Phytosterols Are Present in Pneumocystis carinii - Antimicrobial ...

10 downloads 0 Views 1MB Size Report
Although originally classified as a protozoan, Pneumocystis carinii is now ... phylogenetic relationship between P. carinii and fungi and suggest that these sterols ...
Vol. 38, No. 11

ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Nov. 1994, p. 2534-2540

0066-4804/94/$04.00+0 Copyright X) 1994, American Society for Microbiology

Phytosterols Are Present in Pneumocystis carinii STEPHEN T. FURLONG,1,2* JULIE A. SAMIA,3 RICHARD M. ROSE,4

AND

JAY A. FISHMAN"13

Department of Medicine, Harvard Medical School,' Department of Rheumatology/Immunolov, Brigham and Women 's Hospital,2 Infectious Disease Unit, Massachusetts General Hospital, and Division of Pulmonary Medicine, New England Deaconess Hospital,4 Boston, Massachusetts Received 24 February 1994/Returned for modification 11 July 1994/Accepted 17 August 1994

Although originally classified as a protozoan, Pneumocystis carinii is now considered to have fungal characteristics. Drugs typically used for the treatment of fungal infections target ergosterol. Because P. carinii is an important pathogen in AIDS and other immunocompromised patients, knowledge of the sterol content of this organism may be useful as a basis for developing new treatment strategies or for improving diagnosis. P. carinii organisms were harvested from infected rat lungs and were purified by filtration. Control preparations from uninfected animals were identically prepared. Lipids were extracted from the organisms and control preparations and were separated into neutral lipid, glycolipid, and phospholipid fractions by silicic acid chromatography. The neutral lipid fraction was further treated by alkaline hydrolysis and was analyzed by reversed-phase high-pressure liquid chromatography (HPLC), gas chromatography (GC), and GC-mass spectrometry (GC-MS). As shown by HPLC, the neutral lipid fraction from infected rats contained a minimum of six peaks, while in control preparations a single peak with a retention time identical to that of cholesterol was observed. The predominant sterol in these preparations was positively identified by GC-MS as cholesterol and constituted 80 to 90% of the total. The remaining peaks had relative retention times similar to those of phytosterols by both HPLC and GC, and the similarity of these sterols to those derived from plants and fungi was confirmed by MS. Ergosterol, however, was not present. These results provide further evidence for a close phylogenetic relationship between P. carinii and fungi and suggest that these sterols could be used as targets for drug development and for improving diagnosis.

single most important target for therapeutic fungicidal drugs (33). These drugs act either by binding to fungal ergosterolcontaining membranes, causing altered membrane permeability (amphotericin B), or by interfering with the synthesis of fungal sterols (azoles). Mammalian cell membranes almost exclusively contain cholesterol, while fungal membranes most often contain ergosterol. However, other fungal sterols are also common. Because plant and fungal sterols are often similar, both types are sometimes referred to as phytosterols. A differentiating feature of the sterols of plants, protozoa, and fungi is the side chain configuration at C-24 which separates the algae and fungi from the vascular plants (36). Other lipids have been described in microorganisms that are not present in mammalian cells and could provide clues to phylogeny. Examples include diacylglyceryl-trimethylhomoserine (DGTS) in Acanthamoeba castellanii (15) and phosphonolipids in Tetrahymena, Trypanosoma, and Tetraurelia spp. (1, 11, 25). Few lipid-related studies have been done on P. carinii. Morphological studies with filipin or digitonin and freeze fracture have shown the presence of sterols by electron microscopy in the P. carinii plasma membrane (37, 38), and biochemical evidence for the presence of cholesterol has been reported (19). An unusual outer membrane has also been reported on the cyst wall (7). Previous reports of the phospholipid (26) and the fatty acid (19) contents of P. carinii have shown high degrees of similarity with the lipid contents of the host cells. The only P. carinii sterol described thus far is cholesterol (19). There have been no reports of lipids unique to P. carinii and no reports of significant differences in lipid content between P. carinii and host cells. Insight into the organism's lipid biochemistry would allow prediction of the susceptibility of P. carinii to drugs acting on organism-specific lipids or biosynthetic pathways. Such information would also be useful for clarifying the phylogenetic relationship between P. carinii and other micro-

