Bisgaard's taxa

Taxon 1

All 11 isolates of this taxon originally reported by Bisgaard (20)* were demonstrated in mixed flora from the intestinal tract of apparently normal ducks. These isolates were later classified with Pasteurella sensu stricto based upon DNA:DNA hybridizations (27) and named Pasteurella anatis by Mutters et al. (Int. J. Syst. Bact. 1985, 35, 309-322). P. anatis has subsequently also been obtained from ducks suffering from respiratory disease (50).

Crossed immunoelectrophoresis applied to representative strains from 11 different Pasteurella spp. showed that P. anatis formed a separate cluster branching deeply with P. langaa showing less than 70% similarity with the core group of genus Pasteurella (Schmid et al. Zbl. Bakt. 1991, 275, 16-27). A new genus, Gallibacterium gen. nov., suggested to include the avian [P.] haemolytica – [A.] salpingitidis – [P.] anatis complex was reported by Christensen et al. (119), since these taxa formed a monophyletic unit with 16S rRNA similarities above 95%. Strains originally reported as [P.]anatis are now classified as G. anatis biovar anatis. Korczak et al. (Int. J. Syst. Evol. Microbiol. 2004, 54, 1393-1399) demonstrated that rpoB gene sequence analysis was useful for separating Gallibacterium spp. from other taxa of Pasteurellaceae

Taxon 2 and 3 complex

Organisms associated with salpingitis and peritonitis in ducks were first reported as atypical

Actinobacillus lignieresii by Bisgaard (4). Comparative investigations of avian Pasteurellaceae including additional isolates of atypical Actinobacillus lignieresii obtained from ducks, geese and pigeons subsequently resulted in classification of these organisms as taxa 2 and 3 based upon differences in production of acid from (+)-L-arabinose and dulcitol (20).

A total of 35 strains isolated from the respiratory tract, liver, heart and spleen of pigeons and different species of Psittaciformes were reported by Beichel (1986). Nineteen of these were reinvestigated by Bisgaard et al. (90) and biotyped as previously described (50). The host spectrum was extended to include galliform birds like partridges and pheasants (50). The importance of taxon 2 and taxon 3 in salpingitis in web-footed birds was subsequently confirmed by Bisgaard (61).

On the basis of DNA-DNA hybridizations these taxa were shown to form a large distinct group which seems to represent a new genus with several species within the family Pasteurellaceae Pohl 1981 (27). However, full matrix DNA-DNA hybridizations were not carried out.

De Ley et al. (36) performed hybridizations between labelled rRNA from seven representative members of the family Pasteurellaceae to 53 strains of Pasteurellaceae and showed that strain HIM 730-3 (F420) of taxon 2/3 (according to Bisgaard et al. (51) and not taxon 2 as listed by De Ley et al ., (36) and strain NCTC 11412 of taxon 3 biovar 2 clustered with different nodes of the rRNA branches outlined, indicating major genetic diversity within taxon 2 and 3.

Different polyamine patterns were obtained with a duck isolate of taxon 2 biovar 1 (F150 T ) and a parakeet isolate of taxon 3 biovar 1 (F450 T ) according to Busse et al. (Int. J. Syst. Bact. 1997, 47, 698-708). Both isolates had profiles different from the newly established genera Gallibacterium (119) and

Avibacterium (138) of avian origin. Comparison of phena defined by protein profiling with species/groups previously established by DNA-DNA hybridization, carbohydrate profiling and biovar typing showed that the best correlation existed between DNA-DNA hybridization and biovar typing. A correlation between results obtained from DNA-DNA hybridizations and protein profiling was not observed. Protein profiling, however, indicated a connection between protein profiles and hosts of isolation (51). Host related bacterial lineages of the taxon 2 and 3 complex have subsequently been demonstrated by amplified fragment length polymorphism (AFLP) typing (158).

Phylogenetic analysis by 16S rRNA gene sequence comparison has shown that taxon 2 biovar 1 (NCTC 11414) and taxon 3 biovar 2 (CCUG 15565 T, wrongly referred to as CCUG 15563 in the article) cluster together forming subcluster 3D of Dewhirst together with strains subsequently classified as Gallibacterium and [ Pasteurella ] langaaensis (Dewhirst et al ., Zbl. Bakt. 1993, 279, 35-44). In a subsequent investigation including avian taxa, Bisgaard taxon 2 and taxon 3 formed a separate entity within the avian cluster (127). A single isolate from septicaemia in a budgerigar classified as trehalose positive taxon 3 biovar 1 only showed 94.5% 16S rRNA gene sequence similarity to biovar 2 of taxon 3 underlining the heterogeneity of this complex group.

The accumulated evidence outlined above underlines the uncertain taxonomic position of the taxon 2 and taxon 3 complex of avian origin and the lack of unambiguous diagnostic possibilities associated with these organisms which makes interpretation of studies on these organisms difficult and prevents progress in understanding of their epidemiology and pathogenesis.

For the same reason a subset of 23 strains representing existing biovars characterized by AFLP (158) were investigated by phylogenetic analysis of the 16S rRNA, rpoB , infB and recN gene sequences. Moreover, the recN gene sequences were used for estimation of whole genome similarity to allow comparison with previously published DNA-DNA hybridizations. Selected strains were additionally characterized by polyamine profiling. Clusters outlined by

Bojesen et al. (158) and shown not to belong to Gallibacterium sensu stricto were excluded from the investigation and these taxa will be published separately. These studies allowed the identification of five groups with unique properties allowing the suggestion of five new species of Gallibacterium: G. melopsittaci, G. trehalae, G. columbinum, G. salpingitigis and a not yet named species (181). Isolates of taxon 2 and 3 sharing the properties of Volucribacter are under investigation.

Taxon 4

([Pasteurella] langaaensis)
Organisms classified as taxon 4 were originally obtained in mixed culture from the respiratory tract of apparently healthy broiler breeders (Bisgaard, 1982). These strains shared the cultural and biochemical characters of isolates previously designated Pasteurella sp. by Clark and Godfrey (1960). Unfortunately, the original isolates of Clark and Godfrey were no longer available for comparison.
Differences in acid production from lactose, maltose, trehalose and dextrin separated taxon 4 from [P.] gallinarum, while differences in acid production from glycerol, D (+) xylose, lactose and trehalose separated taxon 4 from taxon 1 (Bisgaard, 1982).

Based upon DNA homology, Mutters et al. (1985) reclassified the genus Pasteurella Trevisan 1887 and proposed five new species. The two strains of taxon 4 included, demonstrated 100% DNA:DNA binding and were named Pasteurella langaa, while only 30, 25, 20, and 14% binding was observed with P. dagmatis, P. multocida ssp. septica, P. multocida ssp. multocida and P. canis, respectively. The highest % DNA:DNA binding observed with other species of Pasteurella was with [P.] volantium (57-49%) and Pasteurella anatis (Bisgaard taxon 1) (51-36%), while only 20% binding was observed with [Pasteurella] gallinarum.

Schmid et al. (1991) used crossed immunoelectrophoresis to compare results from DNA:DNA hybridizations reported by Mutters et al. (1985). These investigations showed that [P.] langaa formed a separate cluster branching deeply with [P.] anatis, showing less than 70% similarity with the core group of the genus Pasteurella. Interestingly, [P.] gallinarum, [P.] avium biovar 1, [P.] sp. A, and [P.] volantium formed a separate cluster, distantly related with Pasteurella sensu stricto. These taxa were subsequently classified as a new genus, Avibacterium (Blackall et al., 2005).

The polyamine pattern of [P.] langaa NCTC 11411T mainly included 1,3-diaminopropane (DAP) (95.5%) in addition to small amounts of putrescine (PUT) (2.7%) and spermidine (SPD) (1.9%) (Busse et al., 1997).
In the 16S rRNA gene sequence based phylogenetic tree for the family Pasteurellaceae, P. langaa clustered distantly with Bisgaard taxon 2 and taxon 3, in addition to [P.] anatis and A. salpingitidis (Dewhirst et al., 1993).

In the neighbour-joining phylogenetic tree based on 16S rRNA gene sequences of members of the family Pasteurellaceae, [Pasteurella] langaa clustered distantly with A. rossii with a bootstrap value of only 37%, while [P.] langaa clustered with [P.] caballi in the phylogenetic tree based on the rpoB gene. Finally, [P.] langaa appeared as a singleton in the phylogenetic tree based on the sodA gene (Korczak and Kuhnert, 2008).

Both crossed immunoelectrophoresis and 16S rRNA-, rpoB-, and sodA gene analysis clearly suggest that [P.] langaa should be excluded from the genus Pasteurella.

Additional investigations, including whole genome sequencing and deduction of more conserved gene sequences, in addition to randomly selected conserved protein sequences for comparison, are needed before reclassification of [P.] langaa is possible (Christensen and Bisgaard, 2018). At the same time a correct latin name, [P.] langaaensis, should be used.

The G+C content of DNA of [P.] langaaensis ranges from 43.9 to 45.3 mol %, while the molecular weights  range from 1.7 to 1.9 x 109d (Bisgaard et al.,1983). The type strain is NCTC 11411T.

Taxon 5

(Caviibacterium pharyngocola)
In the 1980ies knowledge on the occurrence of Pasteurellaceae in guinea pigs was very limited and mainly inferred from indirect evidence, gained from occasional isolations in various types of clinical material. To improve our knowledge on the natural occurrence of members of Pasteurellaceae in healthy guinea pigs, Bisgaard et al. (1983) examined two colonies of conventional, well-managed guinea pigs.

Four new taxa designated 5 through 8 were demonstrated in both colonies. Differences in 13 characters separated the taxa reported. Strains tentatively designated taxon 5 only differed from two strains previously reported by Stewart and Letscher (1976) in acid production from lactose and trehalose. A phenotypical relationship was also demonstrated between taxon 5 and taxon 6, the SP-group (Necropsobacter rosorum), [P.] aerogenes and Av. gallinarum, respectively.

The G+C content of DNA from taxon 5 varied from 40.0% to 42.1% both of which are within the limits for Pasteurellaceae. The genome mol. wt. was 1.8 x 109d. A DNA:DNA affiliation at 48% DNA binding was observed between taxon 5 and A. hominis P575, while 43% DNA:DNA binding was observed between taxon 5 and A. lignieresii NCTC 4189T and 34% between taxon 5 and Av. gallinarum ATCC 13360. These binding values clearly suggest classification with Pasteurellaceae.

Less than 20% DNA binding was observed between taxon 5 and taxon 6 and 8, respectively, while taxon 5 and taxon 7 demonstrated 27% DNA:DNA binding (Bisgaard et al., 1983). Subsequently, these taxa have been confirmed to belong to Pasteurellaceae (Christensen and Bisgaard, 2008; Dewhirst et al., 1993).

Among 14 strains of taxon 5 subsequently investigated, 12 strains showed identical rpoB gene sequences, while a German isolate only showed 93.5% similarity to the type strain of taxon 5 (7.3T). The two strains deviating in rpoB gene sequence demonstrated 95.5% similarity. Among the deviating strains, BJ 2809.4 showed 95.6% rpoB gene sequence similarity to the type strain of taxon 5, indicating classification with taxon 5. Finally, taxon 5 showed 87% ropB gene sequence similarity to the type strain of P. multocida, which is slightly above the general similarity between genera of Pasteurellaceae of 85% (Christensen et al., 2007)

The 16S rRNA gene sequence comparison demonstrated a unique position of the type strain 7.3T of taxon 5 among other genera of Pasteurellaceae. Strain FD 498 was related at 98.5% 16S rRNA gene sequence similarity to the type strain, in accordance with the intra species general level of 97% similarity (Stackebrandt and Goebel, 1994). BJ 2809.4, however, only demonstrated 95.5% 16S rRNA gene sequence similarity to 7.3T, which is below species level.

The closest related genus to taxon 5 based on 16S rRNA gene sequence comparisons was the type strain of Ursidibacter maritinus with 96.1% similarity, which is slightly above the genus level difference for genera of Pasteurellaceae (Christensen et al., 2007). According to Bisgaard et al. (1983) DNA:DNA bindings between strains 3.3 and 10.3 of taxon 5 demonstrated 98% binding, while the closest related taxon compared was A. hominis with 48% binding.

Based upon pheno- and genotypic results obtained, a new genus with at least one species, Caviibacterium pharyngocola, was proposed (Adhikary et al., 2018). Phylogenetic analysis of 31 randomly selected conserved protein sequences has subsequently confirmed the classification of taxon 5 as a new genus of Pasteurellaceae (Christensen and Bisgaard, 2018).

Taxon 6

(remains to be classified and named)
Most vertebrates investigated so far have allowed isolation and identification of one or more taxa of Pasteurellaceae from the mucosal surfaces of the upper respiratory tract, oropharynx and/or the reproductive tract.

Bisgaard et al. (1983) investigated the occurrence of Pasteurellaceae in two colonies of conventional, well-managed guinea pigs. Detailed phenotypical investigations demonstrated four new taxa, provisionally designated 5 through 8, of which taxon 6 might constitute up to 50% of the flora.

Only differences in urease, phosphatase and trehalose separate taxon 6 from taxon 5, while six and ten characters separate taxon 6 from taxon 8 and 7, respectively. Differences in urease and phosphatase separate taxon 6 from the organisms described by Stewart and Letscher (1976), while only production of acid from D (-) mannitol separates taxon 6 from Bisgaard taxon 10 (Mutters et al., 1989).

Polyamine patterns of two strains of taxon 6 showed tat 1,3-diaminopropane (DAP) made up more than 90%, while only small amounts of putrescine (PUT), cadaverine (CAD), sym-norspermidine (NSPD), spermidine (SPD) and spermine (SPM) were detected (Busse et al., 1997).

Only between 4 and 23% DNA binding were observed between taxon 6 strain 7.5 and strains 3.3 and 10.3 of taxon 5, strain 10.4 of taxon 7, and strain 8.3 of taxon 8, indicating that these four taxa probably belong to different genera of Pasteurellaceae (Bisgaard et al. 1983). The G+C content of DNA from strain 7.5 of taxon 6 was 51.2 mol%, while the genome molecular weight was 2.1 x 109d (Bisgaard et al. 1983). For comparison the G+C content of DNA from strains of the SP group, classified as Necropsobacter rosorum, varied between 53.7-53.2 mol%, while the Stewart Letscher strain P625 only showed 48.3 mol% G+C (Christensen et al., 2011).

According to Dewhirst et al. (1993) taxon 6 clustered deeply with [H.] somnus in the 16S rRNA gene based phylogenetic tree. Subsequently, Christensen et al. (2004) showed that taxon 6 and taxon 10 clustered with [A.] succinogenes in the phylogeny of the family Pasteurellaceae based upon maximum-likelihood analysis of 16S rRNA gene sequences. Subsequently, Christensen et al. (2005) showed that taxon 6 (CCUG 15568) and taxon 10 (CCUG 15572) of Bisgaard made up a monophyletic group with P.a.3 (biovar 17 of [P.] aerogenes), [A.] succinogenes ATCC 55618T and [H.] somnus HS 8025T in the phylogenetic tree based on maximum-liklihood analysis of 16S rRNA gene sequences of taxa of the [P.] aerogenes – [P.] mairii – [A.] rossii complex and representative members of Pasteurellaceae. The bootstrap value for taxon 6, taxon 10, and P.a.3 was as high as 99%, clearly indicating that these organisms might represent a new genus of Pasteurellaceae.

Reinvestigations of six strains from the ear and small intestine of guinea pigs and a single isolate from a muskrat, previously classified as Actinobacillus sp. (Boot et al., 1983), were subsequently shown to belong to taxon 6 (Boot and Bisgaard, 1995). Finally, four strains from pandas have been shown to meet the phenotypical characters of taxon 6 (Christensen and Bisgaard, unpublished data).

Additional taxonomic investigations, including a total of 27 strains of taxon 6 from guinea pigs (n=19), a muskrat (n=1), a laboratory rodent (n=1), pigs (n=3) and pandas (n=4), and 13 strains of taxon 10 from horses (n=12) and a horse bite in man (n=1), existing in the authors collection, are highly needed to classify and name these two taxa, to improve our understanding of the biology of these taxa.

Taxon 7

(Conservatibacter flavescens)
Knowledge on the host reservoir of members of the family Pasteurellaceae mainly comes from investigations of taxa associated with disease in man and animals. To extend our knowledge on the epidemiology of these organisms, more systematic investigations within different phylogenetic groups of animals remains to be carried out.

Four new taxa, provisionally designated 5 through 8 were demonstrated in two colonies of conventional, well-managed guinea pigs by Bisgaard et al. (1983). A total of ten phenotypical characters separate taxon 7 from taxon 5 and taxon 6, while 11 phenotypical characters separate taxon 7 from taxon 8.

The polyamine pattern of strain 7.4 (CCUG 24852T) of taxon 7 showed that 1,3-diaminopropane (DAP) made up 88.2% followed by spermine (SPM) 6.5%, spermidine (SPD) 4.1%, putrescine (PUT) 0.8%, and cadaverine (CAD) 0.4% (Busse et al., 1997). Almost similar values were obtained for two strains of taxon 6.

De Ley et al. (1990) demonstrated at least seven rRNA branches among 53 strains known or presumed to belong to the Pasteurellaceae. A single isolate of Taxon 7 (10.4) was located at the common root for the rRNA branches for P. multocida, A. lignieresii, H. influenzae, [H.] aphrophilus and [A.] actinomycetemcomitans.

