V. Genetic heterogeneity of rat-derivedPneumocystis

Two distinct levels of genetic diversity have been reported for Pneumocystis populations harbored by rat colonies. The first, more dramatic level of diversity, was exhibited by 2 populations that were found to co-exist within individual rat lungs and retain genetic identity. These 2 populations were originally called ‘prototype and variant’[1, 2], based on fidelity to and divergence from the small ribosomal subunit gene sequences originally reported for rat derived Pneumocystis[3, 4]. Subsequently, these populations have been awarded ‘special form’ status. Pneumocystis carinii f.sp. carinii and Pneumocystis carinii f.sp. ratti are now used to designate ‘prototype’ and ‘variant’ populations, respectively [5, 6] (discussed below). The second level of diversity was more subtle and characterized by low levels of genetic polymorphisms among karyotypic forms of the ‘prototype’ population. The levels of divergence between ‘prototype’ and ‘variant’ were characterized as Class II and those among the ‘prototype’ populations as Class I using the genetic ranking system described by Stringer [7]. This system of classification is based on sequence comparisons at 5 genetic loci: thymidylate synthase, mitochondrial large subunit rRNA, β-tubulin, arom, and internal transcribed spacer regions (ITS) in the nuclear ribosomal RNA locus. The highest level of divergence was observed among Pneumocystis populations from different mammalian hosts and ranged between 15–50% divergence at the selected loci, with the ITS regions being most divergent. Class II divergence ranged from 4–7% for all genes, but within the ITS regions a 20–30% divergence was observed. This level of divergence approaches that seen among bonafide species of other microbes. Class I sequences differed by 0–0.8% in the 4 gene sequences and by 2–4% in the ITS regions. Differences among Pneumocystis populations obtained from human beings have been shown to only vary at a Class I level, as well as populations of rat P. carinii defined by karyotypic profiles (see below). Because infections with more than 1 Pneumocystis population may hold implications for the prognosis of the host, or, like other microbes, P. carinii populations may manifest different degrees of virulence, a review of the genetic distinctions that differentiate these organism populations should provide the necessary characterizations to begin such evaluations. Eight forms of P. carinii f.sp. carinii have been defined by stable and reproducible band patterns produced by pulsed field gel electrophoresis (Fig. 1) and hybridization with probes to single copy genes [1, 2, 8]. The karyotype profile, form 5, was first reported by another laboratory [9]. The rats harboring forms 1–4 and 6–8 P. carinii f.sp. carinii populations were obtained from commercial vendors in filtered-barrier containers, immediately placed under barrier in the animal facilities at the University of Cincinnati College of Medicine, and subsequently immunosuppressed for 6–15 weeks by weekly injections of 4 mg kg−1 methylprednisolone [1, 2]. Agarose embedded organisms were prepared by homogenization of individual rat lungs in a phosphate buffered solution using a Stomacher 80 lab Blender with subsequent hypotonic lysis of the homogenate to reduce host cell contamination, as previously described [1, 2]. Organisms embedded in 0.8% low melt agarose were subjected to digestion with proteinase K in a detergent based solution for 24–48 h at 55°C. Chromosome sized-DNA bands were separated by Countour Clamped Homogeneous Electrical Field (CHEF) electrophoresis at a 120° angle of reorientation,~3.8 V cm−1, 14°C, in low ionic strength buffer for 5 days. Bands were visualized by ethidium bromide staining. Karyotypic forms of Pneumocystis carinii f.sp. carinii. Lambda concatamers were used as size markers. The first panel represents a single gel run containing forms 1–7; the second panel is another gel run with form 8 only. Lanes: (1) form 1, M3610; (2) form 2, M3616; (3) form 3, M813; (4) form 4, 91–1063; (5) form 5, M1919; (6) form 6, 92–1007; (7) form 7, 92–1047; (8) form 8, M151. Samples 1–3 and 5 resulted from infections induced by intratracheal inoculations of cryopreserved P. carinii populations, while the remaining samples were obtained from uninoculated, barrier housed, immunosuppressed rats (‘natural infection’). Several important concepts evolved from these studies: (i) about 50% of the colonies harbored mixed P. carinii f.sp. carinii-P. carinii f.sp. ratti as the infecting populations; (ii) of the colonies harboring infections of a single P. carinii type, form 1 P. carinii f.sp. carinii was the most common; (iii) form 1 P. carinii f.sp. carinii was the most common form of the organism found in co-infections with P. carinii f.sp. ratti; (iv) the electrophoretic karyotype of a rat colony remained stable over time if unperturbed by changes in