Pelistega Europaea Gen. Nov., Sp. Nov., a Bacterium Associated with Respiratory Disease in Pigeons: Taxonomic Structure and Phylogenetic Allocation

International Journal of Systematic Bacteriology (1 998), 48,43 1-440 Printed in Great Britain

Pelistega europaea gen. nov., sp. nov., a bacterium associated with respiratory disease in pigeons: taxonomic structure and phylogenetic allocation

P. Vandamme,'#' P. Segers,' M. RYII,~J. horn me^,^ M. Vancanneyt,' R. Coopman,' R. De Bae~e,~Y. Van De Peer,5 K. Kersters,' R. De Wachter5and K. H. Him3

Author for correspondence: P. Vandamme. Tel: + 32 9 264 51 13. Fax: + 32 9 264 50 92. e-mail : [email protected]

1 Laboratorium voor Twenty-four strains isolated mainly from infected respiratory tracts of pigeons Microbiologie, were characterized by an integrated genotypic and phenotypic approach. An Ledeganckstraat 35, Universiteit Gent, 6-9000 extensive biochemical examination using conventional tests and several API Gent, Belgium microtest systems indicated that all isolates formed a phenotypically Department of Medical homogeneous taxon with a DNA G+C content between 42 and 43 mol%. Microbiology, University Whole-cell protein and fatty acid analysis revealed an unexpected Hospital Antwerp UIA, heterogeneity which was confirmed by DNA-DNA hybridizations. Four main Belgium genotypic sub-groups (genomovars) were delineated. 165 rDNA sequence 3 Klinik fur Geflugel der analysis of a representative strain indicated that this taxon belongs to the Tierarztlichen Hochschule, Hannover, Germany beta-subclass of the Proteobacteria with Taylorella equigenitalis as its closest neighbour (about 94.8% similarity). A comparison of phenotypic and 4 Provinciaal Verbond voor Dierenziektenbestrijding, genotypic characteristics of both taxa suggested that the pigeon isolates Tor hou t, Be1g iu m represented a novel genus for which the name Pelistega is proposed. In the

5 Department of absence of differential phenotypic characteristics between the genomovars, it Biochemistry, University of was preferred to include all of the isolates into a single species, Pelistega Antwerp UIA, Belgium europaea, and strain LMG 10982 was selected as the type strain. The latter strain belongs to fatty acid cluster Iand protein electrophoretic sub-group 1, which comprise 13 and 5 isolates, respectively. It is not unlikely that the name P. eumpaea will be restricted in the future to organisms belonging to fatty acid cluster I, or even to protein electrophoretic sub-group 1, upon discovery of differential diagnostic features.

Keywords: Pelistega europaea, taxonomic structure, pigeon

INTRODUCTION METHODS Bacterial strains and growth conditions. Pelistega europaea The present study was undertaken to fully characterize strains were grown on Trypticase Soy agar (BBL) and a collection of 24 strains isolated mainly from the incubated at 36-37 "C in a microaerobic atmosphere con- respiratory tract of pigeons in Belgium and Germany. taining approximately 5% 0,, 3.5% CO,, 7.5% H, and These strains were obtained at necroscopy from 84 YO N,. Taylorella equigenitalis strains were grown on diseased pigeons and could not be identified by Columbia agar (Becton Dickinson) supplemented with 5 YO conventional biochemical analysis or whole-cell fatty horse blood and incubated under the same conditions acid analysis. The present report describes the poly- described above, except when stated otherwise. phasic approach which was used to determine the The strains and their sources are listed in Table 1. Bac- taxonomic relationships of these isolates and their teriological purity was checked by plating and examining phylogenetic allocation. living and Gram-stained cells. Fatty acid methyl ester analysis. P. europaea strains were The EMBL accession number for the strain LMG 10982Tsequence reported grown for 48 h. T. equigenitalis strains were grown for 48 h in this paper is Y11890. on a chocolate agar base (Oxoid CM55) supplemented with

00632 0 1998 IUMS 43 1 P. Vandamme and others

Table 1. List of strains studied

~ ~ ~ Strain" Other strain no.? Fatty acid Source clusterlprotein electrophoretic sub-group

~ ~~ Pelistega euvopaea strains LMG 10982T Hommez N57 Pigeon (Belgium) LMG 12980 x2 12-90, HinzH6 Pigeon, lung, pneumonia, airsacculitis (Germany) LMG 12988 x858-83, HinzM19 Pigeon, lung (Germany) LMG 15727 ~993-85 Pigeon, trachea, tracheitis, bronchitis (Germany) LMG 15943 Hommez 4 Pigeon (Belgium) LMG 15725 ~708-82 Pigeon, trachea, tracheitis, bronchltis (Germany) LMG 15726 ~707-82 Pigeon, trachea, rhinitis, tracheitis (Germany) LMG 15728 ~207-85 Pigeon, air sac, airsacculitis, salpingitis (Germany) LMG 15934 Hommez 17 Pigeon (Belgium) LMG 16152 ~706B-82 Pigeon, trachea, exudative rhinitis, tracheitis (Germany) LMG 10983 Hommez 118 Pigeon (Belgium) LMG 11609 ~227-90 Pigeon, lung, pneumonia, airsacculitis (Germany) LMG 10986 Hommez 119 Pigeon (Belgium) LMG 12981 x101-90, HinzH7 Pigeon, palatine cleft, respiratory disease (Germany) LMG 12982 x121-90, HinzH8 Pigeon, lung, pneumonia, oophoritis (Germany) LMG 12985 x554-85, HinzM 11 Pigeon, trachea, respiratory disease (Germany) LMG 12991 x433-84, HinzM24 Pigeon, lung, bronchitis (Germany) LMG 15936 Hommez 25 Pigeon (Belgium) LMG 15942 Hommez 1 Pigeon (Belgium) LMG 15944 Hommez 24 Pigeon (Belgium) LMG 12990 x 567- 86, HinzM2 3 Pigeon, lung, pneumonia, airsacculitis (Germany) LMG 15729 ~636-85 Pigeon, lung (Germany) LMG 15730 ~657- 85 Pigeon, liver (Germany) LMG 15731 ~1-85 Pigeon, lung, crop inflammation, nephrosis (Germany) Taylovella equigenitalis strains LMG 6222T CCUG 10786T Horse, contagious metritis LMG6223 CCUG 16464 Horse (Sweden) LMG 6224 CCUG 16465 Horse (Sweden) B 1938-89 Horse (Germany)

