OUP user menu

Pneumocystis oryctolagi sp. nov., an uncultured fungus causing pneumonia in rabbits at weaning: review of current knowledge, and description of a new taxon on genotypic, phylogenetic and phenotypic bases

Eduardo Dei-Cas, Magali Chabé, Raya Moukhlis, Isabelle Durand-Joly, El Moukhtar Aliouat, James R. Stringer, Melanie Cushion, Christophe Noël, G. Sybren de Hoog, Jacques Guillot, Eric Viscogliosi
DOI: http://dx.doi.org/10.1111/j.1574-6976.2006.00037.x 853-871 First published online: 1 November 2006


The genus Pneumocystis comprises noncultivable, highly diversified fungal pathogens dwelling in the lungs of mammals. The genus includes numerous host-species-specific species that are able to induce severe pneumonitis, especially in severely immunocompromised hosts. Pneumocystis organisms attach specifically to type-1 epithelial alveolar cells, showing a high level of subtle and efficient adaptation to the alveolar microenvironment. Pneumocystis species show little difference at the light microscopy level but DNA sequences of Pneumocystis from humans, other primates, rodents, rabbits, insectivores and other mammals present a host-species-related marked divergence. Consistently, selective infectivity could be proven by cross-infection experiments. Furthermore, phylogeny among primate Pneumocystis species was correlated with the phylogeny of their hosts. This observation suggested that cophylogeny could explain both the current distribution of pathogens in their hosts and the speciation. Thus, molecular, ultrastructural and biological differences among organisms from different mammals strengthen the view of multiple species existing within the genus Pneumocystis. The following species were subsequently described: Pneumocystis jirovecii in humans, Pneumocystis carinii and Pneumocystis wakefieldiae in rats, and Pneumocystis murina in mice. The present work focuses on Pneumocystis oryctolagi sp. nov. from Old-World rabbits. This new species has been described on the basis of both biological and phylogenetic species concepts.

  • Pneumocystis
  • pneumocystosis
  • Pneumocystis taxonomy
  • Pneumocystis phylogeny
  • Pneumocystis morphology
  • Pneumocystis oryctolagi


Pneumocystis is a group of organisms assigned to the Fungal Kingdom (Edman, 1988, 1989; Wakefield, 1992; Calderon-Sandubete, 2002). The genus comprises pathogens dwelling in the lungs of terrestrial, aerial and aquatic mammals (Laakkonen, 1993, Laakkonen & Sukura, 1997; Laakkonen, 1998, 2001; Mazars, 1997b; Guillot, 1999, 2001; Durand-Joly, 2000; Demanche, 2001, 2003). Occasionally they induce severe pneumonitis, particularly in hosts with severe impairment of the immune system. In such hosts, Pneumocystis species develop progressively and may fill pulmonary alveolar cavities, a process that leads to respiratory failure (Dei-Cas, 2000).

The highly ubiquitous occurrence and the marked pathogenic potential of Pneumocystis species, especially of the human-associated Pneumocystis jirovecii, has stimulated a growing interest in these peculiar microfungi. On the basis of morphological, phylogenetic and experimental approaches we are now beginning to realize that Pneumocystis constitutes a highly diversified biological group, with numerous species that are host-specific and well adapted to live inside the lungs of a great diversity of mammal species (Guillot, 2001; Hugot, 2003).

Ultrastructural studies have shown that Pneumocystis species from diverse mammals, especially the trophic forms (formerly called ‘trophozoites’), attach specifically to type-1 epithelial alveolar cells (Dei-Cas, 2004). Trophic forms emit cytoplasmic expansions or filopodia that vary in thickness depending on the species (Dei-Cas, 1994, 2004; Mazars & Dei-Cas, 1998; Nielsen, 1998). Filopodia may penetrate deeply into the cytoplasm of the host cell (Dei-Cas, 1991). However, no disruption of host cell membrane results from either attachment or filopodial activity. In addition, no structural or functional host cell alteration was found in in vitro or in vitro studies using transmission electron-microscopy (TEM) (Dei-Cas, 1991, 2004; Settnes & Nielsen, 1991; Aliouat, 1993a), confocal microscopy (unpublished) or exploring the alveolar epithelium cytophysiology (Beck, 1998).

Molecular genetic studies have revealed that Pneumocystis gene sequences present a marked divergence with the host species concerned. Numerous gene fragments were compared in Pneumocystis from humans and other primates, rodents, rabbits and insectivores. Consistently it was found that specific sequences could be attributed to pathogens from different host species (Banerji, 1995; Mazars, 1995; Laakkonen, 1998; Wakefield, 1998; Denis, 2000; Durand-Joly, 2000; Guillot, 2001). Karyotypic divergence was found among Pneumocystis strains from diverse hosts (Keely, 2004). A multilocus enzyme electrophoresis (MLEE) approach showed that laboratory rats, mice and rabbits harbor dramatically different categories of Pneumocystis genotypes. The high level of linkage disequilibrium found in this study suggested that Pneumocystis genotypes from different hosts have been genetically isolated from each other for a very long time (Mazars, 1997a; Mazars & Dei-Cas, 1998). The evidence of robust Pneumocystis genetic heterogeneity (Dei-Cas, 1998b) has led to the replacement of ‘formae speciales’ by genuine species (Redhead, 2006). The rabbit-associated organism until now has been referred to as Pneumocystis carinii f.sp. oryctolagi (Anonymous, 1994).

A more recent study in primates showed that large subunit of mitochondrial ribosomal DNA (mtLSU-rDNA) sequence divergence among Pneumocystis species was correlated with the phylogeny of their hosts (Demanche, 2001). This observation, which could be extended to other mammals (Demanche, 2001; Guillot, 2001), suggested that cophylogeny can explain the current distribution of pathogens in their hosts. In order to test this hypothesis, aligned DNA sequences of three genes [dehydropteroate synthetase (DHPS), mtLSU-rRNA and small subunit of mitochondrial ribosomal RNA (mtSSU-rRNA)] from strains originating from 20 primate species were subjected to separate phylogenetic analyses, and then combined in a single data set (Hugot, 2003). At least 61%, and perhaps as much as 77%, of the homologous nodes of the cladograms of hosts and pathogens may be interpreted as resulting from codivergence events (Hugot, 2003). Coevolution of Pneumocystis species and their hosts could explain both the remarkable adaptation of these pathogens to the alveolar environment and the close host specificity of Pneumocystis, which was proven by cross-infection experiments (Gigliotti, 1993; Aliouat, 1993b, 1994; Mazars &Dei-Cas, 1998; Dei-Cas, 1998b; Atzori, 1999; Durand-Joly, 2002).

The high divergence among Pneumocystis species, probably resulting from a prolonged process of coevolution with each mammal host and mostly associated with cospeciation (Hugot, 2003), is consistent with the marked phenotypic divergence recently reported to exist among Pneumocystis species from diverse mammals (e.g. the Pneumocystis species selective infectivity). Pneumocystis species show little difference at the light microscopy level, but host species-related divergence was found using TEM (Dei-Cas, 1994, 2004; Nielsen, 1998). Further differences were found in growth rates (Aliouat, 1999) and in vitro behaviour (Aliouat, 1993a). For instance, rat-derived Pneumocystis seemed to have a higher capacity for attaching in vitro to target cells than mouse-derived pathogens, and in vitro attachment of rat Pneumocystis seemed to be more sensitive to pentamidine or cytochalasin-B than attachment of mouse-derived organisms (Aliouat, 1993a).

Combining the above data, the presence of host specific ultrastructural and biological differences among Pneumocystis species strengthen the view of multiple species existing within the genus Pneumocystis (Stringer, 2001). The following species were subsequently described: Pneumocystis carinii Frenkel, P. jirovecii Frenkel, Pneumocystis wakefieldiae Cushion et al, and Pneumocystis murina Keely et al. (Frenkel, 1999; Cushion, 2004; Keely, 2004). As the genus was assigned to the fungal kingdom, these species were described according to rules of the International Code of Botanical Nomenclature (ICBN).