animal husbandry practices; and (v) Pneumocystis were harbored by immune competent rats. This last concept was shown by our standard operating procedure for provocation of pneumocystosis in rats. We received immune competent rats from the commercial vendor in filtered containers. They were immediately placed in barrier housing to prevent contact with the environment, and remained therein throughout the entire experiment. Upon immunosuppression with corticosteroids for 6–10 weeks, a fulminant Pneumocystis pneumonia ensued. The only possible source of organisms was from the immune competent rats themselves. These animals must have harbored the organisms, albeit in low numbers, which then were able to expand after chronic immunosuppressive therapy. To determine the genetic relatedness of the P. carinii f.sp. carinii populations defined by karyotypic profiles, regions of the nuclear small subunit rRNA (srRNA) and mitochondrial large subunit rRNA (lsrRNA) from at least 3 representatives of each of P. carinii f.sp. carinii karyotype forms 1, 2, 3, 4 and 6, and 2 representatives of forms 7 and 8 were amplified by the polymerase chain reaction (PCR) and sequenced for comparative analyses. Three representatives of P. carinii f.sp. ratti were also sequenced at the same loci. Form 5 P. carinii f.sp. carinii was not available for analysis. Primers were directed to the 3′ end of the srRNA encompassing the 390-bp intron [1–3] and to a 346-bp region of the lsrRNA using primers pAZ102, H and E [10]. Purified amplicons were sequenced using dsDNA cycle sequencing (forms 1–4, 6, and P. carinii f.sp. ratti) or submitted for sequencing at the University of Cincinnati DNA Core Facility (Forms 7, 8). Analysis of a~300-bp region of the lsrRNA gene showed no polymorphisms among the karyotypic forms 1, 2, 3, 4, 6, 7, and 8 (Fig. 2). When compared to the P. carinii f.sp. ratti sequence, there were 26-bp changes in 277 bp of the P. c. ratti gene disregarding gaps or insertions (9% divergence) or 40 differences between the 2 populations if gaps and insertions were included (14% divergence). Of the nucleotide changes,~46% were transversions (pyrimidines to purines or vice versa) and 54% were transitions (purine-purine or pyrimidine-pyrimidine changes). More purine switches (AG or GA) occurred in the transitions (~71%) than pyrimidine (TC or CT) (~29%). These changes may be noteworthy because transversions in the mitochondrial rRNA genes appear to have a linear relationship with time over the past 75 million years, whereas transitions in these genes represent more rapid changes and are therefore less reliable for higher level (ordinal or supraordinal) phylogenetic associations [11]. The lsrRNA sequences of the P. c. carinii forms were identical to the 205-bp sequence published for rat by Peters et al. [12], but differed from the 277 bp of rat P. carinii sequence reported by Lee et al. [13]. In the latter sequence, 2 additional Gs were present at nucleotides (nt) 107 and 110 and one base transversion from a C to a G at nt 174 (nucleotide numbers as per Lee et al. [13]). It is unclear if these differences represent real divergence among P. c. carinii genomes as the karyotypic profile(s) of the organisms used by Lee et al. [13] were not reported. Large subunit mitochondrial rRNA sequences of Pneumocystis carinii f.sp. carinii and Pneumocystis carinii f.sp. ratti. Forms 6, 7, and 8 had identical sequences and are not shown. Although P. c. carinii forms 1, 3, 4, 6, 7, and 8 shared identical srRNA sequences, there were 2 base changes in the sequences of form 2 samples (Fig. 3). These changes were not due to sequencing errors, since the changes were apparent in all four of the form 2 amplicons directly sequenced. Subsequent analyses of form 1 and form 2 genomes by Keely et al. [14] supported the subtle but distinct difference between form 2 genomes and the other P. c. carinii forms. Primers directed to an upstream region of the α-repeat, a region containing multiple copies of a telomere-like sequence, were used to amplify products from forms 1–4 and 6–8. Gel electrophoresis of the amplicons revealed a product of about 520 bp sequence among all the forms with the exception of form 2. The amplicons from form 2 genomes had an approximate size of 600 bp. Sequencing of the form 2 amplicons revealed that form 2 contained additional telomere-like repeats that were responsible for the size difference. In addition, further analysis of the α-repeat sequence of forms 1 and 2 revealed several nucleotide differences. No evidence of sequence divergence, however, was found in 2 other genetic loci in the same populations; the ITS2 and lsrRNA sequences were found to be invariant among the prototype forms. Nuclear small subunit rRNA sequences of forms 1–4 Pneumocystis carinii f.sp. carinii. Forms 6–8 had identical sequences and are not shown. Form 2 populations of P. c. carinii organisms appear to exhibit subtle, yet