Genotypes at the Internal Transcribed Spacers of the Nuclear rRNA Operon of Pneumocystis jiroveci in Nonimmunosuppressed Infants without Severe Pneumonia

Abstract

The frequency of Pneumocystis jiroveci (human-derived Pneumocystis) in immunocompetent infants developing acute respiratory syndromes has recently been evaluated and has been shown to be close to 25%. Until now, there have been no data on the genomic characteristics of the fungus in these patients, while molecular typing of P. jiroveci organisms was mostly performed with samples from immunosuppressed patients with pneumocystosis (Pneumocystis carinii pneumonia [PCP]). The present report describes the genotypes of P. jiroveci organisms in 26 nonimmunosuppressed infants developing a mild Pneumocystis infection contemporaneously with an episode of bronchioloalveolitis. The typing was based on sequence analysis of internal transcribed spacers (ITSs) 1 and 2 of the rRNA operon, followed by the use of two typing scores. By use of the first score, 11 P. jiroveci ITS types were identified: 10 were previously reported in immunosuppressed patients with PCP, while 1 was newly described. By use of the second score, 13 types were identified, of which 2 were newly described. The most frequent type was identified as type B₁a₃ (first score), which corresponds to type Eg (second score). Mixed infections were diagnosed in three infants. The occurrence of such diversity of P. jiroveci ITS types, an identical main type, and mixed infections has previously been reported in immunosuppressed patients with PCP. Thus, the P. jiroveci ITS genotypes detected in immunocompetent infants and immunosuppressed patients developing different forms of Pneumocystis infection share characteristics, suggesting that both groups of individuals make up a common human reservoir for the fungus. Finally, the frequency of P. jiroveci in nonimmunosuppressed infants with acute respiratory syndromes and the genotyping results provide evidence that this infant population is an important reservoir for the fungus.

Seroepidemiological surveys have suggested that humans commonly develop a Pneumocystis primary infection early in life (21, 30, 31). Until recently, it was assumed that this primary infection was asymptomatic (32). This hypothesis has been challenged by the results of two recent studies. Vargas et al. (39) have shown that acquisition of serum antibodies to a Pneumocystis sp. by immunocompetent infants can be asymptomatic but can also be contemporaneous with acute respiratory syndromes, during the course of which the fungus can be detected in nasopharyngeal aspirates (NPAs). We have recently reported positive results by PCR for Pneumocystis jiroveci (human-derived Pneumocystis) detection in NPAs from 45 of 178 nonpremature immunocompetent infants who presented with acute respiratory syndromes (26). Although no evaluation of the antibody response to Pneumocystis sp. antigens was performed in our study, the low mean age (4.7 months) of these infants who tested positive for P. jiroveci argues in favor of a first contact with the fungus. The two reports indicated that P. jiroveci occurs in NPAs from symptomatic nonpremature immunocompetent infants at frequencies of from 22 to 32%. Moreover, the results suggest that in this patient population, primary Pneumocystis infection may be revealed by an acute respiratory syndrome.

Molecular typing of P. jiroveci organisms was mostly performed with samples from immunosuppressed patients with pneumocystosis (Pneumocystis carinii pneumonia [PCP]) (3, 8-11, 15-19, 23, 24, 28, 36-38), while there have been no data concerning the genomic characteristics of the organisms involved in the infections of immunocompetent infants at risk for primary Pneumocystis infection. The aim of the present study was to type the P. jiroveci organisms obtained from the 45 infants mentioned above. The typing was performed by sequence analysis of internal transcribed spacers (ITSs) 1 and 2 of the nuclear rRNA operon, one of the most informative regions for P. jiroveci genotyping (19, 36, 37). The P. jiroveci ITS types identified in this infant population were then compared with those previously described in reports concerning immunosuppressed patients with PCP.

(The results of this study were reported in part at the 41st Interscience Conference on Antimicrobial Agents and Chemotherapy [G. Nevez, A. Totet, and C. Raccurt, Abstr. 41st Intersci. Conf. Antimicrob. Agents Chemother., abstr. J-268, 2001].)

MATERIALS AND METHODS

Patients and specimens.