. LMG, BCCMILMG Culture Collection, Laboratorium voor Microbiologie Gent, Universiteit Gent, Belgium. t CCUG, Culture Collection University of Goteborg, Department of Clinical Bacteriology, Goteborg, Sweden.

10% (v/v) defibrinated horse blood. A loopful of well- DNA base compositions. All of the mean mol % G + C values grown cells was harvested and preparation, separation, were determined by thermal denaturation and calculated by identification and numerical comparison of the fatty acid using the equation of Marmur & Doty (16), as modified by methyl esters was performed using the Microbial Iden- De Ley (4). tification System (Microbial ID) as described previously DNA-DNA hybridization experiments. The degree of DNA- (25). DNA binding, expressed as a percentage, was determined Analysis of protein electrophoreticpatterns. P. europaea and spectrophotometrically by the initial renaturation rate T. equigenitalis strains were incubated for 48 and 96 h, method of De Ley et al. (5). Each value is the mean of at least respectively. Preparation of cellular protein extracts, PAGE, two hybridization experiments. Values of 30 % DNA bind- densitometric analysis, normalization and interpolation of ing and less do not represent significant DNA homology. the protein profiles, and numerical analysis were performed The total DNA concentration was about 63 pg ml-', and the as described by Pot et al. (20) using the GelCompar 4.0 optimal renaturation temperature in 1 x SSC was 63.9 "C. software package (Applied Maths). The profiles were recorded and stored on a PC computer. The similarity between 16s rRNA sequencing. Part of the rDNA operon, comprising all pairs of traces was expressed by the Pearson product- the almost complete 16s DNA, was amplified by PCR. The moment correlation coefficient converted for convenience to forward primer was AGAGTTTGATCCTGGCTCAG, a percentage value. corresponding to positions 8-27 in 16s rRNA. The reverse Drimer was TCTGTGTGCCTAGGTATCC. comdemen- Preparation of high-molecular-mass DNA. High-molecular- iary to positions 45-26 in 23s rRNA. The PCR reachon was mass native DNA was prepared as described previously (25). carried out in a volume of 100 pl, using 2.5 U Taq polymerase

