., and M. E. Stanghellini. 1979. Endemic and Tucson. soilborne nature of Erwinia carotovora var. atroseptica, a 27. Sneath, P. H. A., and R. R. Sokal. 1973. Numerical taxonomy. pathogen of mature sugar beets. Phytopathology 69: 1096-1099. W. H. Freeman and Co., San Francisco. 12. Dodd, A. P. 1940. The biological campaign against prickly pear. 28. Standring, E. T. 1940. M.S. thesis. University of Arizona, Commonwealth Prickly Pear Board, Brisbane, Australia. Tucson. 13. Dye, D. W. 1968. A taxonomic study of the genus Erwinia. I. 29. Starmer, W. T., M. A. Lachance, and H. J. Phaff. 1987. A The “amylovora” group. N. Z. J. Sci. 11590-607. comparison of yeast communities found in necrotic tissue of 14. Dye, D. W. 1981. A numerical taxonomic study of the genus cladodes and fruits of Opuntia stricta on islands in the Carib- Erwinia. N. Z. J. Agric. Res. 24:223-229. bean Sea and where introduced into Australia. Microbiol. Ecol. 15. Edwards, P. R., and W. H. Ewing. 1972. Identification of 14: 179-192. Enterobacteriaceae. Burgess Publishing Co., Minneapolis. 30. Starr, M. P., and M. Mandel. 1969. DNA base composition and 16. Fogleman, J. C., and J. L. M. Foster. 1989. Microbial coloniza- taxonomy of phytopathogenic and other enterobacteria. J. Gen. tion of injured cactus tissue (Stenocereus gummosus) and its Microbiol. 56: 113-123. relationship to the ecology of cactophilic Drosophila rnojaven- 31. Stone, W. J. H. 1966. A highly virulent erwinia isolate from sis. Appl. Environ. Microbiol. 55:lOO-105. Arizona vegetables. Plant Dis. Rep. 50:414-418. 17. Fucikovsky, L., and J. F. Jaimes. 1982. Bacterial diseases of 32. Suslow, T. V., M. N. Schroth, and M. Isaka. 1982. Application of agave and cactus in Mexico, p. 27-34. In J. C. Lorano (ed.), a rapid method for Gram differentiation of plant pathogenic and Proceedings of the 5th International Conference on Plant Patho- saprophytic bacteria without staining. Phytopathology 72:917- genic Bacteria, Cali, Columbia, 1981. Centro International de 918. Agricultura Tropical, Cali, Columbia. 33. Verdonck, L., J. Mergaert, C. Rijckaert, J. Swings, K. Kersters, 18. Gavini, F., J. Mergaert, A. Beji, C. Mielcarek, D. hard, K. and J. De Ley. 1987. Genus Erwinia: numerical analysis of Kersters, and J. De Ley. 1989. Transfer of Enterobacter agglom- phenotypic features. Int. J. Syst. Bacteriol. 37:4-18.