This study was approved by the Ethics Commission of Picardy, France. The project was registered in France with the “Direction Générale de la Santé” (no. 990440). Forty-five archival NPAs obtained from 45 nonpremature immunocompetent infants (mean age, 4.7 months; age range, 1.9 to 11.8 months; sex ratio, 26 boys and 19 girls) were examined in this study. They initially tested positive for P. jiroveci by a PCR assay that amplifies a portion of the gene encoding the mitochondrial large-subunit (mtLSU) rRNA (26). DNAs extracted from NPAs were stored at −20°C until they were typed. The infant's characteristics are detailed in Table 1. All infants presented with an acute respiratory syndrome compatible with bronchioloalveolitis and had normal immunological function, as revealed by the absence of defects in blood lymphocytes and immunoglobulins. The presence of P. jiroveci merely reflected a mild infection, since clinical improvement was obtained for all infants with short-term hospitalization (from 1 to 12 days), despite the absence of antibiotic treatment for the fungal infection. The fungus was associated with bacteria or viruses in 35 infants, whereas it was detected alone in 10 infants.

TABLE 1.

Characteristics of 45 infants with bronchioloalveolitis in whose NPAs P. jiroveci was detected by PCR

Infant code	Sex	Age (mo)	Date of NPA retrieval (mo. day. yr.)	Microorganisms identified in NPAs			Duration of hospitalization (days)
				P. jiroveci	Viruses	Bacteria
E8	Ma	3.6	11.15.1999	+	RSV	Branhamella catarrhalis	3
E18	F	3.7	11.19.1999	+	RSV	B. catarrhalis	4
E26	F	9.4	11.24.1999	+	RSV	−	8
E30	F	5.2	11.26.1999	+	−	−	2
E42	F	1.9	12.02.1999	+	RSV	Haemophilus influenzaeStreptococcus pneumoniae	5
E46	F	5.8	12.03.1999	+	−	−	1
E50	F	8.7	12.07.1999	+	RSV	−	1
E55	M	2.1	12.10.1999	+	−	−	1
E57	M	3.8	12.11.1999	+	RSV	−	10
E64	M	4	12.23.1999	+	−	−	6
E71	F	4.8	12.20.1999	+	−	Klebsiella pneumoniae	9
E75	M	4.7	01.03.2000	+	RSV	H. influenzae	3
E76	F	6.5	01.03.2000	+	−	H. influenzae	3
E81	M	4.2	01.11.2000	+	RSV	B. catarrhalis	2
E87	F	3.4	01.18.2000	+	RSV	H. influenzae	2
E88	F	2.7	01.20.2000	+	RSV	S. pneumoniae	5
E89	M	7.7	01.20.2000	+	−	−	1
E90	M	3.1	01.25.2000	+	RSV	H. influenzae B. catarrhalis	3
E93	F	2.4	02.01.2000	+	−	−	4
E98	M	4.7	02.15.2000	+	RSV	H. influenzae	2
E101	F	3.9	02.25.2000	+	RSV	S. pneumoniae	4
E104	M	6.1	03.01.2000	+	−	−	5
E106	M	3.6	03.04.2000	+	−	Bordetella pertussis	7
E120	M	2.8	12.09.2000	+	RSV	−b	3
E122	F	6.1	12.10.2000	+	RSV	+	4
E123	M	2.5	12.12.2000	+	RSV	−	3
E132	M	11.8	12.20.2000	+	−	+	4
E135	F	5.8	12.21.2000	+	RSV	B. catarrhalis	3
E137	M	1.7	12.22.2000	+	−	−	1
E142	M	2.1	12.26.2000	+	RSV	−	4
E146	M	4.9	12.27.2000	+	RSV	−	7
E148	F	4.7	12.27.2000	+	−	B. catarrhalis	5
E149	M	4.5	12.28.2000	+	−	+	12
E150	F	3.7	12.28.2000	+	−	−	7
E152	M	4.2	12.29.2000	+	−	−	1
E155	M	3.9	01.02.2001	+	RSV	−	3
E159	F	4.7	01.02.2001	+	RSV	H. influenzae	3
E164	F	4.5	01.08.2001	+	−	B. catarrhalis	6
E166	F	4.8	01.08.2001	+	−	H. influenzae	3
E181	M	4.3	01.22.2001	+	RSV	H. influenzae	7
E188	M	11.6	02.01.2001	+	−	+	2
E190	M	2.7	02.06.2001	+	RSV	+	1
E195	M	5.2	02.15.2001	+	RSV	B. catarrhalis	6
E205	M	3.6	03.03.2001	+	−	B. catarrhalis	2
E215	M	5.7	04.04.2001	+	RSV	B. catarrhalis	3

^aAbbreviations and symbols: M, male; F, female; RSV, respiratory syncytial virus; +, positive result; −, negative result.