432 International Journal of Systematic Bacteriology 48 Pelistega europaea gen. nov., sp. nov.

(Boehringer Mannheim), about 350ng DNA, 500nM of sterilized chicken serum. Final readings were done after 3 each primer and 200 JAMdNTPs, in the appropriate buffer. and 5 d incubation at 37 "C. Before addition of Tag polymerase and DNA, the mixture Oxidase activity was tested with a freshly prepared 1 YO was irradiated for 5 min on a UV transilluminator. After solution of N,N-dimethyl-p-phenylene monochloride 2min denaturation at 94"C, the following cycle was (Sigma) on reagent-impregnated filter paper as described by repeated 30 times: 1 min denaturation at 94 "C, 1 min Kersters & De Ley (12). The reactions were read within 30 s. annealing at 50 "C and 2 min polymerization at 72 "C. The Catalase was tested as described by Kersters et al. (13) and last cycle was followed by 10 min elongation at 72 "C. by using the ID-COLOUR-CATALASE test (bioM6rieux). The amplification product was purified by agarose gel Urea hydrolysis was determined as described by Lautrop electrophoresis, ligated in T-tailed pBluescript I1 SK( + ) (15) and on Christensen urea agar (Merck) with daily (Stratagene) as described by Holton & Graham (8), and readings up to 5 d. Nitrate reduction was done in nitrate electroporated into Escherichia coli DH5a cells. The plasmid broth cultures (Merck) and indole production was done in was isolated from 13 single clones according to Birnboim & standard I1 nutrient broth (Merck) with and without 2% Doly (2). This mixture was sequenced with the Pharmacia inactivated filter-sterilized chicken serum after 24 and 48 h Sequencing kit using a set of primers described previously by incubation at 37 "C. Wilmotte et al. (31). Part of the sequence was obtained by Determination of the presence of P-galactosidase (ONPG automatic sequencing with the ALF express (Pharmacia). In test) was carried out as described by Kilian (14) using 2- this case we used Cy5 labelled primers with the Thermo nitrophenyl P-D-galactopyranoside. The determination of Sequenase fluorescent-labelled primer cycle sequencing kit ornithine and lysine decarboxylase and arginine dihydrolase (Amersham). activity was performed by the method of Marller (1 7) using Phylogenetic analysis of the 165 rRNA sequence. The 16s meat peptone no. 100 (Gibco) and beef extract (Oxoid). rRNA sequence of strain LMG 10982Twas included in an Final readings were done after 3 d incubation at 37 "C. alignment of small ribosomal subunit sequences (27). Gelatine hydrolysis was tested on gelatine agar [standard I Neighbour-joining trees were constructed by means of the nutrient agar (Merck) supplemented with 0.4 YO gelatine software package TREECON (26). Sequence dissimilarities (Serva)] with and without 2 % inactivated filter-sterilized were computed using all alignment positions, and converted chicken serum by flooding the plate surface with the reagent into evolutionary distances according to Jukes & Cantor of Frazier (7) after 3 and 5 d incubation. (lo), disregarding insertions and deletions. Bootstrap analy- Hyaluronidase and chondroitinsulfatase activities were sis (6) was carried out using 1000 replicas. tested according to KafEa (1 1) by means of a Streptococcus Nucleotide sequence accession numbers. The nucleotide egui strain and according to Smith & Willet (21), respect- sequence accession number of strain LMG 10982Tis Y 11890. ively. The reactions were read after 2 and 3 d incubation, The following strains (with their accession numbers) were respectively. used to illustrate the phylogenetic allocation of strain LMG For the methyl red and Voges-Proskauer test (production of 10982Tas shown in Fig. 3 : Alcaligenes sp. 05-36 (X86580), acetylmethylcarbinol), the glucose/phosphate medium con- Alcaligenes faecalis (M22508), Alcaligenes xylosoxidans sisted of 0.5% peptone (peptone no. 100; Gibco), 0.5% (M22509), Bordetella avium (U04947), Bordetella bronchi- glucose and 0.5% K,HPO,, pH 6.9. Final reactions were septica (U04948), Bordetella holmesii (U04820), Bordetella determined after 2 d incubation by adding methyl red or the parapertussis (U04949), Bordetella pertussis (U04950), T. reagents of Barrit (3). equigenitalis (X68645) and Leptothrix mobilis (X9707 1). Acid production from carbohydrates was tested in phenol Phenotypic tests. Strains were grown on Columbia agar base red broth base (Difco) plus 1% carbohydrate with and (Oxoid) supplemented with 7 % defibrinated sheep blood without 2 YOinactivated filter-sterilized chicken serum. Each unless stated otherwise. Cells used as inocula for biochemical of the tubes was inoculated with 0.05 ml of a dense bacterial tests were grown for 16-24 h under microaerobic conditions suspension. Uninoculated medium with carbohydrate and in a moist chamber. Bacteria were suspended in 0.15 M inoculated medium without carbohydrate served as controls. NaCl solution, pH 7-0& 1, to a concentration of about 10' Acid production was examined daily up to 6 d. Acid viable bacterial cells ml-', except when mentioned otherwise. production from carbohydrates was also tested by using the Growth was recorded daily at least up to 3 d. Standard I buffered single-substrate test [buffered saline (KH,PO,, nutrient broth (Merck) was used to test growth in liquid 0.1 YO; K,HPO,, 0.1 YO; NaCl, 0.5 YO; pH 7.6) containing 2 YO medium and motility. carbohydrate and phenol red as indicator] (15). A heavy Capsule and flagella staining were done according to Moller inoculum (100 pl of a suspension of about 400 mg wet (18) and Walc-Pokrzywnicki (29), by use of bacterial cells bacterial mass in 2 ml 0.15 M NaCl solution) was added to harvested from blood agar cultures after 16 h incubation at 0.5 ml buffered carbohydrate solution in a tube with di- 30 or 37 "C. ameter of 7 mm. The reactions were read after 6 and 24 h Growth was evaluated on MacConkey agar (Merck), incubation in a water bath at 37 "C. Simmons citrate agar (Oxoid) and on litmus lactose agar All test systems were controlled by bacterial strains that are according to the method of Drigalski (Merck). Oxidation known to give either positive or negative reactions under the and fermentation of glucose were examined by using the given conditions. Most of the tests were repeated at least medium described by Pickett (19) and according to Hugh & twice. The following bacterial strains were used as controls : Leifson (9). Inoculated tubes were read daily for up to 7 d. Enterobacter cloacae Hinz 401-92, E. coli Hinz 17-94, Aesculin hydrolysis, phenylalanine deaminase, utilization of Riemerella anatipestger Hinz 175-94 and strain Hinz 176-94 malonate as carbon source, lecithinase and DNase were and Ornithobacterium rhinotracheale LMG 9086 and LMG tested by means of aesculin broth (Merck), malonate 10958. phenylalanine broth according to the method of Shaw & The API 20NE, API ZYM, rapid ID 32 A and API ID 32 E Clarke (Merck), lecithinase agar (Oxoid) and DNase test (for P. europaea strains) and API ZYM and rapid ID 32 A agar (Merck) with and without 2% inactivated filter- (for T. equigenitalis strains) microtest systems were used

International Journal of Systematic Bacteriology 48 433 P. Vandamme and others according to the recommendations of the manufacturer ition in the dendrogram (Fig. 1). Fatty acid cluster I (bioM6rieux). was somewhat heterogeneous and comprised 13 strains (see Fig. 1). Cluster I1 was more homogeneous and RESULTS comprised 11 strains. The mean fatty acid methyl ester composition of cluster I and I1 strains and of T. Fatty acid methyl ester composition equigenitalis LMG 6222T is given in Table 2. Numerical analysis of the whole-cell fatty acid components revealed two clusters of pigeon isolates while PAGE of whole-cell proteins T. equigenitalis LMG 6222T occupied a distinct pos- Duplicate protein extracts of several strains were prepared to check the reproducibility of the growth conditions and the preparation of the extracts. The correlation level between duplicate protein patterns was r > 0.93.