The present paper focuses on a Pneumocystis species identified in meat-, laboratory and wild rabbits (Oryctolagus cuniculus) from the Old World. Genomic, isoenzymatic, ultrastructural, and biological data, obtained mostly in France during the past 15 years, made it possible to distinguish rabbit-derived Pneumocystis from species or formae speciales originating from other mammals. From this work it became clear that rabbit-derived Pneumocystis belongs to a hitherto undescribed species. Genotypic, phylogenetic and phenotypic bases for designating this new species are summarized.

How to set about doing research on rabbit-associated Pneumocystis

Sources of rabbit-associated Pneumocystis

Rabbits hosts were obtained from the following sources: (a) European suppliers of laboratory animals (Charles River, Rouen, France; Iffa Credo, Lyon, France; BioMérieux, Lyon, France; Harlan, Oxon, United Kingdom; Harlan, Zeist, The Netherlands; Rabbit Pathology Unit, INRA, Nouzilly France; Vasseur from Prouzel, Barrois from Nord-Pas de Calais, France), (b) breeders of meat-rabbits (Deregnaucourt and Pollet from Nord-Pas de Calais), and (c) wild rabbits trapped in France and Spain. Most domestic animals were California-New Zealand hybrid rabbits, but Dutch or Chinchilla rabbits (from Harlan, UK and The Netherlands) have also been used. In addition, rabbits from colonies maintained under isolated conditions on farms in different regions in France were used, especially for rabbit Pneumocystis population genetics approaches. Strains and regions were as follows: ‘Hollandais’, ‘brun marron’ and ‘Rex’ from Alsatian farms; ‘Blanc de Bouscat’ from Verdun farms; ‘Nain de Couleur’ and ‘Polonais aux yeux roses’ from Var farms; ‘Argenté de Champagne’ from Saone et Loire; ‘California’ from Dordogne; ‘Sablés des Vosges’ from Bas-Rhin, and outbred rabbits from the cities of Boulogne (Pas de Calais) and Rodez (Aveyron).

The challenge of Pneumocystis-free rabbits

Studies on Pneumocystis pneumonia (PcP) associated pulmonary surfactant changes (Aliouat, 1998) and on local immune response against Pneumocystis infection (Allaert, 1996, 1997; Rajagopalan-Levasseur, 1998) have stimulated a growing need of Pneumocystis-free rabbits (i.e., with repeated negative PCR results and/or no PcP development after continuous corticosteroid administration) to be used as control. These were obtained from the Rabbit Pathology Unit (INRA, Nouzilly, France) by combining prolonged cotrimoxazole administration with careful microbial isolation measures (Cere, 1997b).

Sampling procedures

Pulmonary material to detect Pneumocystis was sampled using noninvasive terminal broncho-alveolar lavage (BAL) or post-mortem homogenization of lungs. Another noninvasive method consisted of gently rinsing the nasal cavities with sterile saline. This method allows the rabbits to be kept alive, and was inspired by the success of noninvasive sampling of nasopharyngeal aspirates from small children (Nevez, 2001; Vargas, 2001). In rabbits, the nasal cavities were rinsed with 1–2 mL of a sterile NaCl 0.9% aqueous solution, using 16G × 2″ I.V. catheters (Terumo Europe N.V., Leuven, Belgium). Secretions were collected in 15 mL capped sterile tubes. After sampling, nasal wash fluids were put on ice for transport to the laboratory. Then, the samples were centrifuged at 2900 g for 10 min at 4°C. The supernatants were removed and the pellet stored at −80°C until DNA extraction and Pneumocystis DNA detection by nested PCR at the mtLSU-rDNA locus (see below) (Wakefield, 1996).

Other sampling methods were performed after rabbit euthanasia by pentobarbital irreversible anesthesia. For terminal BAL the trachea was cannulated and the lungs were rinsed five times with 10 mL of sterile NaCl 0.9% solution. Fluid recovered was pooled on crushed ice and centrifuged at 2900 g for 10 min at 4°C to pellet cells. Pellets and supernatants were stored separately at −80°C. Terminal BAL fluid (BALF) samples were used as source of pulmonary surfactant (Aliouat, 1998), alveolar macrophages (Allaert, 1996) or of host-cell RNA in rabbit-PcP immunology studies (Allaert, 1997). Finally, postmortem sampling of rabbit lungs, which is described in the next section, was used either to evaluate the kinetics of Pneumocystis infection (Soulez, 1989; Dei-Cas, 1990b; Aliouat, 1998, 1999) or as source of Pneumocystis antigen for immunofluorescence (IFA) and Western-blot assays (Soulez, 1988, 1989; Dei-Cas, 1990b).

Staining Pneumocystis organisms for light microscopy

Rabbit-derived Pneumocystis organisms were usually detected in lung impression smears (Dei-Cas, 1989, 1990a; Soulez, 1989), lung-homogenate air-dried smears (Soulez, 1989, 1991; Rajagopalan-Levasseur, 1998) or BALF samples (Aliouat, 1998). Although rabbit pathogens can be identified using phase contrast microscopy (Rajagopalan-Levasseur, 1998), like other Pneumocystis species (Dei-Cas, 2004), current detection was made using toluidine blue O (TBO) (Chalvardjian & Grawe, 1963), Gomori-Grocott's methenamine silver nitrate (GMS) (Grocott, 1955; Rajagopalan-Levasseur, 1998), and methanol-Giemsa or Giemsa-like stains with similar cytological affinities, such as the RAL-555 kit (Réactifs RAL, Paris, France) (Cushion, 1988; Soulez, 1988, 1991; Dei-Cas & Cailliez, 1996, 1998a). Additionally, Pneumocystis-specific fluorescein-labelled antibodies have helped to identify Pneumocystis organisms in impression smears or lung-homogenate air-dried smears (Soulez, 1988).

In order to approach lung tissue changes associated with rabbit pneumocystosis (Dei-Cas, 1990b; Rajagopalan-Levasseur, 1998), conventional histological methods have also been used, as described in detail elsewhere (Creusy, 1996; Dei-Cas, 1998a). In rabbit lung sections Pneumocystis cystic forms were detected using TBO, GMS or even Periodic Acid Schiff (PAS) stains (Emmons, 1977; Dei-Cas, 1998a).

Fixing Pneumocystis organisms for ultrastructural study

The effect of a large range of osmolarities of fixative and washing solutions (190–2580 mOsm) on the structural preservation of Pneumocystis cells was tested in our laboratory (Palluault, 1992a). Tests were made on rabbit-derived Pneumocystis, and revealed that high osmolarity (850–1300 mOsm) of fixative and washing solutions was a critical condition for obtaining well-preserved Pneumocystis cytoplasmic structures. The following protocol was found to be effective for fixation of Pneumocystis-infected lung samples from rabbits, and was therefore used in this work: (1) fixation with a phosphate-buffered 2.5% glutaraldehyde solution (0.1 M pH 7.5) adjusted to about 700 mOsm by the addition of 0.18 M NaCl; (2) many washings with 0.1 M phosphate buffer (same osmolarity); (3) postfixation for 1 h in a 1%-osmium tetroxide solution in phosphate buffer, dehydration in ethanol, and embedding in Epon (Dei-Cas, 1998a). Finally, 2D images from serial ultra-thin sections were used to reconstruct 3D images of two Pneumocystis life cycle stages (trophic form and intermediary sporocyte) (Palluault, 1991b, c; Dei-Cas, 2004).