Strain 10.4 of taxon 7 demonstrated 85% DNA:DNA binding with strain 7.4 taxon 7, documenting that these strains belong to the same species. Between 23 and 27% binding was observed between strain 10.4 of taxon 7 and strains 3.3, 7.5 and 8.3 of taxon 5, 6 and 8, respectively, indicating that taxon 5 through 8 represent different genera of Pasteurellaceae (Bisgaard et al., 1983).

A total of 11 isolates of taxon 7 of Bisgaard and four new urease negative isolates from guinea pigs in Germany were investigated by a polyphasic approach by Adhikary et al. (2018).

The type strain of taxon 7, 7.4T, showed a divergent position among the genera of Pasteurellaceae in the 16S rRNA gene sequence phylogeny. The highest similarity of 95.4% was obtained to the type strain of Bibersteinia trehalose NCTC 10370T, confirming a DNA:DNA binding of 44% observed between taxon 7 strain 7.4 and [P.] haemolytica biovar T NCTC 10624 (Bisgaard et al., 1983). All five strains of taxon 7 examined showed identical 16S rRNA gene sequences. Similarly, all 15 strains of taxon 7 shared the same rpoB gene sequence.

Phylogenetic analysis, including maximum-likelihood and neighbour- joining methods, of 31 randomly selected proteins conserved in all genomes investigated, confirmed the divergent position of Conservatibacter, observed in the 16S rRNA gene based tree (Christensen and Bisgaard, 2018). In the MALDI-TOF analysis, nine strains tested showed 72-83% similarity to the type strain.

Due to the conserved genetic nature of all the strains investigated, taxon 7 was named Conservatibacter flavescens, flavescens referring to the yellow tinge of the colonies (Adhikary et al., 208). Two or more phenotypical characters separate Conservatibacter from other genera of Pasteurellaceae.
The G+C content of the type strain, 7.4T (CCUG 24852T) of the type species is 39.5 mol% as determined by whole genomic sequencing (Adhikary et al., 2018).

Taxon 8

(Remains to be classified and named as a new taxon of Actionobacillus)
A strong host-association was believed to exist for most members of the family Pasteurellaceae in the beginning of the 1980’ies, based upon diseases associated with taxa belonging to the family. Systematic investigations within different phylogenetic groups of animals were, however, lacking to support this perception.

To improve our knowledge on the natural occurrence of members of Pasteurellacae in healthy guinea pigs, Bisgaard et al. (1983) examined to colonies of conventional, well-managed guinea pigs. A total of four new taxa, designated 5 through 8 were demonstrated in both colonies. Differences in five, six, and eleven phenotypical characters separated taxon 8 from taxon 5, 6, and 7, respectively. Using the phenotypic characters, normally used for separation of genera of Pasteurellaceae (Christensen et al., 2020), did not allow separation of taxon 8 from the Actinobacillus genus, and only differences in a single phenotypic character separated taxon 8 from the genera Aggregatibacter, Avibacterium, Gallibacterium, Pasteurella, Rodentibacter, and taxon 35 of Bisgaard (unpublished data). Extending the number of phenotypic characters for comparison and separation of taxon 8 from the genus Actinobacillus was unsuccessfull (unpublished data). Comparison of taxon 8 with the single taxa of the genus Actinobacillus allowed separation from the named taxa, but not from taxon 9 (Genomospecies 2), (Christensen and Bisgaard, 2004).

According to Christensen et al. (2004) taxon 8 formed a monophyletic group with 11 taxa of Actinobacillus in the phylogeny of the family Pasteurellaceae based on maximum-likelihood analysis of 16S rRNA gene sequences. The bootstrap value for this group was 68%.

DNA:DNA hybridizations showed 49, 46, and 40% binding between strain 8.3 of taxon 8 and [H.] pleuropneumonia CCM 5869T, A. lignieresii NCTC 4189T and [P.] ureae NCTC 10219T, respectively, confirming the phenotypical results obtained. Bindings between 8 and 23% were observed between strain 8.3 of taxon 8 and taxon 5 strain 3.3, taxon 6 strain 7.5, taxon 7 strain 10.4, and [P.] aerogenes Carter SS 274-74. The G+C content of DNA from strain 8.3 of taxon 8 was 39.6 mol% and the genome size 1.8 x 109 d (Bisgaard et al., 1983).

Additional investigations, including whole genome sequencing and deduction of more conserved gene sequences for comparison are highly needed, in addition to comparison of randomly selected conserved protein sequences, before a final classification and naming of taxon 8 is possible.

Taxon 9

Trehalose negative isolates of Actinobacillus equuli were established as a separate group tentatively named taxon 9 in 1993 (50). These organisms had been obtained from the mucous membrane of oropharynx of apparently normal horses in addition to arthritis. Additional investigations have also associated these bacteria with cases of septicaemia in a horse and foal (70, 85). Subsequent investigations of 18 isolates showed that taxon 9 represents two novel species (115), A. arthritidis and Actinobacillus genomospecies 2, both of which have been obtained from as well diseased as normal horses. The closest 16S rRNA sequence based relationships for these species were A. ureae and A. hominis, respectively. According to the phylogenetic tree of the family Pasteurellaceae based on partial rpoB sequences A. lignieresii, A. pleuropneumoniae, A. arthritidis and A. genomospecies 1 and 2 form a separate cluster with a bootstrap value of 100 (Korczak et al. Int. J. Syst. Evol. Microbiol. 2004, 54, 1393-1399).

Taxon 10

Organisms tentatively classified with taxon 10 remains to be properly published. However, strains classified with this group have previously been included in other studies for comparison (127, 129, 133, 136). Taxon 10 produces gas from glucose, and has so far only been obtained from horses and horse bites in humans (Bisgaard, unpublished data). 16S rRNA gene sequence phylogeny demonstrated a close relationship with taxon 6 and [P.] aerogenes biovars 17 and 18 (136). These organisms remain to be classified and named to improve proper identification and general understanding on the importance of these bacteria.

Taxon 11

Evidence obtained to indicate that equine isolates of organisms previously reported as A. suis or haemolytic variants of A. equuli might constitute a separate taxon provisionally named taxon 11 was first published by Bisgaard et al. (25). Four biovars were initially reported, and selected DNA:DNA hybridizations did not support classification with A. suis. Additional biovars have been published, and several reports have associated these organisms with pathological lesions. However, other papers seem to point to a true predilection of taxon 11 for the respiratory tract of horses (50). Subsequent studies demonstrated that A. equuli and taxon 11 represent two different genotypes which differ with respect to disease pattern and epidemiology. For the same reasons two subspecies of A. equuli have been proposed: A. equuli subsp. e quuli (former A. equuli) and A. equuli subsp. haemolyticus (former taxon 11) (111).

RTX toxin activity (eqx) has been demonstrated in isolates of A. equuli subsp. haemolytica (Berthoud et al. Vet. Microbiol. 2002, 87, 159-174), and a stronger cytotoxicity towards horse-compared to pig lymphocytes was reported by Kuhnert et al. (Vet. Microbiol. 2003, 92, 161-167). So far, eqx has not been demonstrated in A. equuli subsp. equuli associated with sleepy foal disease and septicaemia in piglets.

Taxon 12

(Muribacter muris gen. nov., comb. nov)
In the early 1980ies simplified diagnostic keys, which did not allow an unambiguous separation of taxa of Pasteurellaceae, were often used in diagnostic laboratories, making interpretation of the literature uncertain. Using an extended number of phenotypical tests for characterization, Bisgaard (1986) investigated the natural occurrence of Pasteurellaceae in four breeding colonies of healthy white laboratory mice.

Nine strains were classified as [P.] pneumotropica sensu stricto (R. pneumotropicus) and 24 strains as [P.] pneumotropica type Heyl (R. heylii), while 19 strains formed a homogeneous group tentatively designated taxon 12. These organisms were demonstrated in two of the three conventional breeding colonies examined.

Phenotypical characters obtained with strains of taxon 12 were compared to those from [P.] ureae NCTC 10219T, [P.] ureae Ackerman 80-443D, [P.] pneumotropica types Jawetz NCTC 8141T, Heyl (P313), and Henriksen (953/60), in addition to [H.] influenzae-murium (NCTC 11146=EO1). Similar reactions were obtained for taxon 12 and [P.] ureae Ackerman 80-443D.

Mutters et al. (1984) re-examined the taxonomy of [P.] ureae by conventional phenotypic characterization and by DNA-DNA hybridization, and located human isolates of [P.] ureae in the Actinobacillus group together with A. hominis. The murine isolate, J.J. Ackerman 80-443D, appeared to represent a hitherto unrecognized species of Pasteurellaceae, demonstrating 30% DNA binding with the type strain of A. ureae NCTC 10219T with a standard deviation of 7.6%, and a G+C content of DNA of 46.9 mol% and a genome size of 1.49x109d (Mutters et al., 1984).

According to Bisgaard (1986) taxon 12 differs in at least 13 or more phenotypical characters from A. ureae sensu stricto, [H.] influenzae-murium and [P.] pneumotropica types Jawetz, Heyl and Henriksen. However, 51% DNA binding was observed with the avian hemolytic Actinobacillus-like complex (taxon 26; Bisgaard, 1993), while 52% DNA-DNA binding was observed between [P.] ureae strain 80-443D and the type strain of A. ureae NCTC 10219T (Piechulla et al., 1985). For the same reason, it was suggested to classify taxon 12 with the genus Actinobacillus Brumpt 1910, as A. muris sp. nov. (Bisgaard, 1982). Extended DNA:DNA hybridizations with the [P.] pneumotropica-complex (Ryll et al., 1991) confirmed that taxon 12 should be classified outside this group.

Investigations on the phylogeny of the family Pasteurellaceae linked [A.] muris with [H.] influenzae-murium in the maximum-likelihood analysis of 16S rRNA gene sequences, while it made up a separate branch in the analysis of infB and rpoB genes (Christensen ad Bisgaard, 2004; Christensen et al., 2004).

In the phylogenetic tree of the family Pasteurellaceae based on partial rpoB sequences, [A.] muris also appeared as a separate branch outside the genus Actinobacillus (Korczak et al., 2004). Finally, Nørskov-Lauritsen et al., (2004) excluded [A.] muris from the genus Actinobacillus sensu stricto based on analysis of partical infB sequences.

The taxonomy of [A.] muris was reinvestigated by Nicklas et al. (2015), based on 474 strains, mainly from mice and rats. All 474 strains were characterized by phenotype, from which 130 strains were selected for genotypic characterization by 16S rRNA (125 strains) – and partial rpoB gene sequencing (127 strains). The type strain was also investigated by whole-genome sequencing.

A total of 16 biovars were identified among the 474 strains investigated. The 16S rRNA gene sequence based phylogeny confirmed the existence of a monophyletic group including the type strain of [A.] muris and all 16 biovars demonstrated. The lowest similarity within the group was 96.7%, slightly below the normal lower limit for a species, however, all members of the taxon were related in a continuum, and neither genotypic nor phenotypic differences justified separation into more species.

The highest similarity outside the group was observed with [H.] influenzae-murium HIM 565-1 with 96.4% similarity, which is below the recognized 16S rRNA gene sequence similarity threshold of 97% for separation of species (Stackebrandt and Goebell, 1994).

In the 16S rRNA gene sequence multiple alignment, all strains of [A.] muris finally demonstrated a characteristic deletion of five nucleotides in the region 211-217 (numbering according to the sequences of P. multocida NCTC 10322T).

Ten groups and a singleton were demonstrated in the rpoB phylogeny. However, a high similarity of 97.2-100% was observed within the taxon. The closest related taxon outside the [A.] muris group was [H.] influenzae-murium with 88.4% similarity, which is below the range of 91-99% between species of Pasteurellaceae (Bisgaard et al., 2012; Korczak et al., 2012).

Based upon the findings stated above a new genus and a new combination, Muribacter muris gen. nov., comb. nov. were proposed. A single strain was found to be V-factor dependent (Nicklas et al., 2015).

Strain 33696Asv8, originally classified as biovar 4 of taxon 26 (non-haemolytic), was recently excluded from taxon 26 (A. anseriformium) (Bisgaard and Christensen, 2012). Nicklas et al. (2015) reclassified this strain as biovar 1 of M. muris. Strain 33696Asv8 also shared an identical rpoB gene sequence with the type strain of M. muris.

Since biovar 4 of taxon 26 was obtained from chickens two reservoirs for M. muris seem to exist as previously demonstrated for R. ratti (Adhikary et al., 2017). Classification of genera of Pasteurellaceae based upon conserved predicted protein sequences confirmed the classification of taxon 12 as a separate genus (Christensen and Bisgaard, 2018).

Taxon 13

Routine phenotypic characterization of P. multocida-like organisms obtained from pneumonic calf lungs often result in a fermentation pattern different from what is generally accepted for P.multocida sensu stricto. A group of indole-, mannitol- and sorbitol-negative isolates was tentatively designated taxon 13 (26). Selected DNA:DNA hybridizations classified ornithine positive taxon 13 isolates with P. multocida biovar 6 ( P. canis) while ornithine negative isolates were classified as V-factor independent isolates of [H.] avium. Additional DNA:DNA hybridizations demonstrated 80-92% DNA binding with P. canis for ornithine positive isolates of taxon 13 while ornithine negative isolates demonstrated 88% DNA binding with V-factor requiering isolates of [H.] avium biovar 1 classified as [P.] avium (now Avibacterium avium). Ornithine negative isolates of taxon 13 also demonstrated 69-81% DNA binding with P. stomatis (Mutters et al. Int. J. Syst. Bact. 1985, 35, 309-322). Additional information of the distribution and importance of these organisms have been published (39, 41,50).

Since routine identification of taxa belonging to Pasteurella rest on a single or a few phenotypic characters, and isolates aberrant for these characters have been reported more focus has been made on improvement of specific diagnostic genetic tools of importance for both the veterinary and medical professions. New diagnostic tools developed, however, questioned the correct classification of taxon 13 as discussed by Christensen et al. (131) who finally demonstrated that taxon 13 represents P. multocida sensu stricto, explaining positive reactions with probes (107) and in PCR- methods developed for detection of P. multocida(see 131 for references). Based upon MLST these organisms form a separate group (Unpublished data).

Taxon 14

A phenotypically homogeneous group of avian Pasteurella-like organisms including two biovars were reported by Bisgaard & Mutters (28). DNA base composition of biovars 1 and 2 varied between 53.1 (biovar 1) and 52.0 mol % (biovar 2) while both biovars demonstrated a genome mass of 1.9 x 10 9 Dalton. Ninety-seven % DNA similarity was observed between biovars 1 and 2. DNA:DNA hybridizations, however, failed to classify these organisms, tentatively named taxon 14. DNA:rRNA hybridizations, classified taxon 14 at the common root of the seven rRNA cistrons outlined for Pasteurellaceae (36). These organisms also demonstrated a separate polyamine pattern (Busse et al. Int. J. Syst. Bact. 1997, 47, 698-708). 16S rRNA gene sequence phylogeny classified taxon 14 with taxon 32 and 40, and [P.] testudinis (127, 129, 133). Similar observations have been reported for rpoB (Korczak et al. Int. J. Syst. Evol. Microbiol. 2004, 54, 1393-1399).

Taxon 14 has been obtained from upper respiratory tract lesions in ducks, turkeys, pigeons, a goose and a peafowl (28, 50) and blepharoconjunctivitis in turkeys (Bisgaard et al. WPSA, 4 th Int. Symp. on Turkey Prod. 2007 pp. 59-50). These organisms still remain to be classified and named , to improve proper identification and general understanding on the epidemiology and importance of these bacteria.

Taxon 15 biovar 1 and 2

Mannheimia varigena

Reports on [P.] haemolytica (M. haemolytica) in hosts other than ruminants have been questioned by Bisgaard (1984). Isolates from pigs were sufficiently different from bovine isolates to constitute a new taxon, tentatively designated taxon 15. Additional investigations including mol % G + C in DNA, genome mass, and DNA:DNA hybridizations confirmed that taxon 15 constitutes a separate group within the family Pasteurellaceae (Mutters et al., 1986). Based upon quantitative evaluation of phenotypic data Angen et al. (1997(2)) outlined two biovars of taxon 15 and demonstrated that taxon 15 biovar 1 and biogroup 6 of [P.] haemolytica from ruminants could not be separated by phenotypic tests. The same was valid for taxon 15 biovar 2 and biogroup 7 of [P.] haemolytica from ruminants.

Subsequent ribotyping using HindIII for digestion of DNA showed that all strains of taxon 15 formed a separate cluster with the exception of two isolates, which clustered together with bovine strains of [P.] haemolytica biogroup 6. This group also included taxon 36 (Angen et al., 1997 (3)). Biovar 1 and 2 of taxon 15 were randomly distributed within the porcine cluster, while the maltose negative strains of biovar 1 and 2 formed a separate subcluster within the porcine cluster.
Results obtained by ribotyping have subsequently been confirmed by AFLP analysis (Axel et al., unpublished results).
Surprisingly, ribotyping of strains classified as [P.] haemolytica biogroup 6, 7, 8, 9 and 10 demonstrated a very heterogenous picture in the cluster analysis. Similarly, MLEE did not allow a clear separation between porcine and bovine strains (Angen et al., 1997(3)). Based upon maximum-likelihood analysis of 16S rRNA gene sequences, taxon 15 and [P.] haemolytica biogroup 6 formed a separate cluster branching into a porcine and bovine line. Strains representing these lines, however, demonstrated a DNA:DNA relationship on species level. For the same reason, these organisms were classified with the same species, Mannheimia varigena (Angen et al., 1999). Taxon 36 demonstrated 88 % DNA binding with M. varigena, confirming the high genetic affiliation between these groups demonstrated by ribotyping (Angen et al., 1997(3)). rpoB based phylogeny also clustered porcine and bovine isolates of M. varigena together (Korczak et al. 2004).