Taylorella equigenitalis 6222 Numerical analysis of the protein profiles of the pigeon i5726 isolates and T. equigenitalis clearly distinguished the 15725 latter from the pigeon isolates. However, although all 15728 - of the pigeon isolates had similar overall whole-cell 15934 protein patterns, several sub-groups could be de- 16152 lineated. Strains belonging to fatty acid cluster I were Cluster I separated into three protein electrophoretic sub- groups and one strain with a separate position (LMG 10986). Sub-group 1 comprised strains LMG 10982T, LMG 12980, LMG 12988, LMG 15727 and LMG 15943; sub-group 2 comprised strains LMG 15725, LMG 15726, LMG 15728, LMG 15934 and LMG 16 152 ;and sub-group 3 comprised strains LMG 10983 15731 - 15729 and LMG 1 1609. Again, strains belonging to fatty acid 15944 I1 - cluster I1 appeared more homogeneous with most of Cluster II 15942 the strains forming a single protein electrophoretic 15936 2 - sub-group (sub-group 4) : LMG 1298 1, LMG 12982, LMG 12985, LMG 12991, LMG 15936, LMG 15942 and LMG 15944. In addition, the following four strains occupied separate positions : LMG 12990, LMG 15729, LMG 15730 and LMG 15731. Fig. 2 illustrates the whole-cell protein profiles of Figrn1. Dendrogram derived from the unweighted pair group mean linkage of Euclidian distances between whole-organism strains representing the protein electrophoretic sub- fatty acid patterns of all strains examined. Roman numerals are groups, of T. equigenitalis strains, and of some of the cluster numbers as discussed in the text. strains with separate positions.

Table 2. Fatty acid composition of the strains studied ...... ,...... Those fatty acids for which the mean amount for all taxa was less than 1 % are not given. Therefore, the percentages for each group do not total 100 YO.Mean percentages and standard deviations are given; TR, trace amount (less than 1 YO);m, not detected.

12:O 14:O 16:O 18:O 16:107c 16:105c 16:0(3-OH) 19:O 10-methyl Summed feature :

Cluster 1(13 strains) 4.4f0.7 9.7f 1.5 15.7k2.9 TR 21.7f3.5 5.2f2-6 1.2f0.3 TR TR 12.7f2.1 27.8f9.4 Cluster I1 (1 1 strains) 4.4 f 1.0 10.8 k 2.5 25.2k 2.8 TR 41.4 f2.7 TR 1.2f0.2 ND 2.7f1.6 12.4f1.1 1.5f0.9 T. equigenitalis - TR 36.8 6.8 TR - TR 2.2 ND 10.7 41.6 LMG 6222T * Summed feature 1 comprises 14: ICO~C,14: lm5t, or both. t Summed feature 3 comprises 14:0(3-OH), 16: 1 is0 I, an unidentified fatty acid with equivalent chain-length value of 10.928, or 12: 0 ALDE, or any combination of these fatty acids. $Summed feature 7 comprises 18: lo7c, 18: lo9t or 18: 1~012t,or any combination of these fatty acids.

434 In ternational Journal of Systernatic Bacteriology 48 Pelistega europaea gen. nov., sp. nov.

Van de Peer et al. (28). This showed that strain LMG 10982T belongs to the beta-subclass of the Proteo- bacteria. Next, a tree was constructed from all the species belonging to the latter subclass, i.e. 183 in total. Strain LMG 10982Tended up in a cluster of 10 species belonging to the genera Bordetella, Alcaligenes and Taylorella. The sister group of the latter cluster contained 16 species, half of which belonged to the genera Leptothrix and Burkholderia. The final tree shown in Fig. 3 shows the topology of the cluster containing strain LMG 10982T,using L. mobilis as an outgroup. Using other species belonging to the Burk- holderia-Leptothrix cluster as an outgroup resulted in the same tree topology in 15 out of 16 cases. T. equigenitalis proved to be the organism most closely related to strain LMG 10982Tfor which a 16s rRNA sequence is known. Sequence dissimilarity values towards its closest neigh- bours were : 5.196 YO towards T. equigenitalis (accession no. X68645), 5.620 % towards Alcaligenes sp. 05-36 (sequence accession no. X86580) and 6.723 % towards A. xylosoxidans (accession no. M22509); all other dissimilarity values were above 7% (data not shown).

Phenotypic tests The analyses were performed on 19 Pelistega strains (not on strains LMG 15934, LMG 15936 and LMG 15942-LMG 15944). The rapid ID 32 A microtest system was performed on all strains. Fig. 2. Whole-cell protein profiles of a representative selection Mostly, when tests were present in multiple API of strains examined. The molecular mass markers used microtest systems, or when tests were performed in a (indicated in the left- and right-hand lanes) are indicated from conventional way and by means of API microtest bottom totop: lysozyme (14500 Da), trypsin inhibitor (20100 Da), trypsinogen (24000 Da), carbonic anhydrase (29000 Da), systems, identical results were obtained. However, glyceraldehyde-3-phosphate dehydrogenase (36000 Da), egg more strains assimilated citrate in the API system albumin (45000 Da) and bovine albumin (66000 Da). The Roman compared to on Simmons citrate agar, and arginine numbers I and II refer to clusters of strains as delineated by dihydrolase activity was not detected in the API fatty acid analysis; the numbers 1-4, and the letter 'sf refer to systems, while this was detectable in 14 out of 19 protein electrophoretic sub-groups as discussed in the text. strains by the conventional procedure. The results of the biochemical analyses of P. europaea DNA base compositions strains are listed below in the Discussion. The DNA G + C content of representative strains per A more restricted set of biochemical tests and API electrophoretic cluster was determined and is listed in microtest systems was performed on all four T. Table 3. equigenitalis strains. Little growth occurred on plain blood agar after 48-72 h incubation at 37 "C under DNA-DNA hybridization results microaerobic conditions. Growth improved when chocolate agar was used, when conventional media The DNA-DNA hybridization results are listed in were supplemented by Oxoid SR 46 Filder extract as Table 3. High hybridization levels (> 82%) were recommended by the manufacturer, or when Brain measured among strains belonging to the same protein Heart Infusion agar (Difco) was supplemented by electrophoretic sub-group ; strains belonging to differ- 15 pg haemin chloride ml-l (equine ; Sigma) or 100 pg ent protein electrophoretic sub-groups or strains with NAD ml-l (Serva). An additional significant improve- separate positions exhibited clearly lower or insig- ment of growth was obtained on Brain Heart Infusion nificant hybridization values in the range of 10-46 YO. agar supplemented with both haemin and NAD. Phylogenetic analysis of the 165 rRNA sequence In general, T. equigenitalis strains did not grow aerobically or anaerobically at 37 "C. In addition, no First, a tree was constructed from the sequences of growth occurred microaerobically at 24 "C. There was LMG 10982T and several representatives of each of 18 no pigmentation and no growth on MacConkey agar large bacterial divisions and subdivisions as listed by or litmus lactose agar. Catalase and oxidase activities