How to separate Pneumocystis organisms from lung tissue

Methods to separate, purify and enumerate Pneumocystis from rabbit-lung tissue were described previously (Soulez, 1991; Dei-Cas & Cailliez, 1996). Briefly, infected lungs are cut into small pieces in Dulbecco Modified Eagle's Medium (DMEM) and homogenized either by squeezing them through a stainless steel mesh or using a magnetic stirrer (4°C, 90 min) or a Stomacher tissue grinder. The first two methods are usually employed in order to keep living pathogens for infectivity or other studies (Soulez, 1991; Aliouat, 1993a, b, 1994). A stomacher is currently used when the aim is simply to evaluate the number of Pneumocystis organisms (Soulez, 1991). In all cases, the homogenate is poured through gauze and centrifuged (2900 g 10 min 4°C). The pellet is incubated in a buffered hemolytic solution (9: 1 solution of 0.15 M NH4Cl in 20 mM Tris-HCl, 10 min 4°C), resuspended in DMEM and filtered successively through 250 and 63 μm stainless steel meshes. The pellet is finally resuspended in DMEM. All procedures are performed under sterile conditions.

Finally, in order to further reduce host cell debris, pathogens are suspended in a polysucrose gradient (Histopaque-1077, Sigma Chemical Co., L'Isle D'Abeau Chesnes, France). Polysucrose solution and pathogen suspension are mixed 1: 1 (v/v) in a 15 mL sterile tube and centrifuged at 1000 g for 15 min at 4°C. The band accumulated at the interface between sterile medium (DMEM or PBS) and polysucrose solution is collected and washed twice with sterile medium (2900 g 10 min 4°C). This supplementary purification step was employed, for instance, to prepare rabbit-derived Pneumocystis samples for Western-blot assays or for pulsed field gel electrophoretic karyotype analysis.

Viable, purified Pneumocystis samples may be used immediately or can be cryopreserved by placing them in fetal calf serum with 10% dimethyl-sulfoxyde (DMSO) at −80°C in a Nalgene 1°C Cryo Freezing Container (Dutscher, Brumath, France) filled with isopropyl alcohol (cooling rate=1°C min−1) (Dei-Cas & Cailliez, 1996). Then, the pathogen samples are stored in liquid nitrogen. Under these conditions, Pneumocystis samples remain infectious for at least six years (Durand-Joly, 2002).

Counting the Pneumocystis organisms in lung samples

For organism counts, cystic forms were counted in 2 or 5 μL air-dried smears stained with toluidin blue O (TBO). Dry smears were fixed with methanol, stained with Giemsa or RAL555 stains, and used to count the relative number of trophic forms, sporocyte and cyst stages of Pneumocystis. The total number of pathogens is calculated as follows: Embedded Image where W is the walled forms (counted on TBO smears); %W the percentage of walled forms (intermediate sporocytes+late sporocytes+cysts); %UW the percentage of unwalled forms (trophic forms+early sporocytes) (Aliouat, 1993b, 1995).

Amplifying and sequencing Pneumocystis DNA

Rabbit lung and nasal wash samples were treated with proteinase K, and genomic DNA was isolated by a phenol–chloroform extraction, or using QIAamp® DNA minikit (QIAGEN, Courtaboeuf, France) according to manufacturer's recommendations. Single or nested-PCR, using oligonucleotide primers described in Table 2, were carried out from rabbit-derived Pneumocystis DNA samples to amplify portions of the following genes: thymidylate synthase (TS), mitochondrial large-subunit rRNA (mtLSU-rRNA), mitochondrial small-subunit rRNA (mtSSU-rRNA), arom locus, manganese-dependent superoxyde dismutase (MnSOD), dihydrofolate reductase (DHFR), dihydropteroate synthase (DHPS), β-tubulin (β-tub), heat-shock-protein 70 (HSP70), and internal transcribed spacer regions (ITS) (Banerji, 1993; Mazars, 1995; Hunter & Wakefield, 1996; Wakefield, 1996; Denis, 2000; Ma & Kovacs, 2001).

View this table:
Table 2

List of explored rabbit-derived Pneumocystis genes

GenesPrimersAccession numberReferences
TSSense: 5′-ATTTATGGGTTTCAATGG-3′UnpublishedMazars (1995)
mtLSU-rDNAFirst round PCR:S42915Wakefield (1996)
Second round PCR:
mtSSU-rDNASense: pAZ112-10 F 5′-TAGACGGTCACAGAGATCAG-3′UnpublishedHunter & Wakefield (1996)
Antisense: pAZ112-10 R 5′-GAACGATTACTAGCAATTCC-3′
Arom locusFirst round PCR:U31054Banerji (1993)
Antisense: 5′-(G,C)(A,T)(T,C)TTICCIGCI(G,C)CIC(G,T)CAT-3′
Second round PCR:
HSP70Sense: 5′-GATGAAAGAATTAGCAGAAACTAA-3′DQ435616Chabé (2004)
ITSFirst round PCR:UnpublishedPresent work
Second round PCR:
β-tubSense: 5′-TGGGCAAAAGGGCATTATAC-3′UnpublishedPresent work

Amplified PCR products were purified before sequencing either directly or after cloning. In the latter case, recombinant plasmids were sequenced in both directions with a model ABI 377 automated sequencer using the Big Dye Terminator Cycle Sequencing kit (Perkin Elmer-Applied Biosystems, Foster City, CA), according to the manufacturer's instructions. Sequences used for comparisons were obtained from the literature and from the GenBank database (http://www.ncbi.nl.nih.gov/Genbank/GenbankOverview.html).

Pneumocystis taxonomy at the species level: sequence comparison and phylogenetic analysis

Sequences compared in this study

In order to distinguish rabbit-derived Pneumocystis organisms from the other Pneumocystis species or formae speciales, portions of the following genes were aligned: manganese-containing cofactored superoxide dismutase (SODA), accession nos AF146752, AF146753, Z79785, AF146751, AF146754; dihydropteroate synthase (DHPS), accession nos AF322064, AF139132, M86602, AF322065, U66283, AY070270, AF362762, AF362761, AF362760, AF362759, AF362758, AF362757; dihydrofolate reductase (DHFR), accession nos AF186097, AF090368, AF322061, AF322063, AF175561, AY017418; large-subunit mitochondrial rRNA (mtLSU-rRNA), accession nos S42915, S42926, U20169, U20173, S42921, AF257179, AF362455, AF362462, AF362461, AF362458, AF362456, AF362464, AF362470, AF362469, AF362468, AF362467, AF362466, AF362465, AF362463, AF362460, AF362459, AF362457, AF362454, AF362453; mtSSU-rRNA, rabbit-derived Pneumocystis, P. jirovecii, P. carinii, P. wakefieldiae, P.carinii f.sp. mustelae, P. murina (Hunter & Wakefield, 1996), P. carinii f.sp. macaca from Indian Macaca rhesus (Durand-Joly, 2000), accession nos AF395579, AF395580, AF395582, AF395584, AF395578, AF395583, AF395574, AF395585, AF395576, AF395575, AF395577, AF395581; thymidylate synthase, rabbit-derived Pneumocystis, P. jirovecii, P. murina (Mazars, 1995), accession nos S77510; 5-enolpyruvylshikimate-3-phosphate synthase (AROM), accession nos U31054, U31055, L18918, U31056, U31053; internal transcribed spacer 1 (ITS1), accession nos DQ010098, AF013806, L27658, AY532651, AF288827; internal transcribed spacer 2 (ITS2), accession nos DQ010098, AF013821, L27658, AY53651, AF288835; Heat shock protein 70 (HSP70), accession nos DQ435616, U80970, U80968, U80969, AY382182; beta-tubulin (β-tub), rabbit-derived Pneumocystis (this paper), accession nos AF170964, X62113.

Matrix and phylogenetic tree construction

Sequences were aligned by use of the BioEdit v7.0.1 package (http://mbio.ncsu.edu:BioEdit/bioedit.html). Introducing a limited number of gaps optimized the alignments. Ambiguous regions in the alignments were not taken into account. The DNA alignments contained 496, 798, 330, 297, 325, 410, 167 and 283 common positions for SODA, DHPS, DHFR, TS, AROM, HSP70, mtLSU-rRNA and mtSSU-rRNA, respectively. The protein alignments contained 165, 265, 110, 99, 108 and 136 common residues for SODA, DHPS, DHFR, TS, AROM and HSP70, respectively. For the phylogenetic analysis of 18 Pneumocystis taxa (see below), mtLSU-rRNA and mtSSU-rRNA sequences were concatenated to form a sequence of 281 nt long. Full-length alignments and sites used in analyses are available upon request from the corresponding author.