So far M. varigena has been isolated from pneumonia, mastitis and septicaemia as well as the oral cavity, rumen and intestines of cattle. Isolates of porcine origin have been associated with septicaemia, enteritis or pneumonia, but isolates from the upper respiratory tract have also been reported (Bisgaard, 1993, Angen et al, 1999).

Further investigations, including whole genomic sequencing are highly indicated to obtain more information on the leucotoxin operon and its possible role in the host specificity of porcine and bovine isolates of M. varigena.

Taxon 16

(Frederiksenia canicola gen. nov., sp. nov.)
Significant variations in the physiologic properties and host reservoir of P. multocida have been reported (Heddleston, 1976), and until the beginning of the 1990ies a nonmotile, nonhaemolytic, oxidase- and indole-positive, urease-negative, fermentative, Gram-negative rod-shaped bacterium that exhibited no special nutritional or atmospheric requirements would have been identified as P. mutocida in routine laboratory diagnosis (Biberstein et al., 1991). For the same reason, most of the early reports on Pasteurellaceae in animals and human beings give rather incomplete descriptions of the species dealt with, just as certain characters of the organisms are not compatible with the species reported according to our present knowledge.

Bisgaard and Mutters (1986) characterized 23 unclassified canine and feline strains of Pasteurellaceae obtained from the oral cavity and compared results obtained with those of P. canis NCTC 11621T, P. stomatis NCTC 11624, [P.] gallinarum ATCC 13361T, P. dagmatis NCTC 11617T, Pasteurella sp. B (P. oralis) (Frederiksen P809), [P.] spp strains Smith 392, NCTC 11051 and P672, and [H.] influenza-murium NCTC 11146.

Eleven strains were classified with P. canis (n=4), P. stomatis (n=2), P. dagmatis (n=4) and P. sp B (P. oralis) (n=1).
Nine canine and three feline strains formed homogeneous group, tentatively designated taxon 16. This group had similar phenotypic characters as the reference strains CCUG 3467C and [P.] sp Smith 392.

Phenotypic results obtained with taxon 16 clearly indicated that taxon 16 should be classified with the genus Pasteurella. The guanine plus cytosine content of DNA varied between 43.5 and 44.3 mol%, while the genome mass varied between 1.6-1.8 x 109d. these values also suggest classification with the genus Pasteurella. DNA-DNA hybridizations between three strains of taxon 16 demonstrated 93-100% similarity. However, only 34% or less DNA binding was observed between taxon 16 (HPA 21) and selected species of Pasteurella sensu stricto, excluding classification with Pasteurella (Bisgaard and Mutters, 1986). Crossed immune electrophoresis also excluded taxon 16 from the genus Pasteurella (Schmid et al., 1991).

Additional isolates of taxon 16 from dogs and cats and humans have subsequently been published by Biberstein et al., (1991), Eckert et al., (1991), Bisgaard (1993), and Muhairwa et al. (2001).

The phenotypic relationships among 160 strains of Pasteurellaceae from free ranging chickens and their animal contacts based upon Euclidian distances and UPGMA, showed that all strains of taxon 16 clustered together (Muhairwa et al., 2001). Ribotyping using the restriction enzyme HpaII for digestion of DNA divided the phenotype cluster into two connected ribotype clusters including 16 and 11 ribotypes, respectively. Both clusters included dog and cat strains (Muhairwa et al. 2001).

rRNA cistron similarities of taxa belonging to the family Pasteurellaceae linked taxon 16 to the common root for the rRNA branches of P. multocida, A. lignieresii, H.influenzae, [H.] aphrophilus and [A.] actinomycetemcomitans (De Ley et al., 1990).

The phylogeny of the family Pasteurellaceae based on maximum-likelihood analysis of 16S rRNA gene sequences finally showed that taxon 16 makes up a separate group unrelated to the existing genera of Pasteurellaceae (Christensen and Bisgaard, 2004; Christensen et al., 2004).

A polyphasic analysis including 24 strains of taxon 16 from five European countries and representing isolates from dogs (n=16), humans (n=4), a cat, a lion, a hedgehog and a banded mongoose, was carried out by Korczak et al. (2014). Phenotypical results were similar to those of the genus Pasteurella and taxon 35 of Bisgaard (Christensen et al., 2020a; Christensen et al., unpublished data). Three or more characters, however, separate taxon 16 from P. multocida, P. canis and P. oralis, while two and a single phenotypic character separate taxon 16 from P. stomatis and P. dagmatis, respectively (Christensen et al., 2020b). α-fucosidase (ONPF) and α-glucosidase (PNPG) finally separate taxon 16 from taxon 35 (Bisgaard, unpublished data).

Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry clearly identified taxon 16 and separated it from other genera of Pasteurellaceae (Korczak et al., 2014).

In all phylogenetic analyses based on 16S rRNA, rpoB, infB, and recN genes taxon 16 formed a monophyletic branch with intraspecies sequence similarity of at least 99.1, 90.8, 96.8, and 97.2%, respectively (Korczak et al., 2014). The closest genetic relationship was with B. trehalosi with 95.9% 16S rRNA-, 89.8% rpoB-, and 74.4% recN gene similarity, while 84.9% infB gene similarity was demonstrated to A. lignieresii (Korczak et al., 2014).

Calculation of genome similarity values based on recN gene sequences, which represents an alternative to whole DNA-DNA hybridizations, demonstrated that the 24 investigated strains belong to the same species (values calculated were ≥0.90). Major respiratory quinones included menaquinone-8, ubiquinone-8, and demethylmenaquinone-8.

Based upon the findings stated above, a novel genus and species within the family Pasteurellaceae, named Frederiksenia canicola gen. nov., sp. nov. with the type strain HPA 21T (CCUG 62410T), was proposed for taxon 16 (Korczak et al., 2014).

Finally, a PCR based on a specific signature sequence in the recN gene was developed to improve identification and separation of F. canicola from other Pasteurellaceae obtained from dogs and cats, and bites inflicted by these animals (Korczak et al., 2014).

Taxon 17

(Rodentibacter rarus)
Some of the organisms from mice and rats received as unclassified Pasteurellaceae or Actinobacillus equuli are phenotypically related with Pasteurella dagmatis. However, differences in D (-) melibiose, α-galactosidase, and β-glucuronidase (PGUA) separate these organisms from P. dagmatis NCTC 11617T (P418). For the same reason, they were provisionally named taxon 17 by Bisgaard (1993).

A single isolate, P434, deviating from taxon 17 in indole, trehalose, and α-fucosidase was listed as [P.] pneumotropica type Henriksen by Frederiksen (1981). However, only 44% DNA:DNA binding was observed between P. dagmatis and P434 according to Ryll (1989), indicating that P434 might represent a new Pasteurella sp.

Only two isolates (HPA 41T and 2005150027) of taxon 17 were investigated by Adhikary et al. (2017). These strains demonstrated 98.5% 16S rRNA similarity, but only 91.2% rpoB gene sequence similarity. DNA:DNA bindings calculated from whole genomic sequences showed 93% binding between the two strains, with the highest binding of 45% within the genus Rodentibacter observed with R. heidelbergensis. Taxon 17 was therefore classified as R. rarus sp. nov.

The type strain suggested was 2325/79T (HPA41T, CCUG 17206T) obtained from a rat is the US. The DNA G+C content of the type strain is 40.4 mol% and the genome size 2.5 Mb as determined by whole genome sequencing.
Whole genome sequencing is suggested to classify strain P434.

Taxon 18

A total of 26 isolates obtained from the rumen of sheep and classified as A. lignieresii were originally received from Dr. J.E.Phillips. Further investigations showed that these isolates might represent non-haemolytic and maltose and dextrine negative [P.] haemolytica and were provisionally named taxon 18 (30). Based upon differences in D (-) sorbitol and L (+) arabinose four different biovars were outlined. An additional biovar was outlined by Angen et al. (74) who also demonstrated that these organisms clustered together with the exception of biovar 5. However, biogroup 8 D of [P.] haemolytica and biovar 2 of taxon 20 also clustered with taxon 18. Although forming a separate cluster biovar 5 of taxon 18 branched deeply with biogroup 9 of [P.] haemolytica. Ribotyping confirmed the relationship between taxon 18 biovars 1, 3 and 4 and [P.]haemolytica biogroup 8 D while other isolates of taxon 18 biovars 3 clustered separately. By MLEE a single isolate of taxon 18 biovar 1 clustered deeply with selected isolates of [P.] haemolytica biogroups 1, 8 D and 10 (78). The phylogenetic relationships of the [P.] haemolytica-complex as revealed by maximum-likelihood analysis of 16S rRNA gene sequences resulted in five major groups of which two strains of taxon 18 biovars 1 and 3 (HPA 81 and 92 T) formed a separate group with HPA 98 ([ P.] haemolytica biogroup 8 D) and UT 26 (atypical biogroup 1 of [P.] haemolytica).

Two strains of taxon 18 biovars 1 and 3 (HPA 92 T and 109) demonstrated 92 % DNA binding and clustered with HPA 98 at 88 % DNA binding. UT 26 finally showed 83 % DNA binding with HPA 98 and 84 % binding with UT 27 (biogroup 10 of [P.] haemolytica). For the same reasons these organisms were named Mannheimia ruminalis (87). Species of Mannheimia also form a separate phylogenetic group by analysis of rpoB(Korczak et al. Int. J. Syst. Evol. Micriobiol. 2004, 54, 1393-1399).

Analysis of the leukotoxin genotype by Southern blot and the corresponding β-haemolytic phenotype on sheep blood agar plates revealed that both characters were present only in strains of biogroup 1 (UT26), biogroup 190 (HPA95, HPA114, and UT27), and taxon 18 biovar 2 (HPA113, HPA90, and UT38), whereas strains of biogroup 8D and taxon 18 biovars 1, 3, and 4 have lost the leukotoxin operon.

Taxon 19

During investigations of porcine Pasteurellaceae (24) a separate group of isolates was outlined and tentatively named taxon 19. These isolates originated from septic infections in piglets and abortion in sows. However, only NCTC 10699 was included in the publication as a reference strain. Six out of 10 isolates investigated represented original isolates from Dr. N.S. Mair after whom these organisms subsequently were named (Sneath & Stevens Int. J. Syst. Bact. 1990, 40, 148-153).

rRNA:DNA hybridizations affiliated [P.] mairii at the common root of the rRNA brances of P. multocida, A. lignieresii, H. influenza and [H.] aphrophilus (36). An emended description of [P.] mairi was reported by Christensen et al. (136). Phylogenetic analysis based on 16S rRNA gene sequence comparison showed that [P.] aerogenes sensu stricto, [P.] mairii sensu stricto and [A.] seminis formed a monophyletic group representing a new genus candidate within the family Pasteurellaceae. Similar observations had previously been observed for rpoB by Korczak et al. (Int. J. Syst. Evol. Microbiol. 2004, 54, 1393-1399).

RTX toxins are predominantly observed in taxa of Pasteurellaceae and often associated with pathogenic representatives. The pax of [P.] aerogenes has been demonstrated in isolates associated with abortion in pigs, but not in those from other clinal conditions (Kuhnert et al. Infect. Immun. 2000, 68, 6-12). The pax gene has subsequently been demonstrated in all isolates of [P.] mairii investigated (155).

Taxon 20

Mannheimia granulomatis
Organisms classified with [P.] haemolytica were reported from 44 out of 81 European hares in France suffering from respiratory tract lesions (Louzis, 1984). Similar organisms were subsequently obtained from mammary and uterine infections and from 10 out of 42 apparently asymptomatic carriers investigated (Louzis et al., 1988). Phenotypically related organisms, provisionally named taxon 20, were subsequently reported from cases of purulent bronchopneumonia and from conjunctivitis in European brown hares in Belgium (Devriese et al., 1991), and rabbitis in Denmark (Bisgaard, 1993).

A simplified dendrogram indicating the phenotypic relationships among 339 strains within the [P.] haemolytica complex based upon average Euclidian distances and UPGNA outlined two biovars of taxon 20, of which biovar 2 clustered with taxon 18, while biovar 1 and biogroup 3J of [P.] haemolytica made up a subcluster of cluster IX, [P.] granulomatis and [P.] haemolytica biogroups 3A-H, representing the other two subclusters of cluster IX, respectively (Angen et al., 1997 (2)). Differences in glycosides, D(+) xylose, and PGUA separate [P.] granulomatis, taxon 20 biovar 1 and taxon 20 biovar 2.

Ribotyping, however, clustered biovars 1 and 2 of taxon 20 together including a single strain of [P.] haemolytica biogroup 3J from a roe deer. MLEE finally classified these organisms with [P.] granulomatis (Angen et al., 1997 (3)).
Taxonomic investigations of the [Pasteurella] haemolytica-complex as evaluated by 16S rRNA sequencing and DNA:DNA hybridizations reclassified [P.] granulomatis, Bisgaard taxon 20 and a roe deer isolate of [P.] haemolytica biovar 3J with the species Mannehimia granulomatis comb. nov. (Angen et al., 1999).

Phenotypic data have subsequently shown that eight strains from lesions in roe deer belong to a distinct biogroup (taxon) within M. granulomatis (Bojesen et al., 2007). 16S rRNA gene sequence analysis confirmed that two 16S sequenced roe deer strains formed a separate cluster within M. granulomatis, suggesting that they might belong to a distinct lineage of M. granulomatis in the 16S rRNA tree, the bovine strains of M. granulomatis (W4672/1 = biogroup 3J of [P.] haemolytica and P1135/26T = type strain of M. granulomatis) being their closest neighbour, while the leporine strains Ph13 (taxon 20 biovar 1) and BJ 1680.3 (taxon 20 biovar 2) made up a more distantly related group. These findings suggest the existence of three distinct lineages of M. granulomatis adapted to bovine, cervine, and leporine hosts, respectively.

Southern blot analysis finally indicated that the IktA gene was present in all seven roe deer isolates. However, the IktA gene was only amplified by PCR in the six β-haemolytic isolates (Bojesen et al., 2007).

Further studies, including whole genome sequencing of selected strains representing the subpopulations demonstrated within M. garnulomatis are clearly indicated to understand the host adaptation observed for many taxa of Pasteurellaceae, M. granulomatis included.

Taxon 21

(Rodentibacter pneumotropicus, R. heylii, and R. mrazii)
D (+) melibiose, raffinose and α-galactosidase positive [P.] pneumotropica type Jawetz received as Actinobacillus equuli, A. lignieresii and atypical [P.] pneumotropica have tentatively been named taxon 21 by Bisgaard (1993). These organisms have been obtained from the nasopharynx of laboratory mice and rats in different countries.

Two strains of taxon 21, 1014/87 and P441, demonstrated 95 and 83% DNA:DNA binding, respectively, with the type strain NCTC 8141T of [P.] pneumotropica type Jawetz, demonstrating that they belong to [P.] pneumotropica type Jawetz (Ryll et al, 1991).

Five strains of Bisgaard taxon 21 were included in the taxonomic investigation of Adhikary et al., (2017), who classified two phenotypically identical strains (3835-84 and 3900-84) with Rodentibacter heylii, while an indole positive (P441) – and negative (3880-84) strain were identified as R. pneumotropicus. Only production of acid from meso-inositol separated the two groups. The fifth isolate of taxon 21, Ppn 418T, only differed from the other four strains in β-glucuronidase (PGUA).

A species-like taxon, including 67 strains, was monophyletic as to rpoB - and 16S rRNA gene sequence comparison with bootstrap supports of 100 and 80%, respectively. The 16S rRNA gene sequence similarity was 96.9-100% within the group and included the type strain of [P.] pneumotropica type Jawetz (Rodentibacter pneumotropicus), strains of biovar Jawetz and taxon 21 (P441 and 3880-80). rpoB gene sequence similarities for 62 strains varied between 97.6 and 100%. The closest related species outside the group was 95% similarity to R. ratti, which is the species limit (Christensen et al., 2007).

DNA:DNA bindings calculated from whole genomic sequences were 80% or higher for the three strains compared within the species, with the highest DNA:DNA binding of 40% with R. ratti within the genus Rodentibacter.

Another monophyletic group based on rpoB – and 16S rRNA gene sequence analysis included [P.] pneumotropica biovar Heyl and two strains of taxon 21 (3835-84 and 3900-84), totaling 68 strains. Both rpoB – and 16S rRNA gene sequence analysis identified the group as monophyletic with 52 and 100% bootstrap support. Between 96.1 and 100% 16S rRNA gene sequence similarity was observed within the group, while 97.0% or less similarity was observed with 3835-84 (taxon 21). The lowest rpoB gene sequence similarity within the group was 96.9%, while 94.0% or less similarity was observed to other taxa. Five strains of the group, including the type strain R. heylii ATCC 12555T shared the rpoB -gene sequence of the type strain of R. pneumotropicus NCTC 4181T. However, 16S rRNA sequences separate the two species. Similar observations have been reported by Dome et al. (2010).