International Journal of Systematic Bacteriology 48 435 P. Vandamme and others

Table 3. DNA-DNA binding values and G + C contents of P. europaea strains

Strain LMG LMG LMG LMG LMG G+C content Fatty acid 10982T 15725 10986 15944 15731 (mol Oh) clusterlprotein electrophoretic sub-group

~ ~~ LMG 10982T 100 43 LMG 15943 100 42 LMG 15725 36 100 43 LMG 15934 95 42 LMG 10986 25 28 100 43 LMG 11609 10 46 42 LMG 15944 34 17 33 100 43 LMG 12985 14 82 42 LMG 15731 15 28 18 41 100 42

Distance 0.05 There was no acidification of mannose or raffinose. No I glutamic acid decarboxylase activity was detected. Arylamidase activity of the following compounds was -00 Bordetella bronchiseptica (U04948) present : arginine, leucyl glycine, phenylalanine, leu- Bordetella pertussis (U04950) cine, tyrosine, alanine, glycine, glutamyl glutamic acid 3 56C Bordetella holmesii (U04820) and serine. However, if the inoculum used had a 99 - - Bordetella avium (U04947) density of only McFarland 4 (compared to McFarland 6), arylamidase activity of leucyl glycine, phenyl- 100 Alcaligenes xylosoxidans (M22 509) alanine, tyrosine, glutamyl glutamic acid and serine

~ Alcaligenes faecalis (M22508) 72 was not detected. Arylamidase activity of proline, pyroglutamic acid and histidine was absent.

DISCUSSION Taxonomic structure of the group of pigeon isolates ...... , ...... Fig. 3. Neighbour-joining tree of a cluster of bacteria A collection of 24 pigeon isolates remained un- belonging to the beta-subclass of the Proteobacteria and identified after preliminary examination by conven- comprising strain P. europaea LMG 10982T. Species names are tional biochemical tests. An extensive biochemical followed by the accession number of the 165 rRNA sequence. The outgroup species L. mobilis belongs to the most closely characterization by using a considerable number of related cluster. The evolutionary distance between two species conventional tests and several API microtest systems is obtained by summing the lengths of the connecting branches further substantiated the high similarity among all of along the horizontal axis, using the scale on top. Bootstrap the strains examined, but was unable to clarify the values (%) are indicated at each node. taxonomic position of the isolates. Cellular fatty acid analysis demonstrated a marked subdivision of the pigeon isolates into two main were present but urease activity was not detectable. clusters (Fig. 1 ; Table 2). Cluster I and I1 strains were Reduction of nitrate, indole production, arginine easily differentiated by comparing the percentages of dihydrolase, production of acetylmethylcarbinol and 16:0, 16:lm7c, 16:lco5c or summed feature 7. The aesculin hydrolysis were all absent. Gelatinase, hya- strains grouping within the fatty acid clusters, how- luronidase and chondroitinsulfatase activities were not ever, were further subdivided into four sub-groups and detected. There was no oxidation or fermentation of several strains with unique patterns by means of one- glucose and no acid production from glucose in dimensional whole-organism protein electrophoresis buffered saline. Alkaline and acid phosphatase, (Fig. 2). The DNA G+C content of representative naphthol-AS-Bl-phosphohydrolaseand leucine aryl- strains of these sub-groups was determined and all amidase activities were present. The activities of values were between 42 and 43 mol% (Table 3). esterase C4 and ester lipase C8 were strain-dependent. Results obtained by whole-cell protein electrophoresis Lipase C 14, valine and cystine arylamidase, trypsin, were a good indicator of the level of DNA-DNA chymotrypsin, a- and P-galactosidase, P-galactosidase- hybridization, as only within the protein electr- 6-phosphate, P-glucuronidase, a- and P-glucosidase, ophoretic sub-groups were high binding levels detected N-acetyl-P-glucosaminidase,a-arabinosidase, a-man- (> 82 YO);all other values were low or not significant nosidase and a-fucosidase activities were not detected. (Table 3).