The relatedness of pairs of aligned sequences for each individual gene was calculated using the ‘Sequence Identity Matrix’ option provided in the BioEdit package. The values were obtained by dividing the number of nucleotide or residues identities by the total number of positions compared and given in percentages. The pairwise distances (%) presented in the matrix were calculated for each pair of aligned sequences as follows: 100−% identity. The ITS and β-tub sequences were not included in the matrix because of the low number of available sequences (β-tub) or unambiguously alignable common positions (ITS).

Phylogenetic analysis of the mtLSU-rRNA and mtSSU-rRNA concatenated sequences dataset was carried out using MrBAYES v3_0b4 (Huelsenbeck & Ronquist, 2001). Bayesian analysis was performed using the GTR (general time reversible) Γ (gamma distribution of rates with four rate categories)+I (proportion of invariant sites) model of sequence evolution, with base frequencies, proportion of invariant sites and the shape parameter α of the Γ distribution estimated from the data. The model used was evaluated using the likelihood ratio test (LRT) implemented in MODELTEST v.3.7. (Posada & Crandall, 1998). LRTs indicated that the GTR+I+Γ model had the best fit to the data.

Starting trees were random; four simultaneous Markov chains (three heated, one cold) were run twice for two million generations, burn-in values were set at 10 000 generations (based on empirical values of stabilizing likelihoods), and trees were sampled every 100 generations. Bayesian posterior probabilities were calculated using a Markov chain Monte Carlo sampling approach (Green, 1995) implemented in MrBAYES version 3.0b4.

Pneumocystis and pneumocystosis in rabbits

Current pneumocystosis of rabbits at weaning

Extensive corticosteroid-induced PcP was reported in rabbits (Oryctolagus cuniculus) as early as in the 1950s by Sheldon (1959). In 1989, however, it was reported that without corticosteroid administration, rabbits developed spontaneous PcP at weaning (about 1 month after birth) (Soulez, 1989). This spontaneous, natural Pneumocystis infection which has been constantly observed in weaning rabbits of several strains (Mazars, 1997a), provokes lung histopathological changes typical of PcP, often associated with blood biochemical abnormalities (Soulez, 1989; Dei-Cas, 1990b; Rajagopalan-Levasseur, 1998). The infection evolves during 7–10 days; afterwards, pathogen levels decrease gradually, becoming very low in 60-day-old rabbits. Almost all animals recover within 3–4 weeks. The regularity of the pattern of this natural infection (abrupt onset at the weaning time, extensive diffuse pulmonary involvement evolving relatively shortly to complete, spontaneous healing) has allowed the development of kinetic studies of the host immune response against PcP (Allaert, 1996, 1997). This research was facilitated by the fact that PcP develops in this model without administration of corticosteroids, as these drugs affect the immunological mechanisms. Corticosteroids also influence the production and composition of pulmonary surfactant. For this reason, the corticosteroid-free rabbit model was recently used to investigate Pneumocystis-surfactant interactions (Prevost, 1997, 1998; Aliouat, 1998).

Morphology of rabbit-associated Pneumocystis at the light microscopic level

TBO and GMS, having a good affinity for components of the cyst wall, stain the cell wall of cystic forms (=intermediate and late sporocytes plus mature cysts) in reddish violet or dark brown, respectively. These techniques are highly sensitive, allowing an easy detection of cystic forms even at low magnifications (Fig. 1). However, trophic forms and early sporocytes remain unidentified with these metachromatic stains (Dei-Cas, 2004). Therefore, in order to detect all the Pneumocystis life cycle stages, methanol-Giemsa or Giemsa-like stains (Fig. 1) have to be associated with TBO or GMS stains. Giemsa and other stains with similar cytological affinities, such as Diff Quick (Cushion, 1985) or RAL-555 (Dei-Cas & Cailliez, 1996), cause pinkish-purple stains on the Pneumocystis nucleus, and blue stains on the cytoplasm. In fact, only methanol-Giemsa and similar polychrome stains (e.g. RAL-555 or Diff-Quick) allow the identification of the different Pneumocystis life-cycle stages (Cushion, 1988; Dei-Cas, 2004). These stains also allow Pneumocystis cells to be distinguished from other organisms (Dei-Cas, 1998a). These dyes do not, however, stain cystic or sporocytic thick cell walls, which appear like a clear peripheral halo around the fungus cell.

Figure 1

Pneumocystis oryctolagi sp.nov.: morphology and pathology at the light microscope level. (a) Cystic forms in an air-dried lung homogenate smear stained by TBO. (b) A mature cyst containing eight ascospores is seen close to the nucleus of a host cell. Air-dried lung smear stained by methanol-Giemsa. (c) Cystic forms (arrowheads) mostly lining an alveolar space. Histological section of the lung of a weaning rabbit stained by TBO (technique for tissue sections). (d–f) Trophic forms (d), a mononucleate sporocyte (e) and a mature cyst (f) containing eight ascospores. Air-dried lung smears stained by methanol-Giemsa. (g) Lung of a weaning rabbit with pneumocystosis. Alveolar septa are thickened; macrophages and other inflammatory cells infiltrate mildly the alveolar lumens. In a severely infected area (at the top, on the right) alveoli are entirely occupied by cell infiltrates. Lung section stained by hematoxylin-eosin stain. (h) Lung of a weaning rabbit with pneumocystosis. A nodular lesion is clearly observed. Lung section stained by hematoxylin-eosin stain. Bar=10 μm (a–f) or 100 μm (g, h).

Trophic forms are irregular in shape and size (2–8 μm diameter). Their cytoplasm contains a unique, homogenous, well-stained nucleus. A spheroid shape and a usually easily visible Giemsa-unstained thick cell wall characterize the 4–7 μm cystic forms (intermediate and late sporocytes and mature cysts). The number of nuclei increased as organisms proceed in their development from the mononuclear trophic form to the mature cyst, which contains eight well-individualized mononuclear spores. Thus, there is one nucleus in the early sporocyte, two to eight nuclei in the intermediate sporocyte, and eight nuclei in the late sporocyte (Fig. 1) (Dei-Cas, 2004).

At the light microscope level, rabbit-derived Pneumocystis cannot be distinguished unequivocally from other Pneumocystis species. In methanol-Giemsa stained smears, however, whereas rabbit-associated Pneumocystis forms are usually well detached from each other, pathogens from rats or primates constitute often large stacks where the different life cycle stages are closely clustered (Table 1). Furthermore, in rabbits with spontaneous PcP the cystic/trophic form ratio is usually higher (about 0.10–0.15) than in immunosuppressed rodents with PcP (about 0.02–0.05) (Table 1).

View this table:
Table 1

Phenotypic differences between rabbit-derived Pneumocystis and other Pneumocystis species

FeaturesRabbit-derived PneumocystisP. carinii and P. wakefieldiaeP. murinaP. jirovecii
Organisms in lung dry smears (TBO or Giemsa stains)Detached from each otherClosely clusteredClusteredClosely clustered
Cystic/trophic form ratio0.10–0.150.02–0.050.02–0.05ND
LocationLining alveolar epitheliumFilling alveolar lumenFilling alveolar lumenFilling alveolar lumen (AIDS) or lining alveolar epithelium (epidemic or infantile PcP)
In vivo doubling time1.7 days (untreated rabbits)4.5 days (P. carinii in corticosteroid-treated rats)10.5 days (SCID mice)ND
Specific hostRabbit (Oryctolagus cuniculus)Rat (Rattus norvergicus)Mouse (Mus musculus)Man (Homo sapiens)
Intraalveolar eosinophilic honeycomb materialRarePresentPresentPresent
  • ND, not determined.