^bMicroscopic examination argued for bacterial infection, despite the absence of positive culture results.

P. jiroveci typing.

P. jiroveci typing was based on sequence analysis of the ITS 1 and ITS 2 regions of the nuclear rRNA operon. The ITS 1 and ITS 2 sequences were amplified by a nested PCR assay. The two rounds of PCR were performed under the same conditions. Each reaction mixture contained the following reagents at the indicated final concentrations: 10 mM Tris-HCl (pH 8.8), 0.1% Tween 20 (vol/vol), 3 mM MgCl₂, 200 μM each deoxynucleoside triphosphate (deoxynucleoside triphosphate set; Eurogentec, Seraing, Belgium), 1 μM each oligonucleotide primer, and 0.02 U of DNA polymerase (Red Goldstar DNA polymerase, Eurogentec) per μl. The first PCR round was done with primer pair N18SF (5′-GGT CTT CGG ACT GGC AGC-3′) and N26SRX (5′-TTA CTA AGG GAA TCC TTG TTA-3′), previously described by Tsolaki et al. (38), for 40 cycles consisting of denaturation at 94°C for 1.5 min, annealing at 55°C for 1.5 min, and extension at 72°C for 2 min. The second PCR round was performed with P. jiroveci-specific primer pair ITSF3 (5′-CTG CGG AAG GAT CAT TAG AAA-3′) and ITS2R3 (5′-GAT TTG AGA TTA AAA TTC TTG-3′) (37) for 40 cycles consisting of denaturation at 94°C for 1.5 min, annealing at 56°C for 1.5 min, and extension at 72°C for 2 min. The PCR products from the first and second rounds were electrophoresed on a 1.5% agarose gel containing ethidium bromide to visualize the expected bands (band sizes, 580 and 530 bp, respectively). To avoid contamination, each step was performed in different areas with different sets of micropipettes. The reagents used in the PCR mixtures were prepared in a laminar-flow cabinet. To monitor for possible contamination, negative controls were included in each PCR round. Second-round PCR products were purified by Microcon PCR (Millipore Corporation, Bedford, Mass.) and cloned into plasmid vector pGEM-T (pGEM-T Vector System II; Promega Corporation, Madison, Wis.), which was used for JM109 cell transformation according to the instructions of the manufacturer. In order to control the transformation, each positive clone was screened by a PCR assay with primers T7 (5′-GTA ATA CGA CTC ACT ATA G-3′) and SP6 (5′-ATT TAG GTG ACA CTA TAG AA-3′), designed, respectively, to the T7 and SP6 promoters flanking the cloning region. The PCR products were also revealed by electrophoresis on a 1.5% agarose gel with ethidium bromide to detect an expected band of 690 bp (530 bp from the insert and 160 bp from the T7 and SP6 promoters). Recombinant plasmids were sequenced from the two strands by using the same T7 and SP6 primers by the dideoxy chain termination method and with a semiautomatic sequencer (BigDye terminator method and 3700 sequencer, respectively; Applied Biosystems, Foster City, Calif.). The ITS 1 and ITS 2 sequences were analyzed by using ABI Prism software (version 3.3.1; model 3700; Applied Biosystems) and aligned by using Clustal W software (version 1.81). The ITS 1 and ITS 2 alleles were subsequently identified by using the typing scores described by Tsolaki et al. (36) and Lee et al. (19). A P. jiroveci ITS type is defined by combination of the types for the ITS 1 and ITS 2 alleles.

Nucleotide sequence accession numbers.

The nucleotide sequences of the new ITS 1 and ITS 2 alleles have been deposited in GenBank. The accession numbers are AY135711 and AY135712, respectively.

RESULTS

ITS amplification and cloning were successful for 26 of the 45 NPAs that initially tested positive for P. jiroveci by PCR for the mtLSU rRNA gene. Three clones were sequenced for 17 of the 26 NPAs, two clones were sequenced for 7 NPAs, and only one clone each was sequenced for each of the 2 remaining NPAs. A total of 67 clones from 26 NPAs were sequenced. The alignments of the ITS 1 and ITS 2 sequences are shown in Fig. 1 and 2, respectively. The results of P. jiroveci ITS type identification are shown in Table 2.