436 International Journal of Systematic Bacteriology 48 Pelistega europaea gen. nov., sp. nov.

In taxonomic practice, the DNA-DNA hybridization isolates. T. equigenitalis strains further differed from level within a bacterial species is mostly above 50-70 YO the pigeon isolates by the absence of urease and (24). This level is determined by the phenotypic arginine dihydrolase activity (present in 14 of 19 pigeon consistency of the taxa studied, as species should be isolates tested) and alkali production from glucose identifiable by phenotypic characteristics (30). Re- (present in all pigeon isolates). cently, the term genomovar was introduced for gen- Considering these various data we feel it is justified to omic species which cannot be distinguished by pheno- include the pigeon isolates in a new genus, for which typic characteristics (23). The group of pigeon isolates we propose the name Pelistega, as Pelistega europaea. is clearly composed of several sub-groups which are Characteristics useful to differentiate P. europaea from genotypically distinct enough to warrant their class- other species that may be encountered in pigeons are ification as different species. The overall similarities in shown in Table 4. Data of reference taxa are primarily biochemical characteristics, whole-cell protein and from our own research groups. fatty acid components, and DNA base composition, and the occasional significant DNA-DNA hybrid- Description of Pelistega gen. nov. ization values between representatives of different sub- groups indicate that all of these sub-groups belong to Pelistega (pe.li'ste.ga. Gr. n. peleia pigeon; Gr. fem. n. a single phylogenetic lineage. In the absence of stege house, stay, residence; M.L. fem. n. Pelistega differential phenotypic characteristics between the refers to the bacteria living in pigeons). genomovars as delineated by DNA-DNA hybrid- Cells are Gram-negative, non-spore-forming, encap- ization studies and whole-cell protein analysis, we sulated, microaerophilic and non-motile with variable preferred to include all of these isolates into a single morphological forms. The major fatty acid com- species, and selected strain LMG 10982 as the type ponents of all strains examined are 12 :0, 14:0, 16 :0, strain. The four protein electrophoretic sub-groups 16: lw7c, 16: lw5c, 16:0(3-OH) and summed features clearly represent four different genomovars. Strain 3 and 7. The DNA base composition is about LMG 10982T is the reference strain of the first 42-43 mol YOG + C and the type species is Pelistega genomovar. We propose strains LMG 15725, LMG europaea sp. nov. Four genomovars and several strains 11609 and LMG 12985 as reference strains for gen- with separate taxonomic positions are delineated, all omovars 2, 3 and 4, respectively. Strains with unique of which are presently included in a single nomen- whole-organism protein patterns (see Tables 1 and 3; species, P. europaea. Strain LMG 10982T,the reference Fig. 2) were not classified into different genomovars strain of genomovar 1, was chosen as the type strain of until the DNA-DNA hybridization levels between the species; strains LMG 15725, LMG 11609 and these strains and the four designated genomovars are LMG 12985 are reference strains for genomovars 2, 3 fully determined. and 4, respectively.

Phylogenetic position of strain LMG 10982T Description of Pelistega europaea sp. nov. Comparison of the virtually complete 16s rDNA Pelistega europaea (eu.ro.pae'a. L. adj. of Europe, sequence of strain LMG 10982Twith representatives because this first collection of strains was isolated in of various bacterial lineages revealed that this or- European countries). ganism belongs to the beta-subclass of the Proteo- Cells are Gram-negative, non-spore-forming, cap- bacteria as defined by Woese et al. (32). In this beta- sulated and rod-shaped with variable morphological subclass, T. equigenitalis was its nearest neighbour forms. Non-motile. In 16-24-h-old cultures on blood with a percentage 16s rDNA similarity of about 94-8YO agar, the mean cell size is 0.2-0.4 pm in width and (Fig. 3). Like the pigeon isolates, T. equigenitalis 1-2 pm in length. Strains produce convex, circular and strains are microaerobic, Gram-negative, non-motile greyish-white to yellowish colonies with entire edge rods which do not acidify carbohydrates and exhibit and smooth surface on blood agar. The strains do not catalase, oxidase and acid and alkaline phosphatase require special growth factors and grow on conven- activity (22). T. equigenitalis strains were subsequently tional media. In standard I1 nutrient broth, growth included as references in the various analyses but were appears as moderate turbidity and pellicle near the always clearly differentiated from all of the pigeon surface of the broth after 24-48 h incubation. In most isolates (Figs 1 and 2; Table 2) and have a different cases, a small button-like sediment can be seen. No DNA G+ C content (about 36-37 % for the former growth on litmus lactose agar at the primary isolation and 42-43 for the latter). The most salient differences from the infected tissues. However, after some pass- between the pigeon isolates and T. equigenitalis strains ages on artificial media some strains show a poor were, however, the fastidious growth requirements of growth. No growth on MacConkey agar. Micro- the latter, while the pigeon isolates grew abundantly aerobic growth at 37 and 42 OC, not at 24 "C. Strains when incubated under microaerobic conditions. In grew aerobically (although most strains less abun- contrast with data reported by Sugimoto et al. (22), we dant), but not anaerobically. All strains produce found a remarkable stimulation of growth of 7'. catalase and oxidase activity ; urease and arginine equigenitalis strains by the addition of growth factors, dihydrolase activity is strain-dependent. No reduction which again differentiates these strains from the pigeon of nitrate, no denitrification, no production of acetyl-

International Journal of Systematic Bacteriology 48 437 P. Vandamme and others

Table 4. Differential characteristics between P. europaea and other bacteria from pigeons ...... + , Characteristic present; - , characteristic absent; v, strain-dependent reaction; w, weak reaction; ND, not determined.