Strain Ppn 418T of taxon 21 was finally classified with six other strains as R. mrazii. All seven strains formed a monophyletic group in the 16S rRNA gene sequence analysis. The lowest 16S rRNA gene sequence similarity was 99.3% between the type strain Ppn 418T (taxon 21) and other strains of this taxon. The highest DNA:DNA binding to other taxa of Rodentibacter was 41% to R. pneumotropicus and R. ratti, confirming the species-like nature of the taxon.

Demonstration of three species within taxon 21, as stated above, clearly indicates the uncertainty of using phenotypic characters for identification of Rodentibacter species.  For the same reason, rapid genotypic tests are highly needed, and research recommended to develop unambiguous tests for identification.

Taxon 22

(Rodentibacter ratti)
Isolates of [P.] pneumotropica from the respiratory tract of apparently healthy chickens were first reported by Bisgaard (1982). Seven phenotypic characters separated the avian isolates from the type strain NCTC 8141T of [P.] pneumotropica type Jawetz.

Schulz et al. (1977) published on eight isolates of [P.] pneumotropica from albino rats, one of which (A5a = HIM 821-6) might resemble the isolates from chickens (Piechulla, personal information 1984). The avian [P.] pneumotropica-like group and strains from the respiratory tract of apparently normal wild rats were subsequently classified as Bisgaard taxon 22 (Bisgaard, 1993).

The taxonomy of the [P.] pneumotropica-complex was investigated by DNA:DNA hybridization by Ryll et al. (1991), who showed that [P.] pneumotropica biovar Jawetz represents a genus-like cluster containing several species including the V-factor dependent [H.] taxon B and Bisgaard taxon 22.

Extended characterization and comparison with relevant type strains reclassified 30 Pasteurellaceae strains from rodents. Two strains from rats received as [P.] pneumotropica were reclassified as taxon 22 of Bisgaard (Boot & Bisgaard, 1995).

Some 600 isolates out of approximately 1000 isolates, mainly from Muridae, have subsequently been reclassified as Rodentibacter ratti, which also included taxon 22 of Bisgaard (Adhikary et al., 2017). Phylogenetic analysis of both rpoB and 16S rRNA gene sequences identified the group as paraphyletic. The lack of recognition of the group as monophyletic in the 16S rRNA gene sequence phylogeny was explained as a result of low sequence divergence, since all strains shared 96.8% or higher 16S rRNA gene sequence similarity.

The lowest rpoB gene sequence similarity was 95.8% between Ppn334 (R. ratti), Ppn 418T (the type strain of R. mrazii) and F75T (the type strain of R. ratti). The similarity to other taxa was 95% or less, which is the species limit (Christensen et al., 2007).

DNA:DNA bindings deducted from whole genomic sequences was 66% or higher for the four strains compared within the species, while the highest DNA:DNA bindings within the genus was 41% with R. mrazii.

The type strain of R. ratti is F75T (CCUG 69665T), and the DNA G+C content of the type strain is 40.2 mol%, while the genome size is 2.9 Mb, determined by whole genome sequencing.

Taxon 23 and 24

Taxon 23 (Hamster group 2 of Krause et al., (1989)
Taxon 24 (Hamster group 1 of Krause et al., (1989)
Taxon 23 Cricetibacter osteomyelitidis gen. nov, sp.nov
Taxon 24 Mesocricetibacter intestinalis gen. nov, sp.nov
Five strains obtained from caecitis in Syrian hamsters and six strains obtained from osteomyelitis (n=3) and conjunctivitis (n=2) in European hamsters and pyometritis in a single Syrian hamster were reported as hamster groups 1 and 2, respectively, by Krause et al. (1989). Phenotypical characters investigated were in accordance with those of Pasteurellaceae. Differences in phosphatase, urease and acid production from maltose separated the two groups, just as differences in acid production from D (-) mannitol and D (-) sorbitol separated the two groups from [P.] pneumotroopica biovars Jawetz (R. pneumotropicus) and Heyl (R. heylii).

DNA-DNA hybridization results confirmed the distinctive nature of the two groups. Intra-group DNA-DNA bindings above 90% were observed for both groups, while very low bindings were observed between the groups and a variety of Actinobacillus and Pasteurella species and several unnamed groups of Pasteurellaceae (Krause et al., 1989). The G+C content of group 1 varied between 47.5 and 48.7 mol% with a genome size of 1.5-1.7x109d compared to 41.9-42.0 mol% and a genome size of 2.0x109d for group 2.

rRNA cistron similarities of taxa belonging to Pasteurellaceae linked the hamster group 1 (HIM 933-7/8) to the common root of the rRNA branches for P. multocida, A. lignieresii, H. influenza, [H.] aphrophilus, and [A.] actinomycetemcomitans (De Ley et al., 1990). These authors also showed that HIM 933-7/8 had a high mol% G+C of DNA (48.7).

According to Bisgaard (1993) hamster strains demonstrating similar phenotypic characters as the hamster groups 1 and 1 of Krause also existed in his collection as taxon 24 and taxon 23, respectively.

Almost similar polyamine patterns were demonstrated for hamster group 1 (Kunstyr 246/85) and 2 (Kunstyr 507/85) according to Busse et al., (1997). 1,3-diaminopropane (DAP) made up 92.9-94.3% of the polyamine content. DAP values above 90%, however, were also observed for A. equuli NCTC 8529T, A. hominis NCTC 11529T, A. lignieresii NCTC 4189T, A. pleuropneumoniae CCM5969T, [A.] rossii NCTC 10801T, A. suis CCM 5586T, A. ureae NCTC 10219T, Bisgaard taxon 26 (A. anseriformium), H. aegypticus NCTC 8502T, [H.] aphrophilus NCTC 5906T, H. influenzae NCTC 8143T, [H.] segnis NCTC 10977T, [P.] bettyae NCTC 10535T, M. granulomatis P1135T, M. haemolytica NCTC 9380T, [P.] langaaensis NCTC 11411T, B. trehalosi NCTC 10370T, Bisgaard taxon 6 MCCM 02144, and N. rosorum P603.
Eleven hamster strains of Bisgaard taxon 23 and taxon 24, also referred to as groups 2 and 1 of Krause, respectively, were characterized by a polyphasic approach by Christensen et al. (2014). Two phenotypical characters separate taxon 23 from the genera Aggregatibacter, Gallibacterium, and Pasteurella, while three or more phenotypical characters separate the remaining genera of Pasteurellaceae from taxon 23 (Christensen et al., 2020).

Two phenotypical characters separate taxon 24 from the genera Gallibacterium and Pasteurella, while three or more characters separate taxon 24 from the remaining genera of Pasteurellaceae (Christensen et al., 2020).

rpoB gene sequence analysis confirmed that taxon 23 and taxon 24 represent unique groups of Pasteurellaceae. Identical sequences were observed for all strains of taxon 23, while a single strain differed slightly (99.0% similarity) from the other strains of taxon 24. Only 83.6% similarity was observed between the two groups. The closest similarity between HIM 933-7/8T of taxon 24 and other genera of Pasteurellaceae was 86.6% to N. rosorum, while 87.6% similarity was observed between HIM 943/7T of taxon 23 and B. trehalosi.

Both taxon 23 and taxon 24 formed monophyletic groups in the neighbour-joining analysis of 16S rRNA gene sequences. The lowest sequence similarity within the groups was 98.8% observed between strains HIM 933-7/8T and 1986057021 of taxon 24. The closest relative to taxon 24 was the type strain of H. haemolyticus with 95.2% similarity.

Taxon 23 strain HIM 943/7T was unrelated to other members of the family Pasteurellaceae and the strains of taxon 24. The closest relative to taxon 23 was M. glucosida with 95.2% sequence similarity.

The 16S rRNA gene sequence similarity between strain HIM 933-7/8T of taxon 24 and strain HIM 943/7T of taxon 23 was 93.9%.

Since taxon 23 and taxon 24 formed two monophyletic groups based on 16S rRNA gene sequence comparison to other members of the family Pasteurellaceae and partial rpoB sequencing, and DNA-DNA hybridizations showed high genotypic relationships within both groups, a new genus with one species, Cricetibacter osteomyelitidis gen. nov., sp. nov. was proposed for taxon 23. Similarly, a new genus with one species, Mesocricetibacter intestinalis gen. nov., sp. nov., was proposed to accommodate members of taxon 24 (Christensen et al., 2014).

The type strain of C. osteomyelitidis is HIM 943/7T (Kunstýř 507/85T = CCUG 36451T). Mol% G+C of DNA from the type strain is 42.0% and the genome size 2.0x109d (Krause et al., 1989). The type strain of M. intestinalis is HIM 933-7/8T (Kunstýř 246/85T = CCUG 28030T). Mol% G+C of DNA from the type strain is 48.7% and the genome size 1.5x109d (Krause et al., 1989).

Taxon 25

(Mannheimia cavia sp. nov.)
Several taxa of Pasteurellaceae, including taxon 25, were reported to be associated with guinea pigs by Bisgaard (1993). The ecology and significance of these taxa, however, remains to be investigated, just as a final classification and naming is needed to allow unambiguous identification.

Phenotypically identical Pasteurella-like organisms were reported by Mannheim et al., (1978) from an outbreak of epidemic conjunctivitis in a large colony of guinea pigs. Isolates from guinea pigs, previously reported as [Pasteurella] gallinarum (later reclassified as Avibacterium gallinarum), were subsequently identified as members of the SP-group (Necropsobacter rosorum), taxon 6 or taxon 25 (Boot and Bisgaard, 1995). Two isolates recovered from guinea pigs suffering from otitis media and classified as taxon 25 have subsequently been shown to share the phenotypical characteristics of a group of organisms represented by  strain T138021-75T as reported by Mannheim et al. (1978) (Bisgaard, unpublished results).

A single lung isolate, P244, obtained from a rabbit and phenotypically related to M. granulomatis clustered with the genus Mannheimia in all phylogenetic trees based on the rrs gene, the consensus tree based on rpoB, infB and recN genes, and individual trees based on these three genes (Kuhnert et al., 2007), without demonstrating a clear species association.

A polyphasic approach was applied to classify and name the four strains mentioned above (Christensen et al., 2011). The type strains of the five species of Mannheimia and the type strain of P. multocida were included for comparison based on their published DNA sequences.

Phenotypic results obtained were in accordance with those previously reported for members of the genus Mannheimia (Angen et al., 1999). Four to nine different characters separated the strains investigated from recognized species of the genus Mannheimia, while differences in ornithine decarboxylase, growth on MacConkey agar, β-glucosidase (NPG), α-galactosidase, β-xylosidase (ONPX), and production of acid from maltose, D (+) melibiose, raffinose, dextrin and glycosides separated the guinea pig isolates from the rabbit isolate (Christensen et al., 2011).

Strains T138021-75T and Pg20 of taxon 25 shared identical 16SrRNA gene sequences and were distantly related to other species of Mannheimia. The highest 16S rRNA gene sequence similarity was to strains of M. glucosida (97.8%), while the type strain of M. varigena showed 97.3% 16S rRNA gene sequence similarity. All other type strains of the genus Mannheimia, including the type strain of the type species (96.9%) demonstrated less than 97% similarity. In the maximum-likelihood analysis of 16S rRNA gene sequences the rabbit strain, P224, clustered with T138021-75T demonstrating a bootstrap value of 68%. These two strains demonstrated 98.6% 16S rRNA gene sequence similarity.

All three strains of taxon 25 shared identical rpoB gene sequences. The closest related strains included the type strains of M. varigena and M. haemolytica biovar 8, both of which showed 93.9% similarity. The highest rpoB gene sequence similarity observed for P224 was 95.1%, which was observed with the type strains of M. haemolytica and M. glucosida. The rabbit isolate, P224, only demonstrated 89.9% sequence similarity with taxon 25 strains

The three guinea pig isolates of taxon 25 also shared identical recN gene sequences, and the close relationship of the rabbit strain and the three guinea pig strains in the 16S rRNA gene sequence based phylogenetic tree was confirmed in the tree based on maximum-likelihood analysis of partial recN gene sequences.

DNA reassociation values between the guinea pig strain T138021-75T and the rabbit strain P224 were 82.6% and 80.5% (mean 81.6%), while the DNA relatedness values between strain T138021-75T and M. glucosida DSM 19638T were 41.8 and 38.7% (mean 40.3%), indicating that the rabbit isolate should be classified with the guinea pig isolates, since species of the family Pasteurellaceae mainly have been defined on the basis of DNA-DNA reassociation values of 80-85% as measured by the spectrophotometric method (Mutters et al., 1989: Christensen et al., 2005).

Finally, all three guinea pig strains of taxon 25 shared the same recN gene sequence, which showed 98.2% similarity with strain P224, while only 78.3-81.3% similarity was observed for species of Mannheimia. Converted to whole genome similarities (Kuhnert and Korczak, 2006) these values correspond to 0.93 between strains T138021-75T and P224 and 0.46-0.62 between strain T138021-75T and recognized species of Mannheimia. According to Kuhnert and Korczak (2006) strains of the same species showed similarity values of 0.9 and higher, while species of the same genus had similarity values above 0.4. Similarity values between 0.85 and 0.9 were considered intermediate, and could indicate different species or subspecies. In addition, threshold similarity values for recN alone, as used above, were comparable to MLSA with very few exceptions.

In conclusion, results obtained with the guinea pig strains of taxon 25 clearly documents that they represent a novel species of the genus Mannheimia, for which the name M. caviae sp. nov. is proposed. The type strain is T138021-75T (CCUG 59995T), and the G+C content of DNA from the type strain is 41.4 mol%.

The differences in hosts, genotypic- and phenotypic characteristics between the rabbit isolate P224 and the guinea pig isolates T138021-75T, Pg19 and Pg20 have parallels in both M. varigena for which two genotypic populations associated with pigs and ruminants, respectively, have been demonstrated (Angen et al., 1999) and M. granulomatis for which three populations have been demonstrated associated with leporine, bovine and cervine hosts, respectively (Bojesen et al., 2007). For the same reason, additional leporine, isolates should be investigated before final taxonomic conclusions can be drawn.

Taxon 26

(Actinobacillus anseriformium sp. nov.)
Hacking and Sileo (1977) first reported on a hemolytic Actinobacillus from waterfowl. Similar organisms were subsequently isolated in mixed flora from the respiratory tract of apparently normal White Pekin ducks. However, a single strain was obtained in pure culture from a duck suffering from conjunctivitis (Bisgaard, 1982). An isolate of A. suis from a Canada goose (Maddux et al., 1987) was subsequently shown to belong to biovar 1 of the avian hemolytic Actinobacillus-complex (Bisgaard, unpublished data). These organisms, tentatively named taxon 26, are normally isolated from the upper respiratory tract of web-footed birds in which they may cause sinusitis, conjunctivitis, and septicemia. Four biovars of taxon 26 exist. Geographically, these organisms have been isolated from web-footed birds in Germany, Belgium, USA and Denmark (Bisgaard, 1993).

DNA-DNA hybridizations with biovars 1-3 were carried out by Piechulla et al., (1985). Eighty-two % DNA-DNA binding was observed between biovars 1 and 3 indicating that acid production from D (-) sorbitol seems of little taxonomic importance. However, only 62% DNA-DNA binding was observed between biovar 2 and 3 indicating that production of acid from D (-) mannitol is of taxonomic importance. A DNA-DNA binding of 51% between biovars 1 and 2 confirms this observation.

Finally, 52% DNA-DNA binding was observed between biovar 3 and A. ureae NCTC 10219T, indicating a close relationship with the Actinobacillus genus.

The G+C content of DNA from taxon 26 strains F64, F66 and F97 varied between 40.0 and 41.1 mol%, while the genome mass varied between 1.8 and 2.0 x 109d (Piecheulla et al., 1985).

According to rRNA cistron similarities of taxa belonging to Pasteurellaceae, taxon 26 biovar 3 (F64=NCTC 11410) was located at the common root of the rRNA branches for P. multocida, A. lignieresii and H. influenzae (De Ley et al., 1990).

1,3 diaminopropane (DAP) made up 93.1% of the polyamine content of taxon 26 biovar 3 (F64). However, DAP values above 90%, have also been observed for A. equuli NCTC 8529T, A. hominis NCTC 11529T, A. lignieresii NCTC 4189T, A. pleuropneumoniae CCM5969T, [A.] rossii NCTC 10801T, A. suis CCM 5586T, A. ureae NCTC 10219T, Bisgaard taxon 26 (A. anseriformium), H. aegypticus NCTC 8501T, [H.] aphropohilus NCTC 5906T, H. influenzae NCTC 8143T, [H.] segnis NCTC 10977T, [P.] bettyae NCTC 10535T, M. granulomatis P1135T, M. haemolytica NCTC 9380T, [P.] langaaensis NCTC 11411T, B. trehalosi NCTC 10370T, Bisgaard taxon 6 MCCM 02144, and N. rosorum P603 (Busse et al., 1997).

Comparative phylogenies of the 16S rRNA and infB genes classified taxon 26 CCUG 28015 (F64) with the genus Actinobacillus sensu stricto (Christensen et al., 2004).

Phenotypical characters separating taxa of Actinobacillus sensu stricto were reported by Christensen and Bisgaard (2004), who showed that taxon 26 is the most distantly related taxon of Actinobacillus s.s., showing between six and 14 divergent phenotypic characters to other taxa belonging to the genus.