Pneumocystis carinii is one of the leading causes of opportunistic infection in AIDS patients (30). Despite the importance of this microorganism as an opportunistic pathogen, much basic information about the parasite is lacking. While originally classified phylogenetically as a protozoan, analyses of P. carinii DNA and mitochondrial and rRNA sequences have generally shown a higher degree of homology to the mitochondrial and rRNA sequences of fungi (yeasts) than to those of protozoa (9, 13, 27, 31, 34, 35). The exact relationship remains to be determined. Analyses of 5S RNA have suggested a somewhat close phylogenetic relationship to members of the Rhizopoda-Myxomycota-Zygomycota group of "protista fungi." Other studies have suggested similarities to Saccharomyces, Candida, and Neurospora spp. and "red yeast" fungi. In general, the arguments in favor of classifying P. carinii as a protozoan are the amoeboid appearance of the trophic form, susceptibility to antiprotozoal drugs such as pentamidine isethionate or trimethoprim-sulfamethoxazole, insusceptibility to antifungal drugs such as amphotericin B, and the lack of ergosterol in Pneumocystis cell membranes. Arguments in favor of including P. carinii among the fungi include similarities of the sporogenous state to ascospore formation in yeasts, poorly developed mitochondria which show ultrastructural similarities to fungal mitochondria, lack of the internal organelles typically found in protozoa, the encoding of the enzymes thymidylate synthase and dihydrofolate synthase by physically separate genes, and the nucleic acid sequence homologies described above (10, 30). Fungal lipids, or more specifically, fungal sterols, are the * Corresponding author. Mailing address: 522 S. G. Mudd Building, 250 Longwood Avenue, Boston, MA 02115. Phone: (617) 432-1593. Fax: (617) 432-2799. Internet: [email protected].

2534

STEROL CONTENT OF P.

VOL. 38, 1994

organisms. The identification of P. carinii-specific lipids may also serve as a marker for infection. In the study described here examination of the lipid composition of P. carinii by highpressure liquid chromatography (HPLC), gas chromatography (GC), and GC-mass spectrometry (GC-MS) demonstrated the presence of sterols in this organism that have not been previously identified. MATERIALS AND METHODS Preparation of P. carinii and Physarum polycephalum. P. carinii isolates were derived from steroid-treated, virus-free, immunosuppressed rats infected by intratracheal inoculation (2, 3, 22). Sprague-Dawley rats (virus and pathogen free; barrier raised; weight, 150 to 174 g; Harlan Sprague-Dawley, Frederick, Md.) were fed water containing dexamethasone (1.2 mg/liter) and tetracycline (500 mg/liter) ad libitum. Rats received a normocaloric, low-protein (8%) rat diet (ICN Biochemicals, Costa Mesa, Calif.). The animals were housed in microisolated cages in a laminar flow rack. After 7 days on this regimen, rats were inoculated intratracheally with 1 x 107 to 3 x 107 freshly isolated P. carinii (including up to 8% cyst forms on Giemsa-stained smears). Inoculation was performed under direct tracheal visualization by using 0.1 ml of P. carinii in sterile phosphate-buffered saline (PBS) and then 0.25 ml of air through a 22- or 23-gauge needle. Control animals were treated with the same dietary regimen and steroids and were inoculated with 0.1 ml of sterile PBS without organisms. Additional controls included animals receiving antibiotic prophylaxis with trimethoprim-sulfamethoxazole (400 mg of trimethoprim and 1 g of sulfamethoxazole per liter of drinking water; Di-Trim; Animal Health, West Des Moines, Iowa). Lungs from each group of animals were harvested under sterile conditions after 6 to 12 weeks. Giemsa-stained impression smears were made from the cut surface of each lung (infected or uninfected) studied. Each lung was cultured for bacteria and fungi on blood agar and Sabouraud's medium and was individually chopped and homogenized in a Stomacher apparatus (Tekmar Co., Cincinnati, Ohio). The supernatant was diluted to 10 ml with Eagle minimal essential medium and was serially filtered through polycarbonate filters with pore sizes of 10, 8, and 5 (twice) jim (Poretics, Livermore, Calif.), and the organisms were collected by centrifugation at 1,800 x g for 15 min (8, 22). Red cells were lysed with ammonium chloride buffer, and the pellet was washed twice and resuspended in the desired medium. Smears of 10 ,ul of organisms were spread over a 1-cm2 premarked slide and were stained with DiffQuik (Baxter, McGraw Park, Ill.) for counting of nuclei and host cells. Organisms were counted by oil immersion microscopy; 10 contiguous fields in three smears of each preparation were counted. No intact nucleated cells or erythrocytes were present after filtration. Typically, less than 5% of the total DNA in these preparations is of mammalian origin. P. polycephalum was grown on agar plates containing rolled oats in the dark at room temperature. Mature cultures were harvested by carefully separating the fungus from the oat flakes with an inoculation loop and then washing the plates with buffer. Extraction, separation, and quantitation of lipids. Procedures for extracting, separating, and quantitating lipids have been described previously (14, 16, 18). In brief, lipids were extracted from organisms with chloroform-methanol (12). Separation of this lipid extract into crude neutral lipid, glycolipid, and phospholipid fractions was achieved by elution from silicic acid columns prepared in 9-in. (22.86-cm) Pasteur pipettes with chloroform (neutral lipids), acetone (glycolipids), and