Character Pelistega Riemerella Ornithobacterium Chryseobacterium Pasteurella europaea anatipestifer rhinotracheale meningosepticum species

Pigmentation - V* - Growth on MacConkey agar - - + - Aerobic growth on blood agar V* W + + at 37 "C Catalase activity + + + Urease activity V - - Nitrate reduction - - + Indole production - V* V Gelatinase activity + + - Chondroitinsulfatase activity - ND V Hyaluronidase activity - ND V Aesculin hydrolysis? - + - P-Galactosidase activity? - + V* Acid production from glucose + V* + Utilization of carbon sources - + - for growth? G + C content (mol %) 29-35 37-39 36-37 40-45 Host spectrum Poultry, Turkeys, chickens, Humans, Mammals, wild birds wild birds birds birds * Over 80 % of the strains contain this feature. ?As determined using the API 20 NE microtest system. methylcarbinol, indole or methyl red, no aesculin dependent. Lipase C14, valine, leucyl glycine, phenyl- hydrolysis, gelatinase (except strain LMG 11609), P- alanine, histidine, glutamyl glutamic acid, serine and galactosidase,DNase, chondroitinsulfatase, hyaluron- cystine arylamidase, trypsin, chymotrypsin, a-galacto- idase, lysine decarboxylase, ornithine decarboxylase, sidase, a-maltosidase, P-glucuronidase, a- and P-glu- lecithinase or phenylalanine deaminase activity. Glu- cosidase, N-acetyl-P-glucosaminidase, a-mannosidase cose is oxidized with the production of alkali using the and a-fucosidase activities are not detected. Proline, medium described by Pickett (19) and Hugh & Leifson pyroglutamic acid and tyrosine arylamidase activities (9); no fermentation of, or acid production from, are strain-dependent. glucose. Malonate is not used as a carbon source. Two major clusters of strains are delineated by means Strain-dependent utilization of citrate on Simmons of cellular fatty acid analysis. The major fatty acid citrate agar. components of all strains examined are 12:0, 14:0, None of the strains produce acid from D-glucose, D- 16:0, 16: 1co7c, 16: lco5c, 16:0(3-OH) and summed fructose, sucrose, lactose, ma1to se, D-galacto se, galac- features 3 and 7. However, the percentages of these turonate, D-mannose, rhamnose, cellobiose, palat- fatty acids vary strongly among the clusters. inose, dextrin, N-acetyl-D-glucosamine, lactulose, Strains have been isolated on common non-selective L-sorbose, adonitol, D-mannitol, L-arabinose, D- and blood agar media under microaerobic (candle jar) L-arabitol, salicin, D-sorbitol, trehalose, D-xylose, conditions mainly from samples of lungs, air sac dulcitol, inositol or myo-inositol. exudate and trachea mucosa and less frequently from There is no assimilation of D-glucose, maltose, other organs such as liver and spleen. They can also be D-mannose, N-acetyl-D-glucosamine, D-mannitol, isolated from swabs taken from the palatine cleft or L-arabinose, D-gluconate, caprate, adipate or phenyl trachea of living acutely diseased pigeons. At present acetate; most strains (18 out of 19 examined) assimi- the knowledge with regard to pathogenicity is in- lated L-malate. complete. The present clinical observations suggest that they are pathogenic and involved especially in Alkaline and acid phosphatase, esterase C4, ester pathogenesis of respiratory diseases in pigeons. How- lipase C8 and leucine, arginine, alanine, glycine and L- ever, the role of co-factors and the interaction with aspartic acid arylamidase activities are present. Ac- other agents are uncertain. Bacteria associated with tivities of naphthol-AS-B1-phosphohydrolaseand li- similar clinical signs in pigeons were described by pase (as present in the API ID 32E system) are strain- Andreasen & Sandhu (1). All morphological and