Strain 20948/1 obtained from sinusitis in a goose (Christensen et al., 2009) shared the phenotypic properties of taxon 26 and represents a V-factor dependent strain of this taxon with 99.8% 16S rRNA gene sequence similarity to strain F66 (biovar 1 of taxon 26).

In the final study for classification and naming of taxon 26, 31 strains were selected for genotypic characterization (Bisgaard and Christensen, 2012).

To evaluate the genotypic coherence of taxon 26, the partial rpoB gene sequence of 31 strains was determined. Thirty strains had the same rpoB sequence, while a single strain diverged in 1nt. The highest rpoB similarity to other taxa was 89.7% to the type strain of A. equuli subsp. haemolyticus and 88.2% to the type strain of the type species A. lignieresii.

A 16S rRNA gene sequence similarity of 99.6-99.8 was observed for biovars 1-3 of taxon 26. The lowest 16S rRNA gene sequence similarity between strains of taxon 26 was 99.6%. The highest similarity to taxa outside the group was 96.4% with the reference strain of Actinobacillus genomospecies 2 and 96.2% with the type strain of A. equuli subsp. equuli; 95.8-95.3% similarity was obtained to the type strain of the type species of Actinobacillus, A. lignieresii.

recN similarities within taxon 26 varied between 99.5 (F66T (biovar 1) and F64 (biovar 3)) and 99.8% (F66T (biovar 1) and F67 (biovar 1)), corresponding to genome similarities of 93.9-94.6%, clearly at the upper limit for species, compared to other members of Pasteurellaceae (Kuhnert and Korczak, 2006; Christensen et al., 2011). The highest recN similarity outside taxon 26 was 83.4% with the type strain of A. capsulatus, where- as the similarity to the type strain of the type species of the genus Actinobacillus, A. lignieresii, was 80.9%, corresponding to genome similarities of 57.7% and 52.0% respectively, confirming a distant relationship of taxon 26 to other species of Actinobacillus. Phylogenetic analysis of recN and 16S rRNA gene sequences of taxon 26 and reference strains Actionobacillus both demonstrated that taxon 26 was the most distantly related member of the genus Actinobacillus sensu stricto.

A previous reported a DNA-DNA binding of 65% with a standard deviation of 7.3% between biovar 3 (F64) and biovar 2 (F97) suggests that biovar 2 represents a separate species, contradicting rpoB, infB and 16S rRNA results. However, DNA-DNA hybridization results have previously resulted in conflicting results (Kuhnert and Koreczak, 2006; Christensen et al., 2007; Bisgaard et al., 2009).

Based upon results stated above, a novel species, Actinobacillus anseriformium sp. nov. was proposed to include biovar 1-3 of taxon 26 (Bisgaard and Christensen, 2012). The type strain F66T (CCUG 60324T) was selected from biovar 1, which is the most frequently isolated biovar. Mol% G+C of DNA from the type strain was 41.1% and the genome mass 2.0x109d.

Taxon 27

(A new Muribacter species?)
Some strains from laboratory rodents received as A. lignieresii, [P.] ureae, P. pneumotropica and unclassified Pasteurellaceae did not meet the phenotypical characters of other Pasteurellaceae existing in the authors collection and were tentatively named taxon 27 to improve the attention to and identification of these organisms (Bisgaard, 1993).

Organisms obtained from mice (n=5), a rat, and a guinea pig and classified as A. lignieresii were reported by Hartmann (1981). These organisms, however, differed from taxon 27 in acid production from trehalose and raffinose.

All species of Rodentibacter are urease positive (Adhihary et al., 2017), separating members of Rodentibacter from taxon 27. A total of five characters (Phosphatase, tween 80, D (+) galactose, dextrin, and arbutin) separate taxon 27 from Muribacter muris (Nicklas et al., 2015), while differences in urease, phosphatase and acid production from glycerol, D (-) mannitol, D (+) melibiose, and raffinose separate taxon 27 from [H.] influenzae-murium NCTC 11146 of Csukás.

In the 16S rRNA gene sequence based phylogenetic tree, taxon 27, M. muris CCUG 16938T and [H.] influenzae-muris NCTC 11146 appeared monophyletic with a bootstrap value of 96%, while the bootstrap value of A. muris CCUG 16938T and strain 3996-85 of taxon 27 was 81%. These three taxa clustered with two strains of Bisgaard taxon 47, demonstrating a bootstrap value of 76%. Whole genome comparison of taxa belonging to this complex is highly indicated to classify and name taxon 27 and taxon 47, in addition to [H.] influenza-murium of Csukás.

Taxon 28

(Ornithobacterium rhinotracheale)
Late in the 1980ies Pasteurella multocida suspect cultures were received from colleagues in England and South Africa. These cultures had been obtained from typical fowl cholera in turkeys and respiratory disease accompanied by airsacculitis in broilers, respectively.

Several similar isolates classified as taxon 28 existed in the authors collection. These isolates had been obtained from turkeys demonstrating uni- or bilateral pleuropneumonia accompanied by focal necrosis and extended consolidation of the lung tissue, typical lesions of fowl cholera.

The colony morphology and biochemical properties of taxon 28, however, deviated significantly form P. multocida and other taxa of Pasteurellaceae. On primary isolation, pin point colonies were observed on blood agar after 24 hours incubation, very much similar to small-colony variants of P. multocida from chronic lesions (Bisgaard, unpublished data). Conventional phenotypic characters were mostly inconsistent, unless suspension cultures (massive inoculation) were used for inoculation of the test media.

Organisms classified as taxon 28 were Gram-negative, nonmotile, catalase negative and oxidase variable. A weak fermentative reaction was observed in Hugh & Leifsons medium with D (+) glucose.

Unlike members of Pasteurellaceae nitrate was not reduced. A positive reaction was observed for urease and phosphatase. A weak production of acid was observed from D (-) ribose, D (-) fructose, D (+) galactose, D (+) mannose, lactose, maltose and dextrin. Positive reactions were observed in the ONPG, PNPG, α-galactosidase, and α-mannosidase tests. Three strains of taxon 28 were deposited at CCUG as CCUG 23170, CCUG 23171T and CCUG 23172. These strains were subsequently included in the final classification of taxon 28 (Van Damme et al., 1994).

Although taxon 28 initially was used to outline this new group of organisms (van Beek et al. 1994; Van Empel and H. Hafez, 1999), classical phenotypic characterization, including carbohydrate analysis, fatty acid methyl ester analysis, and PAGE of whole-cell proteins, in addition to determination of G+C content of DNA, DNA-DNA and DNA-rRNA hybridization and phylogenetic analysis of 16S rRNA gene sequences allowed classification of taxon 28 as Ornithobacterium rhinotracheale gen. nov., sp. nov.

These organisms constitute a genotypically homogenous taxon in rRNA superfamily V, as shown by DNA-rRNA hybridization (van Damme et al., 1994). The G+C contents of members of O. rhinotracheale vary between 37 and 39 mol%, the type strain CCUG 23171T having a mol% of 38%.

Taxon 29

(Heddlestonia cuniculi)
Pasteurella multocida is the most common bacterial pathogen in rabbits, resulting in a variety of disease manifestations (Manning et al., 1989; Stahel et al., 2009). Similar to P. multocida from other animals, P. multocida from rabbits demonstrates major differences in the phenotype (Heddleston, 1976; Manning et al., 1989; Biberstein et al., 1991; Stahel et al., 2009), making an unambiguous diagnosis of this organism problematic (Christensen et al., 2004, Stahel et al., 2009).

Lack of unambiguous phenotypic methods also makes interpretation of early studies difficult, and recognition of new taxa of clinical importance troublesome. However, Arseculeratne (1961, 1962) reported on Actinobacillus capsulatus from caged Angora – and rabbits of mixed breed in Sri Lanka, and closely related organisms have been reported from snowshoe hares (Zarnke and Schlater, 1988) and European brown hares (Bisgaard, 1993). Isolates from European brown hares, tentatively named taxon 20 (Bisgaard, 1993), have subsequently been classified as Mannheimia granulomatis (Angen et al., 1999). Bisgaard (1993) also reported on the isolation of [P.] aerogenes and the SP-group (Necropsobacter rosorum) from rabbits. The occurrence of the SP-group in diseased rabbits was confirmed by Stahel et al. (2009), who also reported on the isolation of P. canis from diseased rabbits. Isolation of [H.] paracuniculus, reported by Targowski and Targowski (1979), has not been verified subsequently.
Christensen et al. (1011) subjected seven strains of Bisgaard taxon 29, obtained from lesions in rabbits, a polyphasic study.

Phenotypical results, obtained were in accordance with those of the family Pasteurellaceae (Christensen et al., 2020). With the exception of the genus Mannheimia, two or more phenotypic characters separate taxon 29 from previously reported genera of Pasteurellaceae (Christensen et al., 2020). However, combinations of urease, ONPG, and production of acid from L (-) fucose, maltose and melibiose allow proper phenotypic separation of taxon 29 from the genus Mannheimia (Angen et al., 1999).

Partial rpoB sequences demonstrated a high similarity within taxon 29, well beyond the limit for separation of species (Christensen et al., 2007).

Sequencing of the 16S rRNA gene of four strains of taxon 29 showed 99.8% similarity within the group, with 95.6% similarity to the closest related taxon, [H.] paracuniculus.

To honour the late American microbiologist, K.L. Heddleston, for his work on P. mutocida, the genus name Heddlestonia is proposed for taxon 29, while H. cuniculi is suggested for the type species of the new genus (Christensen et al., 2011).

Finally, strain Clin 29 reported by Stachel et al. (2009) has been shown to belong to H. cuniculi (unpublished data).

Taxon 30

(Diversibacter marburgensis)
The occurrence and significance of P. multocida and related taxa in rabbits were briefly reported under taxon 29. In addition, the often problematic and unsatisfactory identification of these taxa were briefly discussed.

In the presentation by Christensen et al. (2011) the results of a polyphasic study of Bisgaard taxon 30 from rabbits were briefly described.

Two phenotypic characters, phosphatase and ornithine decarboxylase separate taxon 30 from the genus Ursidobacter, while from three to 15 characters separate taxon 30 from the rest of the genera of Pasteurellaceae (Christensen et al., 2020).

Partial rpoB sequences demonstrated a high similarity within taxon 30. Similarly, the highest 16S rRNA gene similarity within taxon 30 was 99.9%, while only 95.8% similarity was observed with [H.] felis, the closest taxon outside taxon 30.

Since taxon 30 was shown to represent a separate phylogenetic group different from other genera of Pasteurellaceae, the name Diversibacter marburgensis is proposed for this taxon (Christensen and Bisgaard, unpublished data).

Taxon 31

(Avibacterium paragallinarum biovar 2)
Strains obtained from as well young as adult chickens in South Africa suffering from a disease resembling infectious coryza as to both clinical signs as pathology, except that the infra-orbital sinuses are not distended to the same degree (A.J. Morley and R.F. Horner, personal communications), were provisionally designated taxon 31, since they phenotypically deviated from known avian taxa.

Subsequently Mouahid et al. (1992) compared taxon 31 with the V-factor dependent [H.] paragallinarum associated with infectious coryza in chickens. Only differences in V-factor requirement, β-galactosidase (ONPG), and acid production from maltose separated taxon 31 from the type strain of [H.] paragallinarum ATCC 29545T. However, 89% DNA binding was observed between [H.] paragallinarum ATCC 29545T and strain SA4461 of taxon 31. Finally, the DNA base composition and genome mass of taxon 31 strain SA7191 were 41.8 mol % G+C and 1.7 x 109d, compared to 42.3 mol% G+C and 1.7 x 109d for [H.] paragallinarum ATCC 29545T (Mutters et al., 1985).

In 1985 Mutters et al. showed that [H.] avium proposed by Hinz and Kunjara (1977) included three DNA homology groups, which were genetically closer to P. multocida, the type species of the genus Pasteurella, than to H. influenzae, the type species of the genus Haemophilus. [H.] avium was consequently reclassified as [P.] avium comb. nov., while [P.] volantium and P. sp. A were proposed for the remaining two groups. Differences in L (+) arabinose, D (-) mannitol, D (-) sorbitol and ONPG separate these taxa from each other and from [H.] paragallinarum. Surprisingly, the ornithine negative strain K117 of Bisgaard taxon 13 ([P.] avium biovar 2), obtained from pneumonia in a calf, demonstrated 88% DNA:DNA binding with [P.] avium ATCC 29546T. Only differences in V-factor requirement separate these two strains.

Subsequently, Christensen et al. (2004) showed that [P.] avium biovar 2, P. canis biovar 2 and variants in key characters regarded essential for identification of P. multocida all belonged to P. multocida sensu stricto, and that identification of these phenotypically diverse lineages of P. multocida will have to partly depend on genotypic methods.

Phylogenetic investigation of the family Pasteurellaceae based on maximum-likelihood analysis of 16S rRNA gene sequences showed that the type strains of [P.] gallinarum, [P.] volantium, [P.] sp. A, and [H.] paragallinarum made up a monophyletic group with a bootstrap value of 73%, clearly indicating that these organisms should be reclassified with a new genus (Christensen et al., 2004).
These observations were confirmed by Blackall et al. (2005), who reclassified [P.]  gallinarum, [H.]  paragallinarum, [P.] avium and [P.] volantium as Avibacterium gallinarum gen. nov., comb. nov., Av. paragallinarum comb. nov., Av. avium comb. nov. and Av. volantium comb. nov. Key characteristics separating these species include catalase and production of acid from D (+) galactose, D (-) mannitol and maltose.

The DNA fingerprinting work of Miflin et al. (1995) provided evidence that the South African V-factor independent isolates of Av. paragallinarum are clonal, indicating that the DNA:DNA hybridization data for strain SA 4461 are valid for all biovar 2 strains of Av. paragallinarum. In addition, biovar 2 strains serotyped by Miflin et al. (1995) were allocated to the existing Page serotyping system, just as the biovar 2 strains yielded a positive reaction in the Av. paragallinarum specific PCR (Miflin et al., 1999). All these data support the inclusion of biovar 2 strains with the species Av. paragallinarum.

Biovar 2 strains of Av. paragallinarum, which were positive in the Av. paragallinarum-specific PCR and allocated existing Page serovars, have also been reported from Mexico (Garcéa et al., 2004).

Finally, multilocus sequence phylogenetic analysis of Avibacterium, including partially sequenced recN, rpoB, infB, pgi, and sodA genes, confirmed the existence of the species Av. paragallinarum, while a species complex encompassing Av. volantium, Av. avium, Av. gallinarum, Av. endocarditidis, and Av. sp. A could not be resolved. Except for Av. paragallinarum, identification of species of Avibacterium seems problematic, even by DNA sequencing (Bisgaard et al., 2012).

Taxon 32

Spirabiliibacterium pneumoniae
A single unclassified isolate from pneumonia accompanied by septicemia in a goshawk (Accipiter gentilis) (HPA 106) was reported by Bisgaard and Mutters (1986). A negative reaction in D (+) xylose, sucrose and aesculin separated HPA 106 from biovar 1 of taxon 14. Two additional isolates from an European sparrowhawk (Accipiter nisus) with similar lesions as reported for the goshawk and a goshawk with conjunctivitis, respectively, were subsequently classified with HPA 106 and named taxon 32 (Bisgaard, 1993).

The phylogenetic relationships of unclassified, satellitic Pasteurellaceae from different species of birds, as demonstrated by 16S rRNA gene sequence comparison, showed that the kestrel strains were related to taxa 14 and 32 of Bisgaard, demonstrating from 97.5% similarity between strains NCTC 11878 (Kestrel) and 5108/S2/90 (Taxon 32) to 97.9% between strains NCTC 11878 and HPA 106 (Taxon 32). Although the hawk (Taxon 32)- and kestrel strains demonstrated 16S rRNA similarity at species level, they demonstrated major phenotypic differences.
In addition, they originate from different genera of the same bird family and might consequently represent different phylogenetic lines (Christensen et al., 2009). Based upon rpoB and 16S rRNA sequencing data, a new genus, Spirabiliibacterium gen. nov., was provisionally suggested to include taxon 14, 32 and kestrel isolates (Bisgaard and Christensen, 2011).

Finally, a polyphasic approach was used to investigate organisms tentatively reported as taxon 14, taxon 32, and the kestrel taxon (Bisgaard and Christensen, 2021). These investigations showed that the two strains of taxon 32 demonstrated 100% 16S rRNA gene sequence similarity. Between taxon 32 and taxon 14, 96.2-96.9% 16S rRNA gene sequence similarity was observed, while taxon 32 and the kestrel taxon demonstrated 96.5-96.9% similarity. The 16S rRNA gene sequence based phylogenetic analysis allocated the three taxa to a monophyletic group with 100% bootstrap support. These data clearly suggest that these taxa form a new genus within Pasteurellaceae and that each taxon represents a new species.
Phylogenetic comparison of partial rpoB sequence comparison confirmed the 16S rRNA phylogeny. The two strains of taxon 32 showed 99.8% rpoB similarity, while the relationship between taxon 32 and taxon 14 was 88.2-90%. In addition, each of the three taxa represents new species of Pasteurellaceae, since species of Pasteurellaceae diverge by more than 95% partial rpoB gene sequence similarity (Christensen et al., 2007). Partial recN sequencing and prediction of DNA-DNA similarities based upon the formula in Kuhnert and Korczak (2006) resulted in an DNA-DNA similarity of 75% within taxon 14, 91% within taxon 32, and 30 and 34% between the kestrel taxon and taxon 14 and 32, respectively, while 34-37% similarity was estimated between taxon 32 and taxon 14. These results clearly indicate that taxon 14, 32, and the kestrel taxon belong to separate species of new genus.