Finally, histological differences between rabbit and rodent or even primate pneumocystoses were reported (Creusy, 1996). They involved the pathogens, their location in the lung tissue and the pattern of the inflammatory response (Table 1). Host immune status could surely influence the immune and inflammatory responses to the infection, but whether one or other specific difference is due to immunodepression remains unclear. First, pathogens lined the alveolar epithelium in rabbits (Fig. 1), whereas in corticosteroid-treated rats, SCID mice or AIDS patients, they were usually more numerous, closely clustered, and located in the alveolar lumen. Second, pulmonary congestion was less important in rabbits (Fig. 1) than in murine hosts and AIDS patients, where it was widespread and severe. Third, no collagen fibrosis at all was observed in rabbits, though they had not received corticosteroids, whereas diverse degrees of interstitial fibrosis were observed in both AIDS patients and SCID mice. Fourth, among the inflammatory cells, eosinophils and plasma cells were observed in rabbits, a fact that was confirmed using TEM (Fig. 2) (Creusy, 1996; Rajagopalan-Levasseur, 1998), and occasionally in humans with AIDS-related PcP (Fleury-Feith, 1989). Fifth, inflammatory infiltrates were usually diffuse in rodent and human hosts, while they appeared as clearly delimited nodular areas scattered in the rabbit lung (Fig. 1), containing therefore eosinophils and plasma cells. Sixth, the typical eosinophilic foamy honeycomb material, present in rodent and human hosts, was rarely found in rabbits (Creusy, 1996).

Figure 2

Pneumocystis oryctolagi sp.nov.: fungus morphology and associated host cells at the ultrastructural level. (a) Trophic forms and a late sporocyte where ascospores are being generated. Arrowheads indicate filopodia. The arrow is showing the basement membrane of the alveolar epithelium. (b, c) An eosinophil leukocyte (b) and a plasmocyte (c), two cell types often associated with rabbit pneumocystosis. AL, alveolar lumen; SP, sporocyte; TF, trophic form. Bar=1 μm.

Ultrastructure of rabbit-associated Pneumocystis

Most knowledge about rabbit-derived Pneumocystis cell structure resulted from studies made in the 1990s thanks to both significant improvements of TEM-fixation methods (Palluault, 1992a, b), and computer-aided 3D reconstruction studies (Palluault, 1991a, b, c). Interestingly, these studies revealed distinctive ultrastructural features of rabbit Pneumocystis (Dei-Cas, 1994; Mazars & Dei-Cas, 1998; Nielsen, 1998; Durand-Joly, 2000). Most features however can be extended to other Pneumocystis species. The detailed ultrastructure of rabbit-derived Pneumocystis life cycle stages was reviewed recently (Dei-Cas, 2004). It will be summarized shortly here.

All known rabbit-associated Pneumocystis life cycle stages were found in the lungs of infected rabbits, though molecular methods suggested pathogens may spread to other organs (Cere, 1997a). The usually accepted Pneumocystis life cycle involves an amoeboid, thin-walled, mononuclear trophic form, which becomes a thick-walled cystic stage, in which multiple nuclear division leads to the formation of eight spores (Fig. 2). These forms would be able to leave the cyst, presumably by a pore-like zone located at the thickest part of the cyst cell wall, as was suggested for P. carinii (Itatani, 1994), to attach specifically to type-I epithelial alveolar cells and to evolve into the cystic stage.

The transition from trophic form to mature cyst occurs through three consecutive sporocyte stages (early, intermediary and late sporocyte) (Yoshida, 1989; Dei-Cas, 2000). Trophic forms and early sporocytes have a thin cell wall 20–25 nm thick that consists of only one electron-dense layer associated with the outer surface of the plasma membrane (6–7 nm), which extends from the cellular body to the filopodia or tubular expansions. Therefore, these structures — which are frequently observed in cross, oblique or longitudinal sections (Fig. 2) – constantly show cell wall, plasma membrane and cytoplasm levels.

The smallest trophic forms are round to ellipsoid, appearing as eukaryotic cells 1–2 μm in length. Most trophic forms are, however, larger and very irregular in size (4–8 μm long) and shape (Fig. 2), and present filopodia. Their cytoplasm contains one nucleus (up to 1 μm in diameter) that has a typical nuclear envelope with clearly visible 55–80 nm nuclear pores (Palluault, 1990). The chromatin generally appears diffuse. As in small trophic forms, the perinuclear cisterna communicates with well-developed rough (RER) or smooth (SER) endoplasmic reticulum (Palluault, 1990).

The endomembranous system of rabbit-derived Pneumocystis showed two types of endoplasmic structures closely related to RER, SER, and the Golgi complex. The first, which was named a type 1 endoplasmic saccule (ES1) (Palluault, 1990), consisted of one or more coiled endoplasmic saccules that packaged cytoplasm or mitochondria, suggesting autophagic activity. ES1 could therefore be considered as secondary lysosomes. The second type, which was named a type 2 endoplasmic saccule (ES2), consisted of a large, flattened, single endoplasmic saccule present in well-developed trophic forms and in intermediate sporocytes. Although it seems to appear just before nuclear division, its function remains unknown. Furthermore, the cytoplasm contains 50–70 μm osmiophilic granules that are probably lipoid in nature (Palluault, 1990).

A single mitochondrion, as shown by ultrastructural 3D reconstruction, with budding zones occupies an important volume in the cell (Palluault, 1991b). In the intermediate sporocyte the mitochondrion develops active budding, becoming somewhat tree-like (Palluault, 1991c). The organelle evolves apparently by budding into individual mitochondrion of spores. Vesicles of Golgian nature (Dei-Cas, 1989, 2004; Palluault, 1990) develop by budding from either endoplasmic saccules or nuclear envelope. Their number increase from about 20 in trophic forms to about 200 in sporocytes (Palluault, 1990), as organisms proceed in their development from the trophic form to the intermediate sporocyte stage, suggesting that this transition is associated with an increased synthesis of cell wall compounds (glucan, chitin, glycoprotein). The key event of early to intermediate sporocyte transition is the development of the thick cell wall (Fig. 2). The mono-layered cell wall of the early sporocyte becomes thickened by the appearance of a glucan-rich electron-lucent middle layer that results in an increase of the cell wall thickness from 40 to 100 nm (Yoshida, 1989; Dei-Cas, 2000).

The mature cyst is about 5 μm in diameter, round, and thick-walled. Its surface is rather smooth, with rare filopodia. It contains a maximum of eight spores (Fig. 1). These are mononuclear cells, which result from invaginations of the late sporocyte cell membrane. Spores present a single mitochondrion, a well-developed rough endoplasmic reticulum and an electron-dense one-layered cell wall that is externally lined by an outer cell membrane (Fig. 2) (De Stefano, 1990; Palluault, 1992a). Actually, like P. carinii, rabbit-associated organisms have an outer membrane (Fig. 2), a structure apparently absent from the cell wall of other fungi, that appears as a more or less discontinuous osmiophilic deposit that lines the trophic-form plasma membrane or is embedded in the electron-dense outer layer of thick-walled stages (De Stefano, 1990; Palluault, 1992a).

By ultrastructure, rabbit-associated Pneumocystis is not distinguishable from primate Pneumocystis species (Frenkel, 1976; Durand-Joly, 2000) but can be easily distinguished from rodent Pneumocystis (Table 1). Most differences between rabbit and rodent Pneumocystis species involve filopodia. These typical structures of Pneumocystis trophic forms are markedly more numerous, thin, and tree-like in Pneumocystis organisms from mouse than in those from rabbit, human, or macaque (Dei-Cas, 1994, 2004; Creusy, 1996; Nielsen, 1998; Durand-Joly, 2000). Filopodia of rat-derived organisms were also found to be smaller than those from rabbit-derived Pneumocystis. Additionally, the density and diameter of membrane-limited electron-dense cytoplasm granules were found to be respectively higher and larger in mouse- than in rabbit-derived Pneumocystis cells (Nielsen, 1998).