FIG.1.

Alignment of ITS 1 sequences from 67 clones. To the left of each line, the first number represents the NPA number and the second number represents the clone number. The first sequence is the consensus sequence for ITS 1 of P. jiroveci (19). Bases in boldface are implicated in the score established by Tsolaki et al. (36). Underlined bases are implicated in the score established by Lee et al. (19). Bases that are identical to those of the consensus sequence are indicated by periods, missing bases are indicated by hyphens, and bases that are different from those of the consensus sequence are given. The sequences at positions 61 to 70 and 91 to 100 are not shown because no sequence variations were found in these areas. “H,” an ITS 1 allele newly described in this study. Allele B₁ of Tsolaki et al. (36) can be allele E, J, or L of Lee et al. (19). Allele A₂ of Tsolaki et al. (36) can be allele A or B of Lee et al. (19).

FIG.2.

Alignment of ITS 2 sequences from 67 clones. To the left of each line, the first number represents the NPA number and the second number represents the clone number. The first sequence is the consensus sequence for ITS 2 of P. jiroveci (19). Gaps were introduced in the consensus sequence (between positions 60 and 61, 72 and 73, and 165 and 166) to facilitate alignment. Bases in boldface are implicated in the score established by Tsolaki et al. (36). Underlined bases are implicated in the score established by Lee et al. (19). Bases that are identical to those of the consensus sequence are indicated by periods, missing bases are indicated by hyphens, and bases that are different from those of the consensus sequence are given. The sequences at positions 1 to 40, 81 to 90, 101 to 120, and 141 to 150 are not shown because no sequence variations were found in these areas. “a₄” (also designated “f”) represents an ITS 2 allele newly described in this study.

TABLE 2.

Identification of P. jiroveci ITS types in 26 NPAs from 26 immunocompetent infants presenting with bronchioloalveolitis

Infant code	No. of clones	P. jiroveci ITS typea
		Tsolaki score	Lee score
E18	3	B1a3,b B1a4	Eg,b “H”c f
E30	3	B1a4	Jf
E42	3	Ca3,b B1a3	Fg,b Eg
E46	2	A2c1	Bl
E50	2	B1a3	Eg
E55	2	B1a3	Eg
E57	2	B1a3	Eg
E64	3	A2c1	Bl
E71	3	B2a1	Ne
E75	2	B1b2	Ec
E87	3	B2a1	Ne
E89	2	B1b2	Ec
E90	3	B2a1	Ne
E93	3	B2a1	Ne
E106	2	B1b1	Eb
E120	3	B1a4	Jf
E132	3	B1a3	Eg
E155	3	B1a3	Eg
E159	3	A2c1	Al
E164	3	B1“a4”c	J“f”c
E166	3	B1“a4”c	J“f”c
E188	1	B1b1	Eb
E190	1	B2a3	Ng
E195	3	B1d	Ea
E205	3	A2c1,b B1c1	Al,b El
E215	3	A2c1	Bl

^aTsolaki score, P. jiroveci ITS type identification by using the score of Tsolaki et al. (36); Lee score, P. jiroveci ITS type identification by using the score of Lee et al. (19).

^bThe major type, as identified in two of three clones.

^cAllele newly described in this study.

When the score for typing described by Tsolaki et al. (36) was considered, four ITS 1 alleles and eight ITS 2 alleles were found, leading to the identification of 11 P. jiroveci ITS types from the 26 NPAs. Type B₁a₃, which was detected in seven NPAs, was the most frequent type. Types A₂c₁, B₂a₁, and B₁a₄ were detected in five, four, and three NPAs, respectively. Types B₁b₂ and B₁b₁ were each found in two NPAs. Types Ca₃, B₂a₃, B₁d, and B₁c₁ were found in one NPA each. These 10 types have previously been described in reports concerning immunosuppressed patients with PCP (10, 24, 36-38). An 11th type had not been described before. It resulted from the combination of ITS 1 allele B₁ with a new ITS 2 allele, and we temporarily designated it “a₄.” This ITS 2 allele is close to allele a₄, described previously (36), but differs at scoring position 67, resulting in a change from an A to a G residue. The new type B₁“a₄” was found in two NPAs (E₁₆₄ and E₁₆₆).