Finally, whole genomic sequencing and pairwise sequence comparison of the genomes of the type strains of the three taxa demonstrated 20.5 and 26.8% similarity between taxon 14 and the kestrel taxon and taxon 32, respectively, while the kestrel taxon and taxon 32 demonstrated 21.6% similarity, well below the species threshold of 70%.

To investigate the genus like nature of taxa 14, 32 and the kestrel taxon further, the phylogenetic comparison of 41 concatenated conserved protein sequences was performed with 82 taxa including type strains of type species of 28 genera of Pasteurellaceae. The reference strains of taxon 14 and 32 demonstrated 93% identity, while taxon 32 and the kestrel taxon demonstrated 89% identity. Between taxon 14 and the kestrel taxon 90% identity was found. A maximum of 85% protein identity was observed between genera, while a minimum similarity of 85% was observed within genera. The highest identity of the new genus-like group was found to [P.] testudines and Chelonobacter with 77 and 76%, respectively, well below the upper threshold between genera of 85%.

Phenotypically, taxon 32 can be separated from taxon 14 and Kestrel taxon in catalase, V-factor requirement, α-glucosidase, β-glucosidase, aesculin and acid production from D (+) xylose, sucrose and trehalose.

Based upon the data presented a new genus, Spirabiliibacterium gen. nov., including three species: S. mucosae gen. nov., sp. nov. (taxon 14), S. pneumoniae gen. nov., sp. nov. (taxon 32), and S. falconis gen. nov., sp. nov. (Kestrel taxon) was proposed (Bisgaard and Christensen, 2021).

Taxon 33

(Volucribacter)
The ecology and significance of members of the Pasteurellaceae were reviewed by Bisgaard (1993), who stressed that a more accurate basis for diagnosis is a prerequisite for improving our knowledge of the complexity of factors that govern the ecological preferences and virulence of there organisms. Subsequent characterization of 43 strains from pigeons, a goose, and psittacine birds and provisionally classified with Pasteurellaceae demonstrated the existence of a new taxon of Pasteurellaceae provisionally named taxon 33. This group included 12 strains obtained from blue-fronted parrots (Amazona aestiva; n=4), red-lored parrots (Amazona autumnalis; n=2), grey parrots (Psittacus erithacus; n=2), budgerigars (Melopsittacus undulatus; n=3), and a Fischers love bird. Eleven of the strains were obtained from different pathological lesions, mainly associated with the respiratory tract.

Ribotyping using HpaII for digestion of DNA and cluster analysis of the results obtained demonstrated that taxon 33 made up a separate cluster, confirming the results obtained by extended phenotyping. According to Christensen et al. (2003) taxon 33 clustered with the Avian 16S rRNA cluster showing 94.5% 16S rRNA similarity to strain 69 of taxon 34. According to Stenzel (1992), Beichel’s strain 103 (HIM 977-3) demonstrated 50% DNA:DNA binding with taxon 16 of Bisgaard.

In addition to 12 psittacine strains previously characterized phenotypically and ribotyped (nine strains), 13 further strains of which 12 were obtained from lesions in psittacine birds (Budgerigar (n=3), Blunt-tailed parrot (n=1), Blue-fronted parrot (n=2), Amazone sp. (n=1), Grey parrot (n=1), Parrot (n=2), and Parakeet (n=2)) were phenotyped and six strains were selected for 16S rRNA gene sequence comparison and phylogenetic analysis, while DNA:DNA hybridizations were carried out with five strains.

Two biovars of taxon 33 were distinguished. Biovar 1, including 22 strains, is ONPG negative and does not produce acid from L-fucose, maltose and dextrin, while biovar 2, including three strains, is positive for these characters.
16S rRNA gene sequence comparison showed that the four strains of biovar 1 (Gerlach 236/81, 94, 101 and 103) demonstrated 99.2% 16S rRNA similarity.
The lowest similarity observed between biovars 1 and 2 strains was 98.8% between strains 94 and B96/5, while the highest similarity (99.5%) was observed between strains 103 and 146/58/89.
Strain Gerlach 236/81 showed 99.2-99.9% similarity to other biovar 1 strains and 99.4% similarity to biovar 2 strains. A high 16S rRNA similarity between species belonging to the same genus has previously been reported for the type strains of Av. gallinarum and Av. volantium and A. lignieresii and A. pleuropneumoniae (99.7%) corresponding to 72% DNA reassociation (Pohl et al., 1983; Mutters et al., 1985). Similarly, 99.6% 16S rRNA similarity has been reported for the type strains of A. suis and A. equuli, corresponding to 66% DNA reassociation (Christensen et al., 2002).

The highest 16S rRNA similarity outside the taxon 33 was found to strain 69 of taxon 34 of Bisgaard (94.6%) and Av. avium (94.5%) for strain 101.
The high 16S rRNA gene sequence similarity between strains of taxon 33 was confirmed in the phylogenetic analysis, which showed a distinct monophyletic group of taxon 33 supported by a bootstrap value of 99%.
DNA:DNA hybridizations between strains Gerlach 236/81 and 101 of biovar 1 was 94%, while strain Gerlach 236/81 of biovar 1 and strain 146/S8/89 of biovar 2 only demonstrated 47% reassociation, suggesting the existence of two species within taxon 33.

Based upon the results stated above the name Volucribacter gen. nov. was suggested for taxon 33, biovar 1 of which was named V. psittacicida, while biovar 2 was named V. amazonae.

A value of 76% DNA reassociation was observed between the type strain of V. psittacicidae (Gerlach 236/81T) and a V-factor requiring strain, Sittich 1, from a parakeet (Piechalla et al., 1985). However, differences in production of acid from Meso-inositol, salicin and aesculin, in addition to V-factor requirement, separate Sittich 1 from V. psittacicida. Twenty-four % or less DNA reassociation was observed between the type strain of V. psittacicida and five other taxa of Pasteurellaceae (Piechula et al., 1985). According to Stenzel (1992) between 24 and 30 % DNA reassociation was observed between V. psittacicida strain 103 and P. multocida, P. stomatis, Av. Avium and G. anatis, while between 18 and 44% DNA reassociation was observed with taxon 16 of Bisgaard. The DNA G+C content of strain 103 of V. psittacicida was 39.9 mol%, while the genome mass was 1.4 x 109 Da (Stenzel, 1992). The G+C content of the type strain Gerlach 236/81T was 40.8 mol% and the genome mass 2.4 x 109 Da (Piechulla et al., 1985).

In his Inaugural-Dissertation, Beichel (1986) reported on 12 and two strains from psittacine birds sharing the phenotypical characters of V. psittacicida and V. amazonae, respectively.

A single isolate from a blunt-tailed parrot (Hinz 62) clustered with V. psittacicida and V. amazonae in the 16S rRNA phylogeny according to Christensen et al., (2009). Differences in catalase, urease and production of acid from D(-) mannitol and D(-) sorbitol indicate that Hinz 62 might represent a new species of Volucribacter.

To add new insights into the existence, diversity, and identification of Pasteurellaceae in psittacine birds a total of 37 isolates from the choana or crop of clinically affected birds were characterized and compared with the data from 16 reference strains (Gregersen et al., 2009). Phenotypical characters in addition to rpoB and 16S rRNA data classified ten isolates (Grey parrot (n=4); Budgerigar (n=3); Blue-fronted parrot (n=1); Yellow-naped parrot (n=1); red-fronted parakeet (n=1)) and two strains (Blue-fronted parrot (n=1); Yellow-fronted parrot (n=1)) with V. psittacicida and V. amazonae, respectively.

The phylogenetic relationship of partial rpoB sequences of four strains (TIS 2994 (Sulfur-crested cockatoo); TIS 15908 (Red-rumped parrot); TIS 17892 (Grey parrot); KLEIZ 10105 (Grey parrot)) demonstrated at least two groups, distantly related to Volucribacter, and obviously unrelated to the current species, and consequently labelled Volucribacter-like. These isolates showed 87.2-97.5% rpoB similarity and were closely related to four reference strains with 95.9% similarity between TIS 2494 and CCUG 19808, 94.8% between TIS 15908 and B96/37, 97.5% between TIS 17892 and 31564/2, and identity between KLEIZ 10105 and 31564/2. Since 95% rpoB gene similarity separate most species of Pasteurellaceae, while 85% similarity separate most genera (Christensen et al., 2007), it is unclear, how many species the Volucribacter-like isolates represent. 16S rRNA gene similarities supported the existence of two species within the group of Volucribacter-like organisms (Gregersen et al., 2009).

A total of 62 strains obtained from trachea of seven species of psittacine birds not affected by disease were identified by a combination of rpoB and 16S rRNA gene sequencing and by MALDI-TOF (Bisgaard et al., 2017). Nine, 15 and three strains were classified with V. psittacicida, Volucribacter spp. and taxon 34 of Bisgaard, respectively. Phylogenetic comparison of rpoB sequences allocated the nine strains identified as V. psittacicida with the type strain, demonstrating more than 95% rpoB similarity, which is within the species limit. Three strains allocated taxon 34 showed 94.8-97% rpoB similarity to the reference strain of taxon 34. Strains classified with Volucribacter spp. demonstrated rpoB similarities of 89% or higher with the type strains of V. psittacida and V. amazonae, which is within the normal limit for a genus (Christensen et al., 2007). 16S rRNA gene sequencing of a subset of 10 strains confirmed the rpoB analysis. Finally, strain 201433FT1 showed only 92% 16S rRNA similarity to Voluribacter, but 96% 16S rRNA gene similarity to the reference strain of taxon 34, in agreement with the rpoB analysis.

According to Gregersen et al. (2009) and Bisgaard et al. (2017) there is an obvious need to classify and name members of the large and heterogeneous group of Volucribacter-like organisms unrelated to V. psittacicida and V. amazonae to allow a deeper insight and understanding of these organisms associated with psittacine birds. In particular, isolates diagnosed as Bisgaard taxon 3 biovar 4 (ornithine negative P. multocida ssp. spetica; taxon 1 =G. anatis, Inositol positive G. anatis) and Bisgaard taxon 3 biovar 6 (sorbitol positive taxon 3 biovar 4) represent diagnostic problems, since all pigeon isolates belong to Gallibacterium genomospecies 3, while isolates from psittacine birds belong to Volucribacter (taxon 34) (Bisgaard and Christensen, 2019), underlining the importance of host adaptation, when it comes to identification and classification.

Taxon 34

(Volucribacter -related)
Two new taxa of Bisgaard were reported by Christensen et al. (2003). Seven and more phenotypical characters separated taxon 34 of Bisgaard from taxa 22, 26, 32 and 40. Strain 69 of taxon 34 was found closest related to strain 101 of taxon 33 with 94.5% 16S rRNA gene sequence similarity. Only differences in production of acid from D(-) mannitol separate these two taxa.
In the 16S rRNA gene sequence based phylogenetic tree strain 69 of taxon 34 clustered with strain B96/37. However, 15 phenotypic characters separate these two isolates. Strain B96/37 was isolated from a parakeet and showed the highest 16S rRNA gene sequence similarity to taxon 34 with 96.8%. In the phylogenetic tree based on 16S rRNA gene sequence comparison, Bisgaard taxon 34, isolated from Amazona sp. with septicaemia, belonged to the Avian group.

The phylogenetic relationship of 16S rRNA gene sequences of 18 V-factor dependent avian Pasteurellaceae was reported by Christensen et al. (2009). Five strains from budgerigars (n=3), a love bird, and a parakeet were closely related and supported by a bootstrap value of 69. The 16S rRNA gene sequence similarity within the group varied from 95.7% between strains B96/37 (Parakeet) and NCTC 11876 (Budgerigar) to 96.6% between strains CCUG 36050 (Budgerigar) and NCTC 11876. According to Piechulla et al. (1985) 95% DNA:DNA similarity was observed between strains NCTC 11876 and CCUG 36050. These strains should therefore be classified as a new species, although only 96.6% 16S rRNA gene sequence similarity between the strains was observed.
The low 16S rRNA gene sequence similarity observed might be explained as a result of 12 ambiguous positions in strain CCUG 36050.
The new species will probably not include the parakeet isolate, which differs phenotypically and might represent another species.

Two non V-factor dependent reference strains from a parakeet and a parrot and the suggested type strain for Bisgaard taxon 34 (69; parrot) formed a sister group supported by a bootstrap value of 51%. These two groups were linked by a bootstrap value of 86%. The highest 16S rRNA gene sequence similarity within the second group varied from 97.2% between strains 69 and B96/6 up to 98.8% between strains 69 and 31564/2sv, indicating that these three strains make up a third species.

The similarity demonstrated for the entire group, supported by a bootstrap value of 86%, varied from 94.4% between strain 69 (taxon 34) and NCTC 11876 up to 98.2% between strain B96/6 and HP 101. This level of 16S rRNA gene sequence similarity is within the normal range for most genera within Pasteurellaceae (Christensen et al., 2007). Strain NCTC 11876 and the type strains of V. amazonae and V. psittacida demonstrated 91.7% and 93.6% 16S rRNA gene sequence similarity, respectively.

The two clusters supported by a bootstrap value of 86% and tentatively designated taxon 34-like might represent a new genus.

A total of 37 isolates from psittacine birds showing digestive - and/or respiratory disorders were characterized phenotypically and by sequencing of partial rpoB and 16S rRNA genes by Gregersen et al. (2010). The results obtained were compared with the data from 16 reference strains. Four isolates (TIS 15908, Red-rumped parrot; TIS 2494, Sulfur-crested cockatoo; TIS 17892, Grey parrot; KLEIZ 10105, Grey parrot) made up one or more new groups distantly related to Volucribacter and obviously unrelated to the current species, and were consequently named Volucribacter-like. These isolates showed 87.2%-97.5% rpoB similarity and were closely related to four reference strains with 95.9% between TIS 2494 and CCUG 19808, 94.8% between TIS 15908 and B96/37, 97.5% between TIS 17892 and 31564/2, and identity between KLEIZ 10105 and 31564/2. A single isolate, KLEIZ 10105, was 16S rRNA sequenced and showed a close relationship with the two reference strains B96/37 and CCUG 19808. The 16S rRNA similarity between strain KLEIZ 10105 and 31564/2, taxon 34 and 96/6 was 95.8-100% with as much as 99.1% similarity between KLEIZ 10105 and taxon 34 of Bisgaard.

Sixty-two strains of Pasteurellaceae-like organisms isolated from trachea of two populations of healthy captive psittacine birds were identified by a combination of rpoB and 16S rRNA gene sequencing and by MALDI-TOF (Bisgaard et al., 2017).
Three strains were identified as taxon 34 of Bisgaard, while 15 were identified as Volucribacter spp.. Phylogenetic comparison of rpoB sequences showed that strains identified as Bisgaard taxon 34, Volucribacter spp. and V. psittacicida made up a single cluster with the included reference strains and type strains of V. psittacicida and V. amazonae with 100% bootstrap support.

Additional investigations are highly needed to clarify the final taxonomy of strains classified as Bisgaard taxon 34, Volucribacter spp., and Volucribacter-like organisms.

Taxon 35

Misclassified P. multocida from dogs not yet published.

Taxon 36

Mannheimia varigena (Additional information is given under taxon 15).

Numerical taxonomy was used by Angen et al. (1997 (1)) to reinvestigate the phenotypic relationship within the ruminant [Pasteurella] haemolytica complex. A simplified dendrogram of all strains investigated based upon average Euclidian distances and UPGMA showed that the glycoside positive strains differ from the rest of the strains, forming a separate cluster. Two bovine isolates producing acid from glycosides and isolated from bovine mastitis and vaginal exudate formed an outgroup, and was tentatively designated taxon 36. Subsequent studies including taxon 15, 18 and 20 confirmed these observations (Angen et al., 1997 (2)).

Ribotyping of the [P.] haemolytica-complex isolated from cattle, sheep, deer, pigs, hares and rabbits using HindIII for digesting DNA showed that taxon 36 clustered with bovine isolates of [P.] haemolytica biogroup 6 (Angen et al., 1997 (3)).

A polyphasic investigation of the taxonomy of the trehalose-negative [P.] haemolytica complex demonstrated that strain H39 of taxon 36 showed 88% DNA:DNA binding with the type strain of Mannheimia varigena, suggesting that taxon 36 should be classified with this species. However, several phenotypical characters, including production of acid from glycosides, separate taxon 36 from M. varigena. Differences in haemolysis, L(+) rhamnose, maltose, dextrin, and ONPX separate taxon 36 from M. caviae, while differences in gelatinase, D(-) arabinose, D(-) sorbitol, L(+) rhamnose, D(-) melibiose and α-galatosidase separate taxon 36 from M. glucosida. Finally, the third glycoside positive species, M. granulomatis, can be separated from taxon 36 in ornithine decarboxylase, L(+) arabinose, D(+) xylose, D(-) sorbitol, L(+) rhamnose, D(-) melibiose, α-galactosidase, PGUA, and ONPF.

Additional genetic characterization of taxon 36, including whole genome sequencing, is indicated, since DNA:DNA hybridization only has been carried out with the type strain of M. varigena.