CARINII

2535

methanol (phospholipids). The total lipid content from P.

carindi preparations wastneasured by separation with a ternary-gradient HPLC system with mass detection as described previously (16, 18). Samples were resuspended in 50 jl of chloroform-hexane (1:1). A total of 45 jl of sample was injected onto a column (10 cm by 4.6 mm) packed with Spherisorb with a guard column (5 cm by 4.6 mm) packed with 10-jim Spherisorb. Lipids were eluted with a ternary gradient. Mobile phases were hexane-tetrahydrofuran (99:1), chloroform-2-propanol (1:4), and 2-propanol-water (1:1). The eluted lipids were monitored with an Applied Chromatography Systems model 750/14 mass detector with the following settings: evaporator temperature, 90°C; internal air pressure, 27 lb/in2. The nebulizing gas source was a liquid N2 tank equipped with a gaseous N2 outlet and pressure builder. The detector output was digitized with either a model 760 or a model 790 interface and was analyzed with Nelson Analytical chromatography software. Standard curves were prepared from lipid standards purchased from Avanti Polar Lipids (Birmingham, Ala.). Separation and characterization of sterols. To identify the sterols in the neutral lipid fraction, total neutral lipid fractions were saponified in sodium methoxide for 2 h at 70°C (15) and were reextracted with 5 volumes of hexane. For separation by HPLC, samples were resuspended in chloroform-hexane (1:1) and were separated isocratically on a C18 Nucleosil column (250 mm by 5 jim) with a mobile phase of ethanol-methanolwater and were detected with a fixed-length UV detector set at 214 nm, with the data collected and analyzed as described above. By reversed-phase HPLC, sterols have characteristic retention times which are dependent on their structures (24, 28). As one method of identifying these unknown lipids, a series of sterol standards was run under conditions identical to those under which the P. carinii samples were run, and the retention times were measured. To compare better these

3-,um

HPLC data

with

those from other laboratories, a retention

index relative to that of P-sitosterol was calculated by using the following formula K = (Vi -V)/VO, where K is the capacity factor, VO is the retention volume of a nonretained solute, and Vi is the retention volume of the compound of interest. By using the calculated K values, a separation factor (a) was calculated by the equation a = K21K1, here K2 is ,B-sitosterol and K1 is the compound of interest. For analysis by GC and GC-MS, sterol acetates were prepared from the saponified neutral lipid fractions by resuspending the dried samples in pyridine-acetic anhydride (1:1) and the mixture was incubated overnight at room temperature. The solvents were then dried under nitrogen and were resuspended in chloroform. The sterol acetates were separated either on an Ultra-1 column (25 m by 0.22 mm by 0.33 jim) at 280°C with flame ionization detection or on an HP-5 column (30 m by 0.25 mm by 0.25 jim) with temperature programming (75°C for 2 min and then 20°C per min to 310°C). Electron-impact MS conditions were 1.8 scans per s (40 to 650 AMU). Testing for other lipids present in lipid extract. To test for the presence of DGTS (found in Acanthamoeba castellanii), the total lipid extract from P. carinii was separated by either one- or two-dimensional chromatography and was sprayed with Dragendorffs reagent (15). To test for the presence of phosphonolipids the phospholipid (methanol) fraction from silicic acid chromatography was separated by thin-layer chromatography in a solvent system of chloroform-acetic acid (1:1) with visualization by cupric acetate (1, 11, 25).