The P. jiroveci ITS types were more diverse by application of the score of Lee et al. (19), since seven ITS 1 alleles and eight ITS 2 alleles were found, leading to the identification of 13 P. jiroveci ITS types. Among these 13 types, 11 (types Eg, Ne, Ec, Eb, Fg, Ng, Ea, El, and Jf, corresponding to types B₁a₃, B₂a₁, B₁b₂, B₁b₁, Ca₃, B₂a₃, B₁d, B₁c₁, and B₁a₄, respectively, described by Tsolaki et al. [36], and types Bl and Al, both corresponding to the same type [type A₂c₁] described by Tsolaki et al. [36]) were previously reported in immunosuppressed patients with PCP (19). Two other remaining types are newly described in this report. One new type results from the combination of ITS 2 allele f with a new ITS 1 allele that we temporarily designated allele “H.” ITS 1 allele “H” differs from the allele H previously described by Lee et al. (19) by insertion of two T residues at scoring positions 23 and 24. This new P. jiroveci ITS type (type “H”f) was detected in one NPA (E₁₈). A second new type resulted from the combination of ITS 1 allele J with a new ITS 2 allele that we temporarily designated “f.” This type was found in two NPAs (E₁₆₄ and E₁₆₆). ITS 2 allele “f” differs from the allele f previously described by Lee et al. (19) by a change from an A to a G residue at scoring position 68. In fact, this second new type (type J“f”) corresponds to new type B₁“a₄” identified above when the score of Tsolaki et al. (36) was applied.

More than one type was detected in 3 of the 26 NPAs (11.5%), suggesting the occurrence of mixed infections. The association of types B₁a₃ and B₁a₄ (corresponding to types Eg and “H”f, respectively, by using the scoring positions of Lee et al. [19]), C₃ and B₁a₃ (corresponding to types Fg and Eg, respectively, by using the scoring positions of Lee et al. [19]), and A₂c₁ and B₁c₁ (corresponding to types Al and Bl, respectively, by using the scoring positions of Lee et al. [19]) were found in NPAs E₁₈, E₄₂, andE₂₀₅, respectively. Since types B₁a₃ (Eg), Ca₃ (Fg), and A₂c₁ (Al) were identified in two of three clones obtained from NPAs E₁₈, E₄₂, and E₂₀₅, respectively, they were considered major types.

Other base changes were detected at positions other than scoring positions. In the ITS 1 sequence, substitutions from T to C at position 33 and G to A at position 87 were observed in clones E_164.02 and E_164.03. Another substitution, A to G at position 110, was detected in clones E_64.18 and E_106.02. Other sporadic changes in the ITS 1 sequence were also detected in only one clone each: an A-to-G substitution at position 9 in E_155.01, a T-to-C substitution at position 22 in E_75.19, a G-to-A substitution at position 72 in E_64.18, and a T-to-C substitution at position 120 in E_64.16. In the ITS 2 sequence, an insertion of T and A residues between positions 60 and 61 was detected in clone E_18.13; an insertion of A and T residues between positions 72 and 73 was detected in clones E_93.01, E_93.02, and E_93.03; and an A-to-G substitution at position 158 was detected in clones E_55.06 and E_57.08. Other sporadic changes in the ITS 2 sequence were detected in only one clone each: an A-to-G substitution at position 77 in E_120.04, an A-to-G substitution at position 95 in E_93.01, a C-to-T substitution at position 129 in E_71.03, and a T-to-C substitution at position 162 in E_195.01. Moreover, changes were observed in the length of a short homopolymeric (T) tract of the ITS 2 regions in four clones: instead of four T residues, three T residues were found in E_89.09 and E_89.16 and six T residues were found in E₄₂.₀₁ and E_42.03.