Growth rate and host specificity of rabbit-associated Pneumocystis

The growth rates of Pneumocystis species (Aliouat, 1999) and their strong host specificity (stenoxenism) (Aliouat, 1993b, 1994; Dei-Cas, 1994, 1998b) express at best the importance of biological divergence among Pneumocystis species. Actually, the doubling time of Pneumocystis organisms developing in the host lung was highly variable in terms of the host species: 1.7 days for rabbit-derived Pneumocystis, 4.5 days for rat-derived Pneumocystis and 10.5 days for P. murina growing in SCID mice (Aliouat, 1999). In these experiments, the quantitation of Pneumocystis organisms was performed microscopically, as explained in the ‘Methods’ section. Data were plotted in a semi-logarithmic curve, and doubling time (DT) was calculated at the exponential phase as follows: DT=ln 2 μ−1, where μ represents the specific growth rate (i.e. slope of the curve) (Aliouat, 1999).

The strong host specificity of Pneumocystis species was demonstrated in cross-infection experiments that aimed to establish to what extent host species-related genetic variation among Pneumocystis isolates entailed restricted infectious power. Pneumocystis-free SCID mice (infected by nasal route) and Nude rats (infected by tracheal route) have been used as experimental hosts in most experiments (Gigliotti, 1993; Furuta, 1993; Aliouat, 1993b, 1994; Durand-Joly, 2002). Inocula were pathogens isolated from rabbits, rats, mice, ferrets, monkeys or humans. In these experiments, rabbit-associated Pneumocystis, like isolates of the other nonrodent mammals, were not able to develop in the mentioned deeply immunosuppressed experimental hosts. A highly sensitive PCR assay followed by hybridization with DNA-probes specific to each Pneumocystis species (Aliouat, 1995) was used but no Pneumocystis organism was detected in nonspecific hosts (mouse to rat-derived organisms) as soon as 3 days postinfection. Only mouse-derived Pneumocystis developed in SCID mice, and only rat-derived Pneumocystis developed in Nude rats (Aliouat, 1993b, 1994; Furuta, 1993; Gigliotti, 1993; Dei-Cas, 1998b; Wakefield, 1998; Durand-Joly, 2002). These observations attested that Pneumocystis host specificity is an all-or-none event. In short, until now, cross-infection experiments have shown that Pneumocystis organisms from a given mammal cannot infect hosts of another species. Host-species restriction, as established further by molecular identification studies on Pneumocystis species in domestic or wild mammals (Wakefield, 1992, 1997; Peters, 1994a, b; Banerji, 1995; Mazars, 1995, 1997b; Bishop, 1997; Guillot, 1999, 2001; Denis, 2000; Durand-Joly, 2000; Demanche, 2001; English, 2001; Hugot, 2003) seems therefore to be a universal feature of Pneumocystis species. One of the few exceptions was the discovery of macaque-derived Pneumocystis strains, likely to be different species (as mean mtLSU-rDNA sequence divergence was 4.3±1.4%), recovered from both Macaca mulatta (rhesus macaque) and M. fascicularis (cynomolgus macaque) (Guillot, 2004). However, this apparent transgression of the species barrier might indicate that rhesus and cynomolgus macaques are closely related and may belong to a single species. Actually, rhesus and cynomolgus monkeys might have had a large overlapping geographic distribution in the past (Fooden, 1980). Consistently, hybrids between rhesus and cynomolgus monkeys have been reported in Thailand (Fooden, 1964) and in captivity, and are known to be fertile (Bernstein, 1980). There is also some phylogenetic evidence of hybridization between M. fascicularis and M. mulatta (Tosi, 2000), though mitochondrial DNA analyses indicate these two species are separated (Hayasaka, 1996; Morales & Melnick, 1998).

Species-specific gene sequences in rabbit-associated Pneumocystis

All studied DNA sequences of rabbit-associated Pneumocystis were found to be significantly different from the homologous fragments of the other known Pneumocystis species and formae speciales. Targeted genes are shown in Table 2, with PCR primers used, and Table 3, showing the pairwise distances (%) of eight Pneumocystis genes among the Pneumocystis species and formae speciales. ITS1 and ITS2 loci and β-tub genes were not included in the matrix because the number of sequences available was too small (β-tub) or because the variability of sequences was too high (ITS). Considering the compared gene portions, the extent of genetic divergence at the mtLSU-rDNA locus ranged from 18% (between rabbit-derived Pneumocystis and P. jirovecii) to 26.4% (between rabbit-derived Pneumocystis and P. carinii) (Table 3). Distances reported previously were comparable (Wakefield, 1998; Durand-Joly, 2000). In a more recent study, divergences at the same locus were 18.1% between rabbit-derived Pneumocystis and P. jirovecii, and 29.1% between rabbit-derived Pneumocystis and P. carinii (E. Dei-Cas & A.E. Wakefield, unpublished). At the arom locus 25% divergence was found between rabbit-derived Pneumocystis and P. carinii at the deduced amino acid sequence level (present study and Banerji, 1995). Divergence between rabbit-derived Pneumocystis and P. jirovecii was similar at the deduced amino acid sequence level of DHFR locus (25%), and higher (38–39%), at the same locus, between rabbit-derived Pneumocystis and P. carinii (present study and Ma, 2001). Over 99 deduced amino acids of a fragment of the TS locus, divergence was lower: 14.2% between rabbit-derived Pneumocystis and P. carinii, and 9.1% between rabbit-derived Pneumocystis and P. jirovecii (present study and Mazars, 1995). In all the cases, levels of divergence are indicative of species-level variation (Stringer, 1996).

View this table:
Table 3

Pairwise distances (%) of eight Pneumocystis genes

P. oryctolagi/P. jirovecii15.014.612.59.134.433.115.016.324.025.517.316.718.017.0
P. oryctolagi/P. carinii24.429.716.514.235.735.317.618.933.439.121.925.026.417.0
P. oryctolagi/P. wakefieldiaeNANANANA16.436.1NANANANANANA22.218.1
P. oryctolagi/Pc f.sp. mustelaeNANANANANANA17.823.132.832.818.826.018.017.0
P. oryctolagi/P. murina24.029.115.915.223.516.217.219.333.735.521.025.021.616.0
P. oryctolagi/Pc f.sp. macaca15.613.4NANANANA15.217.425.228.2NANA19.814.5
P. jirovecii/P. carinii23.025.516.511.
P. jirovecii/P. wakefieldiaeNANANANA19.114.0NANANANANANA23.420.5
P. jirovecii/Pc f.sp. mustelaeNANANANANANA18.522.318.228.217.319.519.818.1
P. jirovecii/P. murina21.824.316.213.233.530.215.117.432.533.717.319.521.618.4
P. jirovecii/Pc f.sp. macaca7.11.9NANANANA9.911.718.821.9NANA12.510.3
P. carinii/P. wakefieldiaeNANANANA14.910.3NANANANANANA9.68.9
P. carinii/Pc f.sp. mustelaeNANANANANANA18.023.428.833.718.523.221.616.0
P. carinii/P. murina5.13.714.83.131.531.75.96.516.
P. carinii/Pc f.sp. macaca22.625.5NANANANA15.518.534.340.0NANA28.314.9
P. wakefieldiae/Pc f.sp. mustelaeNANANANANANANANANANANANA19.215.2
P. wakefieldiae/P. murinaNANANANA33.033.1NANANANANANA9.67.8
P. wakefieldiae/Pc f.sp. macacaNANANANANANANANANANANANA26.417.4
Pc f.sp. mustelae/P. murinaNANANANANANA16.821.927.326.417.621.318.615.6
Pc f.sp. mustelae/Pc f.sp. macacaNANANANANANA18.323.830.031.9NANA23.714.9
P. murina/Pc f.sp. macaca21.624.3NANANANA15.318.531.334.6NANA27.013.5
  • The pairwise distances (%) presented in the matrix were calculated for each pair of aligned sequences as follows: 100−% identity.