Taxon 37

Isolates from psittacine birds (90) classified as Psittacella sp. (170).

Taxon 38

(Atypical isolates of P.stomatis?)
Investigations on the relationships among Pasteurellaceae obtained from free ranging chickens and their animal contacts in Tanzania demonstrated that four unclassified surcrose negative isolates from the oral cavity of dogs made up a separate phenotypic cluster (IIC), unrelated to known Pasteurellaceae spp. and Bisgaard taxon 16 (Frederiksenia canicda) (Muhairwa et al., 2001). Ribotyping using HpaII for digestion of DNA showed that these strains clustered deeply with the type strain of P. stomatis NCTC 11623T demonstrating some 45% similarity, while strains classified as P. canis only demonstrated some 35% similarity.

Differences in sucrose and α-fucosidase (ONPF) separate 14 strains genetically approved to belong to Canicola haemoglobinophilus (Christensen et al., 2021 submitted for publication) from taxon 38. Similarly differences in ornithine decarboxylase and acid from sucrose separate taxon 38 from the type strain of P. canis NCTC 11621T, while differences in acid from sucrose and trehalose, in addition to α-glucosidase (PNPG), separate taxon 38 from the type strain of P. stomatis NCTC 11623T.

Strain KD14C of taxon 38 was closest related to Pasteurella canis by 16S rRNA gene sequence comparison (99.3%), however, taxon 38 diverged from P. canis in ornithine decarboxylase and in acid formation from sucrose. In the 16S rRNA gene sequence based phylogeny, taxon 38 also clustered with the type strain of P. canis. A high 16S rRNA similarity between species belonging to the same genus has previously been discussed under taxon 33. All five strains of taxon 38 characterized shared rpoB sequence and showed 99% similarity to P. stomatis, but only 94, 93 and 90% similarity to P. multocida, P. dagmatis and P. canis, respectively. All strains were also identified as P. stomatis by MALDI-TOF MS (Kuhnert 2021, personal information). Further investigations, including whole genomic sequence comparisons are, however, needed to classify taxon 38 properly at the species level.

Taxon 39

Bovine isolates classified with Mannheimia (104).

Taxon 40

According to Christensen et al. (2003), phylogenetic analysis by 16S rRNA gene sequence comparison of avian taxa of Bisgaard demonstrated the existence of two new taxa, provisionally named Bisgaard taxon 34 and 40, respectively. Six isolates obtained from the respiratory tract of seagulls made up taxon 40. Some phenotypic variation was observed for taxon 40. Three strains produced acid from maltose and dextrin, but not from D (-) sorbitol, while three isolates produced acid from D (-) sorbitol, but not from maltose and dextrin.
Strain B 301529/00/1 of taxon 40 (D (-) sorbitol positive) was found rather distantly related to all other taxa of Pasteurellaceae with the closest similarity to [P.] testudines with 93% 16S rRNA gene sequence similarity.
In the phylogenetic tree based on maximum likelihood analysis of 16S rRNA gene sequences, taxa 40 and 32 of Bisgaard made up a monophyletic group with the  previously reported taxon 14, which also included the turtle associated [P.] testudines, formerly described as the Testudinis group. The maximal variation between the avian members within the Testudinis group was 7.8% as observed between taxon 40 and taxon 32.

Two groups within the phylogenetic tree contained taxa mainly isolated from birds, the rest being associated with mammals.
The two major monophyletic bird-associated groups named “Avian” and “Testudinis”, included 11 and four species-like taxa, respectively. Three of the four taxa within the “Testudinis” cluster represented avian taxa, and a predominantly avian origin of this group might be assumed. It was also assumed that the “Avian” and “Testudinis” clusters were independently associated Aves as divergence within these groups were 8.5 and 7.8%, respectively, which is lower than the maximal difference between these groups of 9.7%, indicating that the colonization of Aves or its ancestor might have happened independently by these groups.

In 2019 Knowles et al. reported on a mortality event in Rhinoceros Auklets (Cerorhinca monocerata) in Washington, USA, from which taxon 40 was isolated in pure culture. Lesions included pneumonia and pleuritis accompanied by septicemia.
However, the authors were uncertain whether taxon 40 acted as a primary pathogen or simply as an opportunistic pathogen.

Taxon 41

(Rodentibacter myodis)
Considering the wide range of taxa of Pasteurellaceae which have been reported in rodents and other laboratory animals, and the difficulties associated with proper identification and classification of these organisms before molecular methods were implemented, published data based upon too few characters should be interpreted with great care. For the same reason, the reservoir and significance of the taxa have remained uncertain.
A large comparative investigation including some 1000 isolates mainly form Muridae was carried out by Adhikary et al. (2017). Based upon phenotypic characterization, partial sequencing of the rpoB gene of 346 strains was performed. A total of 344 strains selected based on the rpoB gene sequence comparison were subsequently 16S rRNA gene sequenced. Whole genome sequencing included 24 strains, from which DDN’s infB gene sequences, mol% G+C of DNA and genome size were deducted.

Phylogenetic analysis of rpoB and 16S rRNA gene sequences identified a monophyletic group of three strains with 100% bootstrap support. These strains belonged to taxon 41 of Bisgaard and shared rpoB gene sequence similarities of 98.0-99.0%, and 89% or less to other taxa. 16S rRNA gene similarities were 98-100% within the group, while the highest level of similarity outside the group was 96.4% to R. pneumotropicus.
DDN deducted from the whole genomic sequences was 24% or less with other species compared within the genus Rodentibacter. Taxon 41 was proposed as R. myodis sp. nov., named after the animal from which these organisms were isolated, Myodes glareolus (bank vole).
The type strain is Ac151T (CCUG 69666T), and the DNA G+C content of the type strain is 41.0 mol%, while the genome size is 2.5 Mb determined by whole genome sequencing (Adhikary et al., 2017).

Taxon 42

Porcine isolates of [P.] caballi (151).

Taxon 43

Melibiose negative and L (+) arabinose positive A. equuli. These organisms might represent Mannheimia isolates from horses not yet published.

Taxon 44

Psittacine isolates not yet published (B96/3 and B96/4 etc.).

Taxon 45 and 46

(Taxon 45: A new sucrose negative species of Pasteurella sensu stricto, which remains to be named)
(Taxon 46: A new, not yet named species of Pasteurella sensu stricto)

In diagnostic tables phenotypical characters are normally stated as + (90%or more of the isolates positive within 14 days), - (90% or more of the isolates negative within 14 days) or d (less than 90%, but more than 10% positive within 14 days). In addition, key characters are often suggested for separation of specific taxa.

Phenotypic properties of Pasteurella multocida are very diverse, making diagnostics based upon classical phenotypic characterization quite problematic (Heddleston, 1976; Christensen et al., 2002). Add to this that P. multocida has been isolated from a multitude of hosts (Bisgaard, 1993; Frederiksen, 1993).

Key characters for identification of Pasteurella spp. often include urease, ornithine decarboxylase and fermentation of dulcitol, D (-) mannitol, D (-) sorbitol, maltose, sucrose, trehalose and dextrin, in addition to production of α-glucosidase (PNPG) (Mutters et al., 1985; Christensen et al., 2012). However, these characters, in addition to a complete set of phenotypic characters, did not allow a safe phenotypic separation of either the A (P. multocida ssp. multocida and gallicida) – or B (P. multocida ssp. septica) lineages outlined by multilocus sequence analysis of P. multocida, calling into question the use of phenotypic characters for identification, and the validity of subspecies within P. multocida (Bisgaard et al., 2012). Similarly, phenotypic characterization of isolates from the feline oral cavity classified these isolates with P. dagmatis, although 16S rRNA – and sodA gene sequencing clearly showed that these isolated should be classified as a new genomospecies within Pasteurella sensu stricto (Sellyei et al., 2010).

Among 1,268 cultures of P. multocida characterized by Heddleston (1976), none were reported sucrose negative. However, 17 sucrose negative P. multocida out of some 1500 isolates were reported by Bisgaard in Christensen et al. (2005). For the same reason, P. multocida is listed as sucrose negative, since 90% or more of the isolates are sucrose negative. Schmid (1987) reported on six sucrose negative isolates from ovine nasal swabs in Syria, while Muhairwa et al. (2001) reported on sucrose negative isolates from dogs, subsequently named taxon 38 and classified with Pasteurella sensu stricto (See taxon 38). With the exception of lack of Carter capsular antigens, further information on the sucrose negative ovine isolates were not stated by Schmid (1987).

To validate the identification and classification of sucrose negative P. multocida, the 17 isolates of Bisgaard (see above) were phenotypically and genotypically characterized. Isolates were obtained from bovine pneumonia (n=6), sheep (n=2), large cat bites in humans (Tigers, n=3; Lions, n=2; Leopards, n=2), and a chipmunk (n=1), in addition to an isolate of unknown origin. For comparison the type strains of the three subspecies of P. multocida and four bovine isolates identified as P. multocida ssp. septica were included.

With the exception of seven sucrose negative isolates from tiger (n=3) and lion (n=2) bite wounds in humans, a chipmunk, and an unknown source, and two isolates from leopard bite wounds in humans, phenotypic characters obtained were in accordance with P. multocida as reported by Mutters et al. (1985). The seven and two sucrose negative isolates were named taxon 45 – and 46 of Bisgaard, respectively.

Ribotyping using HpaII for digestion of DNA showed that taxon 45 and 46 shared at least four of 13 positions demonstrated, while a single band was common to all isolates.

Comparison of 16SrRNA gene sequences showed that sucrose positive – and negative isolates might share identical 16S rRNA gene sequences.

The phylogenetic analysis based on 16S rRNA gene sequence comparison showed that all strains of P. multocida, including the sucrose-negative variants of both P. multocida and taxon 45, were monophyletic, while the two isolates of taxon 46 clustered with the type strain of P. dagmatis. The level of 16S rRNA gene sequence similarity within the monophyletic group was 98.4% or higher, and the level of similarity was at least 98.6% between all the strains and the type strain of P. multocida. Isolates classified as taxon 45 and the type strain of P. multocida spp. septica showed at least 99.4% 16SrRNA gene sequence similarity.

The two isolates of taxon 46 formed a monophyletic group with the type strain of P. dagmatis if only the type strain of P. multocida ssp. septica was included. The highest degree of similarity found for strain CDC A996 (taxon 46) was to strain F4646 of taxon 46 (99.6%) and to the type strain of P. dagmatis (98.4%). At most, 97.4% similarity was found to the sucrose negative – and positive strains of P. multocida investigated.

infB and rpoB sequence comparison demonstrated that strains provisionally named taxon 45 formed their own monophyletic group, while isolates classified as taxon 46 and represented by strain CDC A996 did not cluster with P. dagmatis.

Isolates classified as taxon 45 all showed DNA:DNA binding values below the species level (31-51%) with the strains representing the three subspecies of P. multocida, while 90% DNA reassociation was observed between two isolates of taxon 45. Surprisingly, all strains of taxon 45, except a single isolate, tested positive in the P. multocida PCR test of Miflin and Blackall) (2001), while capsular typing resulted in variable reactions for this taxon.

ITS analysis confirmed the existence of two groups of sucrose negative isolates (Bisgaard taxon 45 and 46) recognized by ribotyping, rpoB and infB sequence comparison, and DNA:DNA hybridization.
A third group of sucrose negative P. multocida-like isolates, mainly associated with bovine pneumonia, however, belonged to P. multocida sensu stricto.

Based upon the data stated above, Christensen et al. (2005) concluded that two groups of sucrose negative isolates, taxon 45 and taxon 46, represents two new species of Pasteurella sensu stricto, of which taxon 45 mainly is obtained from large cat bites in humans. Phenotypically, a combination of polyamine patterns (Busse et al., 1997), fermentation of sucrose and possible gas production from D (+) glucose might be used for separation of taxon 45 from other species of Pasteurella sensu stricto.

Strain SSI P876, obtained from a lion bite, is suggested as the type strain for taxon 45. The G+C contents of strains SSI P876 and Schmid 351 were 37.2 and 41.8 mol %, respectively, while the genome masses were 1.8 and 1.7 GDa, respectively (Eckert et al., 1981; Stenzel, 1992).

Two other sucrose negative strains obtained from bite wounds caused by leopards also formed a new taxon (46) of Pasteurella sensu stricto. These organisms are urease positive and might be misidentified as P. dagmatis. However, taxon 46 differs from the type strain of P. dagmatis in acid production from D (-) sorbitol, sucrose, and trehalose.

Whole genome sequencing remains to be carried out before naming taxon 45 and 46. This might possibly also explain the positive PCR tests for P. multocida stated above.

Taxon 47

(Remains to be classified and named)
Schulz et al. (1977) reported on three strains obtained from albino rats suffering from pneumonia. Similar organisms have been obtained from rats in Denmark. These organisms are phenotypically related with P. dagmatis, Av. gallinarum and Bisgaard taxon 16. However, their host reservoir is different. For the same reason, the rat isolates have been named taxon 47. A single rat isolate, Rat 44sv., differs from the rest of the group in urease, indole and D (+) xylose.

Taxon 47 differs from Muribacter muris in tween 80, D (+) galactose, dextrin, and arbutin (Nicklas et al., 2015). However, only a single character separates taxon 47 from Rodentibacter pneumotropicus, R. heidelbergensis and R. genomospecies 1 (Adhikari et al., 2017).

According to Ryll (1989) strains NCTC 11051 (Rat A/1a)  and NCTC 11052 (Rat A/4a) demonstrated 96% DNA:DNA binding, while only 36% binding was observed between NCTC 11051 and the type strain of Muribacter muris NCTC 12432T. Similarly, only 32% DNA:DNA binding was observed between NCTC 11052 and [H.] influenzae muris NCTC 11146 of Csukás, indicating that taxon 47 is likely to be classified outside the genus Muribacter. Finally, the R13 strain of Mannheim (HIM 974-2) demonstrated 70 and 68% DNA:DNA binding with strains NCTC 11051 and NCTC 11052, respectively, indicating that this strain also belongs to taxon 47.

Although only a single phenotypical character separates taxon 47 from R. pneumotropicus, DNA:DNA bindings of only 2% between the type strain NCTC 8141T of R. pneumotropicus and NCTC 11051 clearly shows that these taxa are unrelated (Ryll, 1989).

16S rRNA gene sequence comparison showed that NCTC 11051sT and the deviating strain R44 sv. of taxon 47 were monophyletic in the phylogenetic analysis with a bootstrap support of 100%. This group was connected to M. muris (CCUG 16938T), Bisgaard taxon 27, and [H.] influenzae-murium NCTC 11146 by a bootstrap value of 76% (Christensen and Bisgaard, unpublished data).

The DNA G+C content of the suggested type strain of taxon 47, NCTC 11051sT, is 42.8% and the molecular size 1.6 x 109 D (Ryll, 1989).

Taxon 48

(Genomospecies 1 of Rodentibacter)
Five strains (Ppn 415, Ppn 416T, Ppn 419, Ppn 426 and Ppn 428) obtained from Apodemus sylvaticus and Mus musculus and received as [P.] pneumotropica, deviated from the type strain of [P.] pneumotropica type Jawetz NCTC 4181T in seven phenotypic characters, and were therefore allocated a separate group, designated taxon 48.

Phylogenetic analysis of rpoB and 16S rRNA gene sequences identified 46 strains as monophyletic with 100% bootstrap support (Adhikary et al., 2017). The group shared a rpoB gene sequence similarity of 98.5-100% with 96.8% or less similarity to other groups, and 16S rRNA gene sequence similarity of 98-100% within the group, and 96% or less similarity to other groups. Only a single phenotypical character separated the group from R. heylii (D (+) xylose), R. rarus (Glycerol), and R. trehalosifermentans (Indole), and taxon 48 was left as an unnamed genomospecies 1.

Most isolates classified with genomospecies 1 were obtained from Apodemus spp.

The DNA G+C content of the type strain Ppn 416T was 41.0 mol%, and the genome size 2.4 Mb as determined from whole genomic sequencing.

Taxon 49

(Avibacterium volantiuim)
Knowledge on the occurrence of Pasteurellaceae in wild and domesticated partridges and pheasants is only fragmentary to the knowledge of the author. However, one might speculate, if the taxa of Pasteurellaceae affecting chickens also may affect partridges and pheasants, since they all belong to galliforme birds

Grebe and Hinz (1975) reported on two isolates from pheasants, which were classified with Haemophilus, while none of the 156 isolates of Pasteurellaceae reported by Peters (1989) originated from partriges or pheasants. Only a single isolate of P. multocida out of 130 Pasteurellaceae from birds, originated from a pheasant (Beichel, 1986).

Two isolates from the trachea (P. sp 29) and sinusitis (55.000) in pheasants were reported by Petersen et al. (1998). Both strains, however, remained unclassified.

Five strains of P. multocida and a single isolate of Av. gallinarum from partridges and two isolates of P. multocida and a single isolate of Av. gallinarum from pheasants exist in the authors strain collection together with five strains of taxon 3 biovar 3 from partridges (n=2) and pheasants (n=3). In addition, a single trehalose positive isolate of taxon 3 biovar 6 from a partridge exists om the collection. All isolates were associated with lesions. Finally, several isolates of taxon 14 from pheasants exist in the strain collection, a few of which have been published (Bisgaard and Mutters, 1986).