DISCUSSION

In this study, the first data concerning P. jiroveci ITS types from nonpremature immunocompetent infants at risk for primary infection were obtained. ITS 1 and ITS 2 loci were chosen for typing because the results obtained by use of these loci are considered more informative (19, 36, 37) than the results obtained by use of the mtLSU rRNA gene. However, ITS amplification failed to give positive results for 19 of the 45 NPAs that initially tested positive for P. jiroveci by use of the mtLSU rRNA gene (26). This difference in sensitivity between the two PCR assays has been observed previously, particularly for respiratory samples obtained by noninvasive means (23, 38). The high sensitivity of the mtLSU rRNA-based PCR assay has been explained by the fact that the mtLSU rRNA gene is present in many copies within each P. jiroveci genome, whereas ITS regions are present in only one copy (7, 35). Furthermore, because the fungus primarily infects the alveolar spaces and noninvasive sample collection essentially recovers cells from the upper respiratory tract (40), the amount of P. jiroveci organisms present in an NPA is usually low. Consequently, despite PCR amplification, the amount of P. jiroveci DNA remains small and the DNA cannot be directly sequenced. ITS PCR products were cloned in order to increase the quantity of amplified DNA to be sequenced. The cloning procedure was also performed in order to detect mixed infections more easily. Despite the relative lack of sensitivity of ITS amplification, our results show above all that genotyping of P. jiroveci by use of the ITS 1 and ITS 2 loci can be performed with NPAs from infants, as previously established by Tsolaki et al. (38) and Miller et al. (23) for other respiratory samples that can be obtained by noninvasive means, such as oropharyngeal samples and induced sputum from adults.

As the scoring nucleotide positions of Tsolaki et al. (36) are not strictly identical to those of Lee et al. (19), the four ITS 1 alleles identified by using the first score correspond to seven ITS 1 alleles identified by using the second score. Allele B₁ of Tsolaki et al. (36) can be allele E, J, or L of Lee et al. Allele A₂ of Tsolaki et al. (36) can be allele A or B of Lee et al. (19). One of these seven ITS 1 alleles is close to allele H previously described by Lee et al. (19), but it has a TT insertion at scoring positions 23 and 24. Because they are used as scoring positions, this allele was considered distinct from allele H and was designated “H.” Because nucleotide positions 23 and 24 are not included in the score of Tsolaki et al., this ITS 1 sequence corresponds to allele B₁ of Tsolaki et al. (36). Whatever score was used, eight ITS 2 alleles were identified. One of them, which we designated “a₄” (or “f” by application of the score of Lee et al. [19]), was also considered a new allele since a substitution from A to G at scoring position 67 (corresponding to scoring position 68 of the score of Lee et al. [19]) was not previously reported. Insertions of TA between positions 60 and 61 (clone E_18.13) and AT between positions 72 and 73 (clones E_93.01, E_93.02, and E_93.03) were not considered to result in new ITS 2 alleles since these base changes do not occur at scoring positions. However, they will probably have to be taken into account to define a new scoring system if their presence is confirmed by examination of other P. jiroveci organisms during future genotyping studies. Conversely, the significance of sporadic substitutions in the ITS 1 and ITS 2 sequences remains unclear. Other base changes observed in lengths of the homopolymeric (T) tract were excluded from analysis since PCR-induced error at this region has been reported previously (37).

By using the score of Lee et al. (19), two new P. jiroveci ITS types, types J“f” (corresponding to type B₁“a₄” by using the score of Tsolaki et al. [36]) and “H”f (corresponding to type B₁a₄ by using the score of Tsolaki et al. [36]) were identified. The hypothesis that P. jiroveci organisms of these types may affect only infant populations cannot be strictly ruled out. However, the majority of ITS types identified in this infant population were previously described in reports concerning immunosuppressed patients with PCP (10, 24, 36-38). The present data showing that identical P. jiroveci ITS types can be identified either in immunocompromised patients with PCP or in immunocompetent infants with mild Pneumocystis infection suggest the absence of a correlation between P. jiroveci ITS types and clinical profiles and other factors like age and immune status. Furthermore, no particular type was identified in infants infected only with P. jiroveci in the absence of bacteria or viruses. These findings are consistent with the fact that studies exploring whether there is a specific association of P. jiroveci ITS types in immunosuppressed patients with a defined clinical context have given contradictory results (8, 24).

The spectrum of polymorphism of the P. jiroveci ITS types found in the present work is similar to that previously reported for isolates from immunocompromised patients with PCP. Actually, applying the score of Tsolaki et al. (36), we have detected 11 types in 26 samples. Using the same score, Tsolaki et al. (37) and Miller et al. (23) have observed 10 types in 24 samples and 21 types in 43 samples, respectively. Applying the score of Lee et al. (19), we have detected 13 types. Using this score, Lee et al. (19) and Helweg-Larsen et al. (8) have observed 59 types in 207 samples and 49 types in 162 samples, respectively. Despite this diversity, three main types, types B₁a₃, A₂c₁, and B₂a₁ (corresponding to types Eg, Bl, and Ne, respectively, described by Lee et al. [19]), were identified in the infant population. Identical main types have previously been reported among nonepidemiologically linked isolates from immunosuppressed patients with PCP from diverse regions of Europe and the United States (10, 24, 36-38).