  • Common positions of the DNA alignments and common residues of the protein alignments:

  • SODA: nt: 496 AA: 165. TS: nt: 297 AA: 99. HSP70: nt: 410 AA: 136. DHPS: nt:798 AA: 265. DHFR: nt: 330 AA: 110. AROM: nt:325 AA: 108. mtLSU-rRNA: 167. mtSSU-rRNA: 283. NA, nonaligned.

No gene flow between rabbit- and rodent-associated Pneumocystis species

In an early study investigating the genetic diversity of Pneumocystis organisms isolated from rabbits, mice and rats, multilocus enzyme electrophoresis (MEE) was used to analyse five Pneumocystis enzyme systems: malate dehydrogenase (MDH), glucose phosphate isomerase (GPI), leucine aminopeptidase (LAP), malic enzyme (ME) and 6-phosphogluconate dehydrogenase (6PGDH) (Mazars, 1997a). This study revealed high genetic divergence among the Pneumocystis isolates from the three targeted host species (22 weaning rabbits, 30 corticosteroid-treated rats and 17 corticosteroid-treated mice). Isolates from different host species exhibited clearly distinct isoenzyme patterns (near maximum genetic distance possible) (Mazars, 1997a). Within a given host species, Pneumocystis isolates from mice and rabbits showed very little or no genetic diversity. Especially, the 22 Pneumocystis isolates from rabbits of diverse strain or geographic origin did not show genetic diversity. Rat-derived pathogens diverged by the MDH system that showed three distinct profiles, which were however closely related to each other. However, the most important outcome was that this work enabled us to evaluate the degree of genetic isolation between Pneumocystis genotypes. All linkage disequilibrium tests showed considerable departures from panmictic expectation, and strongly suggested that Pneumocystis genotypes from different hosts species have been genetically isolated from each other for a very long time, representing dramatically distinct gene pools (Mazars, 1997a).

Pneumocystis strains from Old-World rabbits represent a common gene pool

Comparing Pneumocystis genotypes from different hosts with classical biological species (Mayr, 1963; Taylor, 2000), it was shown that Pneumocystis isolates derived from a given host species exchange genes freely and represent therefore a common gene pool. This is clearly the case of the rabbit-associated Pneumocystis. Actually, studies performed between 1992 and 2005 revealed no significant genetic polymorphism in rabbit-derived Pneumocystis isolates from both domestic and wild Old-World rabbit populations. In these studies, in addition to the MLEE-based work discussed above (Mazars, 1997a), numerous gene sequence comparisons have been performed. Targeted loci were TS (Mazars, 1994, 1995), mtLSU-rRNA (Wakefield, 1992; Peters, 1994a; Guillot, 1999, 2001; Durand-Joly, 2000), mtSSU-rRNA (Durand-Joly, 2000; Guillot, 2001), arom (Banerji, 1995), SODA (Denis, 2000), DHFR and DHPS (Ma, 2001).

Within some Pneumocystis species, such as P. carinii, P. jirovecii, ferret-derived Pneumocystis- (Wakefield, 1998) and shrew-derived Pneumocystis (Peters, 1994a; Laakkonen & Sukura, 1997), intraspecies polymorphism was shown. Regarding rabbit-derived Pneumocystis, no difference was found among isolates from diverse domestic rabbit strains and/or different geographic origins (Mazars, 1994, 1995, 1997a; Dei-Cas, 1998b). Some divergence at the mitochondrial ribosomal RNA loci was found between isolates from wild rabbits of diverse European geographic regions on the one hand, and strains of domestic rabbits on the other (E. Dei-Cas & A.E. Wakefield, unpublished data), but the divergence levels were low, indicative of class II (strain-level divergence but not species-level variation) according to the Stringer criteria (Stringer, 1996). Data on the same locus in Pneumocystis from European wild rabbits published previously by other authors were consistent with these observations (Guillot, 1999), though the number of molecular studies on Pneumocystis species in wild mammals is still limited (Peters, 1994a; Bishop, 1997; Mazars, 1997b; Guillot, 1999).

Phylogeny of rabbit-associated Pneumocystis: concordance of gene genealogies

Both simple gene sequence comparison studies (mtLSU-rDNA, TS and arom gene) (Wakefield, 1992; Peters, 1994a, b; Mazars, 1995) and Pneumocystis phylogenetic trees constructed on the basis of sequence divergence at many loci (mtLSU-rDNA, mtSSU-rDNA, SODA, DHPS or DHFR genes), either individually (Wakefield, 1998; Guillot, 1999; Denis, 2000; Durand-Joly, 2000; Ma, 2001) or concatenated (this paper; Guillot, 2001; Keely, 2004), provided concordant results on the relationships of rabbit-associated with other Pneumocystis species, and on its place in the phylogeny of Pneumocystis. In sequence comparison studies, rabbit Pneumocystis gene sequences positioned constantly closer to P. jirovecii ones than to those of rodent Pneumocystis species (Fig. 3). Consistently, in phylogenetic trees rabbit-associated isolates emerge regularly as a monophyletic clade placed close to primate-derived Pneumocystis species, and relatively far from rodent-associated Pneumocystis strains (Fig. 3). Thus, the requirement of genealogical concordance (Taylor, 2000), comprehensively discussed by Keely (2004) in their recent description of P. murina, is clearly fulfilled for the rabbit-associated Pneumocystis species.

Figure 3

Maximum likelihood phylogeny of 18 Pneumocystis taxa inferred from mtLSU-rRNA and mtSSU-rRNA concatenated sequences. Bayesian posterior probabilities are given as percentages near the individual nodes. Nodes with values of <50% are not shown. Scale bar=0.1 substitutions (corrected) per base pair.

Rabbit-associated Pneumocystis: a new Pneumocystis species

The rabbit-associated Pneumocystis species presents a high level of genetic and phenotypic divergence from existing Pneumocystis species or formae speciales. Results summarized in this paper demonstrate that genetic divergence from Pneumocystis from hosts other than rabbits occurs throughout the genome as shown by the detailed analysis of eight independent loci, and five isoenzyme systems (Mazars, 1997a). High concordance between gene trees associated with the results of the population genetics approach suggests that the entity has been genetically isolated from the other Pneumocystis species or formae speciales for a very long time, and that it has undergone, therefore, a prolonged genetic and functional adaptation to the rabbit host (Oryctolagus cuniculus). Consistently, in cross-infection experiments, the entity was found to be a stenoxenous species (Aliouat, 1993b, 1994; Dei-Cas, 1994, 1998b), similar to other studied Pneumocystis species associated to specific animal hosts (Furuta, 1987; Gigliotti, 1993; Durand-Joly, 2002). Other phenotypic differences examined in this paper (ultrastructure, growth rate) illustrate the divergence between rabbit-associated and other Pneumocystis species or formae speciales.

According to the biological concept (Mayr, 1963), which hold a prominent place in mycology (Taylor, 2000), species are ‘groups of actually or potentially interbreeding natural populations, which are reproductively isolated from other such groups’. Though mating tests are impossible to apply in Pneumocystis due to their limited in vitro growth, present data indicate that rabbit-derived Pneumocystis natural populations share the same gene pool and do not exchange genes with other Pneumocystis populations. Consequently the rabbit-associated entity is a biological species. Furthermore, the concordance of topologies of multiple gene trees shows that this entity, which emerges as a clear monophyletic clade, is consistent with the concept of phylogenetic species, i.e. an evolutionary lineage that has a unique combination of DNA orthologue sequences (Taylor, 2000). Therefore, the entity satisfies the criteria, at the operational level, of both Biological Species Recognition (BSR), and Phylogenetic Species Recognition (PSR), in the sense of Taylor (Taylor, 2000).