reproducible genetic distinctions that differentiate them from the other populations defined by karyotypic profiles. It is interesting to note that form 2-form 1 dual infections have been observed in commercial rat colonies and can be experimentally induced in immunosuppressed, P. carinii-naive rats, which is similar to the observation of naturally infected rats with P. c. carinii-P. c. ratti co-infections [1, 2]. In contrast, few other P. c. carinii-P. c. carinii dual infections have been observed by us in surveys of commercial rat colonies. It is tempting to speculate that the genetic distinctions between form 1 and form 2 populations, although not as dramatic as the P. c. carinii-P. c. ratti divergence, are different enough to permit establishment of the 2 populations within a single environment, the rat lung, as has been reported for the latter 2 populations. The rRNA gene sequences presented in Figs. 2 and 3 have been submitted to GenBank under the following names: lsrRNA for forms 1–4, U20169–U20172; variant lsrRNA, U20173; srRNA for forms 1–4, U20174–U20177. Until the 1990s, the Pneumocystis organisms found within the lungs of immunosuppressed rats were considered as a single population. Morphologically indistinguishable at light or electron microscopic levels, there were few techniques that would permit genuine discrimination without interference from contaminating host cells. In 1993, 2 clearly resolved electrophoretic karyotypes produced by pulsed field gel analyses provided the first indication that rats from commercial vendors that were chronically immunosuppressed to provoke fulminant pneumocystosis could harbor 2 distinct Pneumocystis populations [1] (Fig. 4). In another report, these 2 populations were shown to be genetically distinct by sequence divergence of the nuclear small ribosomal subunit genes (srRNA); localization of the ribosomal locus to different sized chromosomes in each karyotype; and hybridization differences of a probe to the repetitive major surface glycoprotein gene family (MSG) between the Southern blotted karyotypes [2]. The level of divergence observed between the 2 populations in a variable region of the srRNA gene was approximately 6.6% in 120 bp of sequence. More striking was the absence of a 360-bp self-splicing intron found in the 3′ region of one population and not in the other, suggesting that these Pneumocystis organisms were quite distinct. The degrees of difference observed between these 2 P. carinii populations approached those observed between bonafide fungal species (e.g. Candida albicans, C. tropicalis). To date, the prototype and variant populations of P. carinii have been shown to exhibit significant genetic polymorphisms in the genes and regions within the nuclear ribosomal RNA locus [2, 15]; the large subunit mitochondrial locus (Fig. 2); in 3 single copy gene loci, the TATA-binding protein [16], the α-subunit of the G protein [17], and the thymidylate synthase [18]; and in the major surface glycoprotein genes [19] as well as in antigenic reactivities [20]. Profound antigenic differences were detected by immunoblotting studies with monoclonal antibodies and polyclonal antisera. Vasquez et al. [20] showed that monoclonal antibodies raised against whole or solubilized ‘prototype’P. carinii did not react with blotted P. carinii proteins of ‘variant’ separated by SDS-gel electrophoresis. The immunoblotting technique was shown to detect low levels of P. carinii f.sp. carinii in a mixed infection of predominantly P. carinii f.sp. ratti using monoclonal antibodies directed to antigens of P. carinii f.sp. carinii, providing a candidate diagnostic tool. Karyotypic profiles of Pneumocystis carinii f.sp. carinii and Pneumocystis carinii f.sp. ratti. Lambda concatamers were used as size markers. Lanes: (1) P. carinii f.sp. ratti, ‘variant’; (2) P. carinii f.sp. carinii form 1, ‘prototype’; (3) mixed infection of P. carinii f.sp. ratti and P. carinii f.sp. carinii form 1 in approximately equal proportions. Distinct profiles produced by pulsed field gel electrophoretic techniques permit unequivocal discernment of the identification of Pneumocystis populations infected rat colonies [1, 2]. Hybridization of the Southern-blotted karyotypes with the putative single copy genes of P. carinii, inosine monophosphate dehydrogenase (IMPDH) [21] and enolase, revealed striking differences in the sizes of chromosomal sized DNA bands to which the probes hybridized. IMPDH hybridized to a band of 430 kb in prototype forms 1–5 and to a band of 630 kb in variant [21]; enolase hybridized to bands of 530 kb and 340 kb, respectively, in protoype and variant populations (Fox, D. and Smulian, A.G., unpublished data). In recognition of these differences, the 2 genetically distinct populations were designated with a ‘special form’ status by the Pneumocystis community of investigators at the 3rd International Workshop on Pneumocystis, Cleveland, OH, 