It has been suggested that Pneumocystis pneumonia in immunocompromised patients is not necessarily clonal (1, 37). The diagnosis of mixed infections, established for three infants (11.5%), suggests that mild Pneumocystis infection in immunocompetent infants is also not clonal. However, the rate of mixed infections was lower than that previously observed (27) in our hospital in immunosuppressed patients with PCP (11.5 versus 66%). It may be partially related to the relative low efficiency of detection of minor types in samples recovered by noninvasive means, like NPAs. This hypothesis is probable, since it has been suggested that the types obtained by use of respiratory samples may be assumed a priori to represent only the major types present within the lungs (9). The low rate of mixed infections may also partially be due to the inability to sequence more than one clone for two infants, which made the detection of mixed infections impossible.

This is the first report of P. jiroveci genotypes in nonpremature immunocompetent infants developing mild Pneumocystis infection. Comparison of the present data with those previously reported for immunocompromised patients with PCP makes it possible to establish that in both patient populations (i) identical P. jiroveci ITS types can be found, (ii) similar degrees of P. jiroveci ITS type diversity and similar main types can be observed, and (iii) mixed infections can occur. These shared features of the P. jiroveci ITS types detected in both patient populations suggest that fungus acquisition results from common sources.

Airborne transmission of the fungus from host to host has been demonstrated in rodent models (5, 12, 34), and several observations suggest that interindividual transmission occurs in humans (for a review, see reference 10). Moreover, it is now widely accepted that the Pneumocystis organisms infecting each mammalian species are host specific and that the existence of an animal reservoir for P. jiroveci can be excluded, according to the available data (6). Although an environmental reservoir remains possible, these data argue in favor of the fact that in humans PCP is an anthroponosis, with humans as the reservoir for P. jiroveci.

New detection tools such as PCR assays have revealed that humans with Pneumocystis infections can have a large spectrum of clinical presentations, of which PCP in immunocompromised patients may represent only a small part, while other clinical presentations may constitute the major part (4). Actually, it has been shown that pulmonary colonization with P. jiroveci occurs frequently in immunocompromised patients (20, 25) and less frequently in persons who are apparently immunocompetent but who are suffering from lung disease (2, 33). In a recent report, it was also shown that P. jiroveci organisms can transiently parasitize immunocompetent health care workers in close contact with PCP patients (23). The ITS typing method, which shows that the characteristics of the P. jiroveci organisms in the populations described above are similar, suggests that all of the populations can be a common reservoir of the fungus.

The combination of the high frequency of acute respiratory syndromes in immunocompetent infants and the high frequency of P. jiroveci infection in the course of these syndromes (26, 39) suggests that these usually undiagnosed cases of mild Pneumocystis infection in these infants may in fact represent the majority of P. jiroveci infections in humans. Consequently, immunocompetent infants may play a major role in the poorly understood human reservoir of the fungus. The present data, showing a commonality of the P. jiroveci types with those in other individuals infected with P. jiroveci, are consistent with this hypothesis.

Until recently, investigations of P. jiroveci transmission by genotyping have mainly been based on analyses of clusters of PCP cases among severely immunocompromised patients (10, 18, 22, 29). None have considered the potential role of other human populations that develop mild Pneumocystis infections and that therefore contribute to the circulation of the fungus. Miller et al. (23) have recently suggested that immunocompetent asymptomatic health care workers in close contact with PCP patients may have a potential role in the further circulation or transmission of the fungus. This hypothesis was prompted by the results of Dumoulin et al. (5), who have shown that immunocompetent and apparently asymptomatic mice can be transiently parasitized by Pneumocystis after a brief contact with SCID mice with PCP and were able to transmit the fungus to other susceptible mice. Thus, experimental results have established that Pneumocystis organisms are highly transmissible between hosts, which develop diverse forms of Pneumocystis parasitism (5, 12, 34). Moreover, the roles of neonates and nonimmunosuppressed adult rat populations colonized by the fungus were recently pointed out (13, 14). For a better understanding of the epidemiology of P. jiroveci in humans, immunocompetent infants developing a mild Pneumocystis infection as well as other individuals parasitized by the fungus, whatever the clinical profile, will have to be considered in further investigations of P. jiroveci circulation and transmission.