438 International Journal of Systematic Bacteriology 48 Pelistega europaea gen. nov., sp. nov. biochemical characteristics reported by the latter Fibrolysin und Phosphatase bei Staphylokokken. Zentralbl authors corresponded with our findings. Bakteriol Parasitenkd Infektionskr Hyg Abt I Orig Reihe A 178, 381-385. The DNA base composition is 42-43 mol% G+C. The type strain is LMG 10982, which was isolated 12. Kersters, K. & De Ley, J. (1971). Enzymatic tests with resting from a pigeon in Belgium. Its G+C content is cells and cell-free extracts. Methods Microbiol5A, 33-52. 43 mol %. Strain LMG 10982T belongs to fatty acid 13. Kersters, K., Hinz, K.-H., Hertle, A., Segers, P., Lievens, A,, cluster I and protein electrophoretic sub-group 1, Siegmann, 0. & De Ley, 1. (1984). Bordetella avium sp. nov., isolated from the respiratory tracts of turkeys and other which comprise 13 and 5 isolates, respectively. It is not birds. Int J Syst Bacteriol34, 56-70. unlikely that the name P. europaea will be restricted in the future to organisms belonging to fatty acid cluster 14. Kilian, M. (1976). A taxonomic study of the genus Haemo- I, or even to protein electrophoretic sub-group 1, upon philus. J Gen Microbiol93, 9-62. discovery of differential diagnostic features. 15. Lautrop, H. (1960). Laboratory diagnosis of whooping cough or Bordetella infections. Bull WHO 23, 15-3 1. All P. europaea strains have been deposited in the 16 Marmur, 1. & Doty, P. (1962). Determination of the base BCCM/LMG Culture Collection. composition of deoxyribonucleic acid from its thermal denaturation temperature. J Mol Biol5, 109-1 18. ACKNOWLEDGEMENTS 17 Msller, V. (1955). Simplified tests for some amino acid P.V., Y.V.D.P., R.D.W. and K.K. are indebted to the decarboxylases and for the arginine dihydrolase system. Fund for Scientific Research, Flanders (Belgium), for Acta Pathol Microbiol Scand 36, 158-172. positions as a postdoctoral fellow (P. V. and Y. V. D. P.) and 18 MiAler, 0. (1951). A new method for staining bacterial research and personnel grants (R. D. W. and K. K.). Our capsules. Acta Pathol Microbiol Scand 28, 127-1 3 1. research was also supported by the Prime Minister’s Services, 19. Pickett, M. 1. (1980). Nonfermentative Gram Negative Bacilli. Federal Office for Scientific, Technical and Cultural Affairs, A Syllabus for Detection and IdentiJication. Los Angeles: Belgium. We thank Rita Leise for excellent technical Scientific Development. assistance. We are grateful to E. Falsen (CCUG Culture Collection, Department of Clinical Bacteriology, University 20. Pot, B., Vandamme, P. & Kersters, K. (1994). Analysis of of Goteborg, Goteborg, Sweden) for supplying us with electrophoretic whole-organism protein fingerprints. In reference strains of T. equigenitalis. Modern Microbial Methods. Chemical Methods in Pro- karyotic Systematics, pp. 493-521. Edited by M. Good- fellow & A. G. O’Donnell. Chichester: Wiley. REFERENCES 21. Smith, R. F. & Willet, N. P. (1968). Rapid plate method for 1. Andreasen, J. R. & Sandhu, T. (1993). Pasteurella anati- screening hyaluronidase and chondroitin sulfatase-produc- pestifer-like bacteria associated with respiratory disease in ing microorganisms. Appl MicrobioEl6, 1434-1436. pigeons. Avian Dis 37, 908-91 1. 22. Sugimoto, C., Isayama, Y., Sakazaki, R. & Kuramochi, 5. 2. Birnboim, H. C. & Doly, 1. (1979). A rapid alkaline extraction (1983). Transfer of Haemophilus equigenitalis Taylor et al. procedure for screening recombinant plasmid DNA. 1978 to the genus Taylorella gen. nov. as Taylorella Nucleic Acids Res 7, 15 13-1 523. equigenitalis comb. nov. Curr Microbiol9, 155-162. 3. Cowan, S.T. (1974). Cowan and Steel’s Manual for the 23. Ursing, 1. B., Rossello-Mora, R. A., Garcia-Valdes, E. & Lalucat, Identijication of Medical Bacteria, 2nd edn. Cambridge : 1. (1995). Taxonomic note: a pragmatic approach to the Cambridge University Press. nomenclature of phenotypically similar genomic groups. Int 4 De Ley, 1. (1970). Reexamination of the association between J Syst Bacteriol45, 604. melting point, buoyant density, and chemical base com- 24. Vandamme, P., Pot, B., Gillis, M., De Vos, P., Kersters, K. & position of deoxyribonucleic acid. J Bacteriol 101,738-754. Swings, 1. (1996). Polyphasic taxonomy, a consensus ap- 5 De Ley, 1.. Cattoir, H. & Reynaerts, A. (1970). The quantitative proach to bacterial classification. Microbiol Rev 60, measurement of DNA hybridization from renaturation 407438. rates. Eur J Biochem 12, 133-142. 25. Vandamme, P., Vancanneyt, M., Pot, B. & 10 other authors 6 Felsenstein, 1. (1985). Confidence limits on phylogenies : an (1992). Polyphasic taxonomic study of the emended genus approach using the bootstrap. Evolution 39, 783-79 1. Arcobacter with Arcobacter butzleri comb. nov. and Arco- 7. Frazier, W. C. (1926). A method for the detection of changes bacter skirrowii sp. nov., an aerotolerant bacterium isolated in gelatin due to bacteria. J Infect Dis 39, 302-309. from veterinary specimens. Int J Syst Bacteriol42,344-356. 8. Holton, T. A. & Graham, M. W. (1991). A simple and efficient 26. Van de Peer, Y. & De Wachter, R. (1994). TREECON for method for direct cloning of PCR products using ddT-tailed Windows: a software package for the construction and vectors. Nucleic Acids Res 19, 1156. drawing of evolutionary trees for the Microsoft Windows environment. Comput Appl Biosci 10, 569-570. 9. Hugh, R. & Leifson, E. (1953). The taxonomic significance of fermentative versus oxidative metabolism of carbohydrates 27. Van de Peer, Y., Jansen, J., De Rijk, P. & De Wachter, R. (1997). by various Gram-negative bacteria. J Bacteriol66, 24-26. Database on the structure of small ribosomal subunit RNA sequences. Nucleic Acids Res 25, 11 1-1 16. 10. Jukes, T. H. & Cantor, C. R. (1969). Evolution of protein molecules. In Mammalian Protein Metabolism, vol. 3, pp. 28. Van de Peer, Y., Neefs, J.-M., De Rijk, P., De Vos, P. & De 21-132. Edited by H.N. Munro. New York: Academic Wachter, R. (1994). About the order of divergence of the Press. main bacterial taxa during evolution. Syst Appl Microbiol 11. Kaffka, A. (1957). Erfahrungen mit einfachen Unter- 17, 32-38. suchungsmethoden zum Nachweis von Hyaluronidase, 29. Walc-Pokrzywnicki, 2. (1973). Flagella staining. Zentralbl

International Journal of Systematic Bacteriology 48 439 P. Vandamme and others

Bakteriol Parasitenkd Znfektionskr Hyg Abt 1 Orig Reihe A Structure of the 16s ribosomal RNA of the thermophilic 223,240-243. cyanobacterium Chlorogloeopsis HTF (' Mastigocladus lam- 30. Wayne, L. G., Brenner, D. J., Colwell, R. R. & 9 other authors inosus HTF') strain PCC75 18, and phylogenetic analysis. (1987). International Committee on Systematic Bacteri- FEBS Lett 317,96-100. ology. Report of the ad hoc committee on reconciliation of 32. Woese, C. R., Kandler, 0. & Wheelis, M. L. (1990). Towards a approaches to bacterial systematics. Znt J Syst Bacteriol37, natural system of organisms: proposal for the domains 463-464. Archaea, Bacteria, and Eucarya. Proc Nut1 Acad Sci USA 31. Wilmotte, A., Van der Auwera, G. & De Wachter, R. (1993). 87,4576-4579.

440 international Journal of Systematic Bacteriology 48