1994 in a session dedicated to taxonomy and nomenclature [5]. Thus, the rat organism population with fidelity to the original srRNA sequence published (‘prototype’) was designated ‘Pneumocystis carinii f.sp. carinii’, while that population which showed polymorphisms in the srRNA gene (‘variant’) was termed ‘Pneumocystis carinii f.sp. ratti’. These designations were recently discussed and upheld at the 5th International Workshop on Pneumocystis, Lille, France, 1997 [6]. Observations in the fluctuations of the number of rats co-infected with P. c. carinii and P. c. ratti and the relative ratios of those populations within a single rat lung in 2 rat colonies harboring these 2 populations led us to postulate that they were in competition within these fixed environments. Densitometric analyses of the ethidium bromide intensities of bands within karyotypic profiles of organisms produced from Brown Norway and Long Evans rat colonies sampled over several years showed that, indeed, the populations were in flux, and fulfilled predictions made by the Lotka-Volterra Competition model [22], suggesting that a bonafide competition was occurring between these 2 genetically distinct P. carinii populations. From the calculations it was surmised that the P. c. carinii populations were more dominant than the P. c. ratti populations and eventually the latter organism population would be eliminated if the closed environmental conditions persisted. The impact of the competition on the rat host has not yet been addressed, but preliminary studies comparing the sensitivity of the 2 populations to pentamidine and atovaquone showed no differences in minimum inhibitory concentrations for either drug, suggesting that drug susceptibility may not be a phenotype manifested by either of these populations. In the reported surveys of commercial rat colonies, few (if any) infections consisting of 2 karyotypic forms of P. carinii f.sp. carinii were observed [1, 2]. In some cases (e.g. form 1, form 4), discernment of such a co-infection would be difficult because the superimposed karyotypic profiles may not be fully appreciated as a dual infection because of the similar band sizes and thus may have gone undetected. My laboratory conducted a series of studies to determine if co-infections with different karyotype forms were experimentally possible (unpublished data). Forms 1, 2 and 6 were co-inoculated in equal numbers, inoculated 10 days apart, and inoculated 20 days apart. Results showed that form 2 organism populations could establish co-infections with form 1 organisms if inoculated 0 or 10 days apart, but not after 20 days. Form 1 and form 6 co-infections were only established if co-inoculated at the same time; form 2 and 6 co-infections were not attempted. These studies showed that there were specific compatibility differences among the karyotypic forms that could explain the lack of dual infections caused by P. c. carinii-P. c. carinii populations in commercial rat colonies. Distinct genetic populations of P. carinii have been defined by gene sequence variance, chromosomal patterns, and immunologic reactivities. These 2 populations, P. carinii f.sp. carinii and P. carinii f.sp. ratti, are capable of coincidently infecting an individual rat lung. These populations do not appear to mix genetically, suggesting that they may indeed be different species, although the availability of genetic markers with which to test this hypothesis is currently limited. These 2 populations appear to obey the tenets for competition outlined by the Lotka-Volterra Competition model. The P. carinii f.sp. carinii population is predicted to be the eventual dominant population in the surveyed colonies. P. c. carinii populations could be divided into 8 distinct populations based on karyotypic profiles produced by CHEF. These populations were relatively invariant at 2 genetic loci with the exception of the form 2 populations which had low level divergence in the srRNA and α-repeat loci. P. c. carinii-P. c. carinii co-infections were not observed in rat colony surveys although form 2 and form 1 and form 1 and form 6 co-infections could be experimentally induced and were dependent upon timing of inoculations and karyotypic populations. The prognostic implications of infection of a host by P. carinii population of Class I or Class II or a mixed infection are not yet appreciated and require further investigation. These studies unequivocably show that 2 genetically distinct populations within the same organism group are able to grow and compete within the same environment. The studies described here were supported by grants from the NIH and the VA Merit Review system. Many individuals participated in these projects and for their participation I am grateful: Michael Kaselis, Marnie Ryan, Sally Orr, Margaret Collins, Michael J. Linke, A. George Smulian, Peter D. Walzer, James R. Stringer, Paul Steele and Jonathan Arnold.