Pneumocystis oryctolagi Dei-Cas, Chabé, Moukhlis, Durand-Joly, Aliouat, Stringer, Cushion, Noël, de Hoog, Guillot & Viscogliosi, sp. nov. MycoBank MB 500511. Figs 1 and 2

Pneumocystis oryctolagi (de Oryctolagus cuniculus, in quo fungus reperitur), quae ante Pneumocystis carinii f.sp. oryctolagi noscebatur (Anonymous, 1994)

Non filiosi extracellularesque fungi qui in Oryctolagus cuniculus alveolis pulmoneis inhabitant. Ad Typum Primum epithelii alveolas in alveolare epithelio sparsas haesi sunt. Quod trophicae sporocyticae cysticaeque formae congregatae sunt, quod quidem in Pneumocystis speciebus, quae e mordacibus primatibusque oriuntur, frequenter evenit, contra in Pneumocystis oryctolagi rarius est. Hujus speciei enim fungi alveolare lumen petunt in solis corporibus quae permulti parasitis inhabitant. Trophicae formae sunt 1–4 μm statura, uninucleatae, irregulares, tenuitunicatae plasmaticisque cum membranis in duobus lateribus. Asci (cysti) autem, 4–6 μm statura, crassitunicati, globosi duabus cum plasmaticis membranis sunt, in quibus octo rotundi ad ovatiles ascopores sunt, quisque 1–2 μm statura. Vacui, falciformi aut irregulares videntur. Pneumocystis oryctolagi morphologia non distinguitur ab aliis Pneumocystis speciebus, facili microscopio uso, sed ultrastructuralibus studiis patet Pneumocystis oryctolagi filipodia crassiora valdeque minus copiosa quam soricinis Pneumocystis (Dei-Cas, 1994; Nielsen, 1998). Quin etiam cysticae trophicaeque formae crebriores in Pneumocystis oryctolagi (0.10–0.15) quam in mordacum Pneumocystis speciebus (0.02–0.05). In vivo autem ad duplicandum celerior Pneumocystis oryctolagi est (1.7 dies) quam Pneumocystis carinii (4.5 dies) Pneumocystis murina (10.5 dies) (Aliouat, 1999). Ad alias externas dissimilitudines cognoscendas tabula prima consulenda est.

Pneumocystis oryctolagi dissimilior DNA ordine est quam aliae Pneumocystis species. Cujus mtLSU-rDNA loco genera inter 18.1% (a Pneumocystis jirovecii) et 29.1% (a Pneumocystis carinii) differunt. DNA ordinibus regionalia genera Pneumocystis oryctolagi SODA ordine differunt a Pneumocystis carinii 29.6% atque a Pneumocystis jirovecii 5.5%; DHPS ordine a quibusque 18% et 15%, DHFR ordine a quibusque 31% et 23% (Ma, 2001). Aminis acidis 97 conjectis e TS loci parte Pneumocystis oryctolagi differebat a Pneumocystis carinii 14.4% vel a Pneumocystis jirovecii 9.2% (Mazars, 1995). AROM loco Pneumocystis oryctolagi differt 25% sine amini acidi ordine a Pneumocystis carinii (Banerji, 1995). Cujus HSP70 inter 16.2% (a Pneumocystis murina) et 36.1% (a Pneumocystis wakefieldiae) differunt. DNA ordinibus regionalia genera Pneumocystis oryctolagi mtSSU-rRNA ordine differunt a Pneumocystis f.sp. macaca 14.5% atque a Pneumocystis wakefieldiae 18.1%.

Holotypus IPL-3609, e menstrui cuniculi pulmonibus (Officina Charles River, Rouen, Francia). Cryoservata exempla et electronae traditionis microscopii imagines primae servata sunt in Pastore Instituto apud Lillam (Lilla, Francia). Isotypus in Centraalbureau voor Schimmelcultures depositur.

Description of Pneumocystis oryctolagi sp. nov. Dei-Cas, Chabé, Moukhlis, Durand-Joly, Aliouat, Stringer, Cushion, Noël, de Hoog, Guillot & Viscogliosi, sp. nov. MycoBank MB 500511. Figs 1 and 2

Pneumocystis oryctolagi (L. adj. oryctolagi, of the rabbit, after the host in which the organism is found, Oryctolagus cuniculus)

Formerly known as Pneumocystis carinii f.sp. oryctolagi (Anonymous, 1994).

Nonmycelian extracellular fungal organisms resident in the pulmonary alveoli of Oryctolagus cuniculus. They attach to Type 1 epithelial alveolar cells lining the alveolar epithelium. Clustering of trophic, sporocytic and cystic forms, which occurs frequently in rodent- and primate-derived Pneumocystis species, is rather rare in P. oryctolagi. In this species, organisms extend into the alveolar lumen only in extensively parasitized hosts. The trophic forms, measuring 1–4 μm, are uninucleate, of irregular shape, thin-walled and with inner and outer plasma membranes. Asci (cysts), measuring 4–6 μm, are thick-walled, spheroid, with two plasma membranes and contain eight round to ovoid ascospores, each 1–2 μm. When empty, asci appear falciform or irregular. Pneumocystis oryctolagi is morphologically indistinguishable at the light microscopic level from other Pneumocystis species, but ultrastructural studies show the filopodia of P. oryctolagi to be thicker and clearly less abundant than those of Pneumocystis from mice (Dei-Cas, 1994; Nielsen, 1998). In addition, cystic-trophic form ratio is higher in P. oryctolagi (0.10–0.15) than in rodent Pneumocystis species (0.02–0.05). Likewise, in vivo doubling time of P. oryctolagi is shorter (1.7 days) than in P. carinii (4.5 days) or P. murina (10.5 days) (Aliouat, 1999). Other phenotypic differences are shown in the Table 1.

Pneumocystis oryctolagi is very different at the DNA sequence level from other Pneumocystis species. Genetic divergence at the mtLSU-rDNA locus ranged from 18.1% (between P. oryctolagi and P. jirovecii) to 29.1% (between P. oryctolagi and P. carinii). DNA sequences from regions in the genes of the P. oryctolagi SODA diverged from those of P. carinii by 29.6%, and P. jirovecii by 5.5%; DHPS by 18% and 15%; DHFR by 31% and 23% (Ma, 2001). Over 99 deduced amino acids of a fragment of the TS locus, divergence was 14.4% between P. oryctolagi and P. carinii, and 9.2% between P. oryctolagi and P. jirovecii (this paper, (Mazars, 1995). At the arom locus P. oryctolagi diverged by 25% (deduced amino acid sequence level) from P. carinii (this paper, Banerji, 1995). Divergence was similar in HSP70 (16.2–36.1%) and in the DNA sequence of the mtSSU-rRNA gene (14.5–18.1%).

The type strain is IPL-3609. Extracted from lungs of 1-month-old rabbits (Charles-River, Rouen, France). Cryopreserved samples and original transmission electron micrographies are stored at the Lille Pasteur Institute and at the Centraalbureau voor Schimmelcultures (Utrecht, The Netherlands).


We would like to thank Professors Walter Gams (CBS, Fungal Biodiversity Centre, Utrecht, The Netherlands) and Scott Redhead (Agriculture and Food Ottawa, Canada) for helpful advice in the field of fungal nomenclature and taxonomy, as well as Professors Maria-Lucia Taylor (UNAM, Mexico DF, Mexico), Maria-José Mendez Giannini (UNESP, Sao Paulo, Brazil), Rosely M. Zancopé-Oliveira (FIOCRUZ, Rio de Janeiro, Brazil), and François Delaporte (‘Jules Vernes’ University, Amiens, France) for providing valuable papers on Pneumocystis taxonomy published in the early 20th century. We thank Miss Nausicaa Gantois for technical assistance. The French Ministry of High Education and Research (EA3609 Lille 2 University), Lille Pasteur Institute, Spanish Ministry of Research and Technology (SAF2003-06061, 2003–2006), and European Commission (‘Eurocarinii’ FP-5-QLK2-CT-2000-01369) have supported this work.


View Abstract