Three strains from pheasants (P. sp. 29, P 40, and B69/96/3) and a single partridge strain (C67) were subjected further characterization by Christensen et al. (2003). Strain B69/96/3 was phenotypically related with Av. gallinarum and taxon 14. Strain B69/96/3, however, only demonstrated a high 16S rRNA gene sequence similarity to Av. gallinarum (99.5%), while only 92% similarity was observed to taxon 14. This strain therefore represents a maltose- and dextrin negative isolate of Av. gallinarum. Strains C67 (partridge) and P. sp. 29 (pheasant) demonstrated 99% 16S rRNA gene sequence similarity, and P. sp. 29 showed 97.2 and 97.1% 16S rRNA gene sequence similarity with Av. sp. A and Av. paragallinarum, respectively. However, several phenotypic characters separate P. sp. 29 and C67 from the type strains of Av. paragallinarum and Av. sp. A. For the same reason both strains remained as an unclassified Avibacterium sp. Finally, strain P40 (pheasant) showed 99.9% 16S rRNA gene similarity to HPA7 of taxon 14, and was consequently classified as an ornithine decarboxylase positive strain of taxon 14.
A partridge strain (F296) and a pheasant strain (39199/1) representing biovar 3 of taxon 3 made up a separate phenon 3a, when the protein profiles of 37 strains representing all biovars of the taxon 2 and taxon 3 of Bisgaard were compared (Bisgaard et al., 1993)

These strains also made up a separate AFLP cluster 7, when a subset of 60 strains of the taxon 2 and 3 complex was investigated by amplified fragment length polymorphism (AFLP) (Bojesen et al., 2007).

A polyphasic study of 23 strains, finally classified the taxon 2 and taxon 3 complex with Gallibacterium, including five new species of which two remained unnamed (Bisgaard et al., 2009). Strains F296 and 39199/1, represented by 39199/1, made up group V, which showed a maximum of 96.6% 16S rRNA gene sequence similarity with the other groups. Whole-genome similarities calculated from recN sequences confirmed that this group should be classified as a new species of Avibacterium. Since only two strains have been investigated so far, the authors desisted from naming group V (Bisgaard et al., 2009).

Christensen et al. (2009), finally characterized three strains from pheasants (46671, 47483/5 and HP202) and a single strain from a partridge (84611) all of which were associated with lesions. All strains were V-factor dependent.
All four strains were classified with Avibacterium according to 16S rRNA analysis. However, 16S rRNA gene sequence analysis does not allow separation of all species within the genus (Blackall et al., 2005).

Phenotypically, all four strains shared the phenotypic characters of Av. volantium (Mutters et al., 1985) and demonstrated 98.6 (46671), 99.4 (47483/5), 98.0 (HP 202), and 99.1% (84611) 16S rRNA similarity with the type strain of Av. volantium, and should therefore by classified with this species. According to Hinz (Hinz, 1975; Mutters et al., 1985), most strains of [H.] avium appear as harmless mucosal commensals, which are spread via direct contact or airborne transmission. Some of the L (+) arabinose positive strains, however, may cause sinusitis and edema of the head.

Since all four strains were L (+) arabinose negative and obtained from coryza-like disease, they were believed to be different from Av. volantium based upon their disease potential. For the same reason, these organisms were classified as taxon 49, the final classification of which awaits whole genome sequencing.

Taxon 50

Chicken isolates classified as Av. endocarditidis (164).

Taxon 51

(Remains to be classified and named)
As discussed previously under taxon 45 and taxon 46 of Bisgaard, the phenotypic properties of P. multocida are very diverse, making diagnostics based upon classical phenotypic characterization quite problematic.

Two strains identified as P. multocida and isolated from pharynx of a long-tailed field mouse (Pm110) and trachea of a common European white-toothed shrew (Pm 111) were received for further characterization in 1999. Extended phenotypical characterization showed that they shared the phenotypical characters examined, and that they differed from P. multocida in lysin decarboxylase, meso-inositol, α-fucosidase (ONPF), and β-glucuronidase (PGUA).

Differences in urease, production of acid from D (-) sorbitol and maltose, in addition to α-fucosidase (ONPF) separate the strains from the genus characters of Rodentibacter (Adhikary et al., 2017).

Finally, a total of nine, ten, and thirteen phenotypic characters separate the organisms from M. muris, taxon 47, and taxon 27, respectively.

Based upon phenotypical characters obtained, classification with known taxa was not possible, and the strains were therefore designated taxon 51 to improve attention to and identification of these organisms.

In the 16S rRNA gene sequence phylogeny, taxon 51 clustered with taxa belonging to Rodentibacter, although differing from this genus in four phenotypical characters, as stated above.

To classify and name taxon 51, additional investigations are needed, including whole genome sequencing and deduction of additional conserved gene sequences for comparison, in addition to DNA:DNA hybridizations.

Taxon 52

(Remains to be classified and named)
Strains received as Pasteurella sp. or Actinobacillus sp. and obtained from nasopharynx of apparently healthy Clethrionomys glareolus have remained unclassified so far.

Four to seven phenotypical characters separate all Rodentibacter species from these organisms, tentatively designated taxon 52. Six phenotypical characters separate taxon 52 from Muribacter muris and taxon 27, while only differences in acid production from D (+) arabitol and lactose separate taxon 47 and taxon 52.

In the 16S rRNA gene sequence based phylogenetic tree, taxon 52 clustered distantly with the type strains of A. lignieresii, M. haemolytica and Bisgaard taxon 5, indicating that these organisms might represent a new genus.

To finally classify and name taxon 52, additional investigations are needed, including whole genome sequencing and deduction of more conserved gene sequences for comparison with other taxa, in addition to deducted DNA:DNA bindings with relevant taxa.

Taxon 53 and 54

Taxon 53, unrelated to any known groups of Pasteurellaceae
Taxon 54, unrelated to any known groups of Pasteurellaceae

Grebe and Hinz (1975) reported on the occurrence of a diverse group of bacteria obtained from various avian species and classified with Haemophilus. Subsequently, Hinz and Kunjara (1977) described a new Haemophilus species from chickens, [H.] avium, which is unable to cause infectious coryza in chickens.

According to Piechulla et al. (1985) several previously unreported bacterial strains from budgerigars, pigeons, kestrels, and a goose made up four groups, respectively, which should be regarded as new species of Haemophilus, since they can be separated from each other and from recognized species of the family Pasteurellaceae.

DNA-DNA hybridizations showed that [H.] avium included three DNA homology groups, which are genetically closer to P. multocida, the type species of the genus Pasteurella than to H. influenzae, the type species of the genus Haemophilus. Three new species of Pasteurella, [P.] avium comb. nov., [P.] volantium sp. nov. and [P.] sp. A, were therefore proposed for strains formerly identified as [H.] avium (Mutters et al., 1985).

A total of 155 isolates of V-factor dependent Pasteurellaceae obtained from poultry, cage- and wild birds were reported by Peters (1989), who included the following reference strains for comparison: [P.] avium IPDH 0002 (chicken), [P.] volantium HIM 841-3 (chicken), [P.] sp A IPDH 280 (chicken), IPDH 312 (NCTC 11877) from a pigeon, IPDH 2-76 from a kestrel, and IPDH 195 (NCTC 11876) from a budgerigar. Eighty isolates were classified with species belonging to Avibacterium, and 20 isolates with the taxon 2 and 3 complex, while 18 isolates were related to Volucribacter. The remaining 37 isolates were grouped as Haemophilus-like, among which 21 and eight were related to the taxa from budgerigars and pigeons, respectively, mentioned previously (Piechulla et al. 1985). Finally, Haemophilus-like organisms have been reported from salpingitis in geese according to Bisgaard (1995).

The phylogenetic relationships of unclassified, V-factor dependent Pasteurellaceae obtained from different species of birds, were investigated by 16S rRNA gene sequence comparison by Christensen et al. (2009).
Two isolates 19987/1 Salp (goose) (Bisgaard, 1995) and Peters 120 (1989) (MCCM 1291) (Budgerigar group) were unrelated to any known groups of Pasteurellaceae and tentatively designated taxon 53 and taxon 54, respectively. The highest similarity observed for taxon 53 was to the type strain of H. pittmaniae with 94.5%, while the highest similarity of taxon 54 was to the type strain of Nicoletella semolina with 95.1%. The 16S rRNA gene sequence similarity between taxon 53 and taxon 54 was 94%.

Both taxon 53 and 54 meet the phenotypical characters of Pasteurellaceae (Christensen et al., 2020). In addition, the phenotypic characters of taxon 53 are in accordance with those of the genus Avibacterium, while only differences in requirement for V-factor separate taxon 53 from taxon 35. Similar phenotypic characters were observed for taxon 54 and the genus Aggregatibacter (Christensen et al., 2020). However, the data for Aggregatibacter include different reactions in seven characters in addition to lacking data for another seven characters. A single character separates taxon 54 from the genera Actinobacillus (urease), Avibacterium (D (-) sorbitol) and Gallibacterium (V-factor).
Differences in acid production from meso-inositol, D (-) sorbitol, D (+) galactose and maltose separate taxon 53 and taxon 54.

Interestingly, the three V-factor dependent isolates from geese belonged to three different genera, two of which remains to be defined and named.

Christensen et al. (2001) presented a detailed report on the trend of describing novel taxonomic units based on a very small number of isolates or even on a single isolate. A subsequent survey by Felis and Del- laglio (2007) revealed that too many taxonomic descriptions are still based on a single or a few strains, which is not considered good taxonomic practice, since the descriptions based on a single isolate might result in a highly biased scenario. For the same reasons, Christensen et al. (2009) desisted from naming taxon 53 and taxon 54, although they represent new genera of Pasteurellaceae.

Taxon 55

(Atypical R. ratti)
This group originally included strains from apparently healthy rats and a single strain, Beichel 28 (Beichel, 1986), obtained from otitis externa of a chicken, and received as [P.] pneumotropica type Heyl.

These isolates differed from Bisgaard’s taxon 22, originally associated with chickens, in L (+) arabinose, D (-) arabinose and L (-) fucose (all negative), and were believed to represent a separate taxon from chickens,  named taxon 22-like (taxon 55).

Subsequent DNA:DNA hybridizations (Ryll, 1989; Ryll et al., 1991) demonstrated that the type strain of taxon 22 (F75T) and Taxon B of Kilian, Mannheim R1 and Rat A5/a (taxon 55), Nicklas 78/1, and Wüllenweber-Schmidt 456a showed 74-94% DNA:DNA binding, while only 8% binding was observed between Rat A5/a and rat A1/a (NCTC 11051) of taxon 47. These observations clearly document that the phenotypical differences observed between taxon 22 and taxon 55 do not have a taxonomic impact.
Both taxa were subsequently classified as Rodentibacter ratti by Adhikari et al. (2017).

Taxon 58

Unrelated to any known genera of Pasteurellaceae
Infections in pigeons by Pasteurellaceae have mostly dealt with P. multocida, although routine laboratory diagnosis of this organism does not exclude the incorporation of other taxa of Pasteurellaceae (Biberstein et al., 1991).

Grebe and Hinz (1975) reported on a [Haemophilus] sp. from rhinitis in pigeons. Subsequently, Piechulla et al. (1985) described two strains, IPDH 312 and IPDH 230, from the respiratory tract and liver of pigeons, which demonstrated distinct cultural, morphological, and biochemical characteristics. Since V-factor was required for growth, these isolates were tentatively assigned to the genus Haemophilus Windslow et al. 1917. These isolates demonstrate 96% DNA-DNA binding with the highest DNA-DNA binding of 37% outside the group to strain IPDH 195 from a budgerigar. The G+C content of DNA from the two pigeon isolates varied between 37.0-39.4 mol%, while the genome mass was 1.6 x 109d for both strains. De Ley et al. (1990) located IPDH 230 at the common root for the rRNA branches of P. multocida, A. lignieresii, H. influenzae, [H.] aphrophilus and [A.] actinomycetemcomitans.

Sixty-seven isolates of Pasteurellaceae tentatively designated taxon 3 were reported by Bisgaard (1982). These isolates originated from ducks (n=21), a goose, and pigeons (n=45). Eleven of the pigeon isolates originated from purulent inflammations in the pharynx.

Eighteen out of 130 strains reported by Beichel (1986) originated from pigeons. These strains were classified with P. multocida (n=1) and the taxon 2 and 3 complex of Bisgaard (n=15), while two strains remained unclassified.

Phenotypic characterization of 156 V-factor dependent avian Pasteurellaceae was reported by Peters (1989). Fourteen strains from pigeons were classified as [H.] avium (n=2), Volucribacter sp. (n=1), and Taxon 3 (n=1), while ten strains made up a heterogenous group differing from the organisms reported by Piechulla et al. (1985) in several phenotypic tests.

Based upon differences in production of indole, urease and production of acid from L (+) arabinose, dulcitol, D (-) sorbitol and D (+) melibiose 19 biovars were recognized within the taxon 2 and 3 complex associated with ducks, geese, pigeons, partridges, pheasants and psittacine birds (Bisgaard, 1993).

Thirty-seven strains representing all biovars established within the taxon 2 and taxon 3 complex of Bisgaard were characterized by one-dimensional SDS-PAGE of cellular proteins (Bisgaard et al., 1993).
Numerical analysis of protein patterns obtained resulted in six major and 12 minor groups (phena). Nine pigeon strains made up phenon 2b, while five out of nine strains in phenon 2c originated from pigeons. These phena demonstrated more than 80% similarity. With a few exceptions, a connection was demonstrated between protein profiles and hosts from which the strains belonging to the respective phena originated. A correlation between protein profiling and previous DNA-DNA hybridizations was, however, not observed (Bisgaard et al., 1993).

Fifteen pigeon strains obtained from lesions were characterized by Bisgaard et al. (1999). Thirteen strains were classified with taxon 3 biovar 1 (n=1), biovar 2 (n=1), biovar 4 (n=5), biovar 5 (n=3), and biovar 6 (n=3), while two strains belonged to taxon 14. Ribotyping using HpaII for digestion of DNA and subsequent cluster analysis showed that five pigeon isolates representing taxon 3 biovars 1, 2, 4, 5 and 6 made up a separate cluster, different from isolates of taxon 3 and taxon 2/3 from budgerigars (Bisgaard et al., 1999).

A subset of 60 strains of the taxon 2 and 3 complex, selected to include the broadest diversity with regard to phenotype and host spectrum, was investigated by amplified fragment length polymorphism (AFLP) (Bojesen et al., 2007). Results obtained suggested a statistically clear association between AFLP clusters and the host species families Columbidae, Anatidae, and Psittacidae, confirming the results of whole cell protein profiling. However, a subsequent polyphasic taxonomic study including AFLP clusters 1, 4, 6 and 7 demonstrated that isolates from ducks (taxon 3 biovar 2 and 4), a turkey (taxon 3 biovar 7), a parakeet (taxon 3 biovar 2) and a budgerigar (taxon 2/3 biovar 2) should be classified with the pigeon isolates as Gallibacterium genomospecies 3 (Bisgaard et al., 2009).

A total of 76 strains from pigeons (n=69) and other birds (n=7) representing biovar 1-8 and 10 of taxon 3 were finally characterized genetically by partial sequencing of the rpoB gene (Bisgaard and Christensen, 2019). With the exception of two pigeon isolates classified as P. multocida (taxon 3 biovar 7 and 10), all the other isolates from pigeons (n=67) were classified with G. genomospecies 3. Similarly, isolates from a duck, two turkey isolates, and a canary isolate were classified with G. genomospecies 3, while three isolates from a budgerigar (taxon 3 biovar 4), a parakeet (taxon 3 biovar 4) and a partridge (taxon 3 biovar 6) surprisingly were classified as taxon 44, Volucribacter sp., and taxon 14, respectively, underlining the importance of the host and the uncertainty of the phenotype.

Finally, Christensen et al. (2009) characterized four V-factor dependent strains from Australian (HP 288 and HP 289) and German pigeons (IPDH 312 and Peters 143), one of which (IPDH 312) had been previously investigated by Piechulla et al. (1985). These four strains formed a distinct group within the Testudinis 16S rRNA group, supported by a bootstrap value of 100%. The 16S rRNA gene sequence similarity within this group varied from 98.2% between the German strains to 99.6% between the two Australian strains. This group was only distantly related to taxon 40 of Bisgaard, demonstrating between 93.7% (CCUG 18783) and 94.5% (Peters 143) similarity with B 301529/00/1 of taxon 40.

The overall 16S rRNA gene sequence similarity for the pigeon group and the distance demonstrated to related taxa clearly indicate that strains belonging to this group represent a new species of Pasteurellaceae from a genotypic standpoint. Add to this that 96% DNA-DNA similarity was observed between IPDH 312 and another pigeon isolate (IPDH 230), and the G+C content of these strains varied between 37.0 and 39.4 mol% as stated previously.

Phenotypically, the pigeon group shared the characters of Pasteurellaceae as stated by Christensen et al. (2020a). However, the pigeon group also shares the phenotypes of the genera Avibacterium and Aggregatibacter, while only urease separate the group from the genus Actinobacillus, and V-factor from the genus Gallibacterium. However, four to seven phenotypic characters separate the pigeon group from related taxa (Christensen et al., 2009).

A final classification and naming of the pigeon group depend on the results deducted from whole genome sequencing of the type strain or descriptions of additional strains. Meanwhile, this group is tentatively designated taxon 58 to allow unambiguous communications on this particular group of V-factor dependent organisms from pigeons.