ANTIMICROB. AGENTS CHEMOTHER.

FURLONG ET AL.

2536

CE

PC U

CH

20 .

75-3 -0

> 15

C

0 0.

@i

@3

50

C

0

i

10

0

25-

a 0 U

.-W

C1

5

10 2 0

-1 40

20

10

0

.-Pmwol'A-Me 60

(minute.) FIG. 1. HPLC separation of P. carinii total lipids. Rat-derived P. carinii was extracted with chloroform-methanol (2:1) and separated by HPLC on a silicic acid with a ternary-gradient solvent system with light-scattering detection. The major phospholipids were phosphatidylcholine (PC) and phosphatidylethanolamine (PE). Also shown are the glycolipid peaks that eluted between the cholesterol and phosphatidylethanolamine and the sterol peaks, the most prominent of which is cholesterol (CH). Cholesterol esters (CE) were observed eluting near the solvent front. The results shown here are representative of those obtained in three separate preparations.

5

30

20 Time (minutes)

Time

B

75-

IDo 0

250

0

RESULTS

Total lipid content of P. carinii. HPLC analysis showed that the major lipids in P. carinii include phosphatidylcholine, phosphatidylethanolamine (26), and cholesterol (19) (Fig. 1). Lesser amounts of phosphatidylserine, phosphatidylinositol, and sphingomyelin were also observed. In addition, several unknown peaks were found to elute between the cholesterol and phosphatidylethanolamine peaks, where glycolipids would be expected to elute. Examination of the HPLC trace showed several smaller peaks which eluted close to the major cholesterol peak. Because of the potential importance of defining sterols in preparations containing P. carinii which were not present in host preparations from uninfected animals, the sterol fraction was characterized further. HPLC analysis of P. carinii sterol content. A reversed-phase HPLC separation of a saponified neutral lipid fraction from P. carinii is shown in Fig. 2A. A prominent peak with a retention time identical to that of cholesterol was observed. Five additional peaks with retention times characteristic of those of sterols were also detected. Unlabeled peaks close to the solvent front most likely represent free fatty acids resulting from saponification and were not characterized further. By comparison, samples from control rats (which received corticosteroids but were not infected) prepared in an identical manner showed only a single sterol peak identified as cholesterol (Fig. 2B). Identical results were obtained from animals receiving dexamethasone with trimethoprim-sulfamethoxazole prophylaxis. By integration of peak areas from the unknown peaks from the P. carinii-containing samples, these peaks constituted approximately 10 to 20% of the total sterols. Some variation between preparations was evident. The retention times of the unknown sterols were very similar to those observed for plant sterols. As shown in Fig. 2C, when a mixture of plant sterols containing ,-sitosterol, stigmasterol, campesterol, and brassicasterol was chromatographed, the retention times were similar to the retention times for the sterols found in the P. carinii preparations. Stigmasterol and campesterol did

10

20 Time (minutes)

30

20-

~15@3~~~~~~~~ 0.

FIG. 2. HPLC separation of saponified neutral lipids from P. carinii. The saponified neutral lipid fractions from either a P. carinii preparation (A) or a control preparation (B) from uninfected lungs were treated identically. For comparison, a mixture of plant sterols was analyzed under identical conditions (C). Numbered peaks in panel A correspond to those in Table 2. The arrow in panel B indicates the cholesterol peak. The numbered peaks in panel C correspond to brassicasterol (peak 1), campesterol (peak 2), and ,8-sitosterol (peak 3). A small amount of stigmnasterol was also present in this mixture, but it did not separate from campesterol under these conditions. Separations were carried on a C-18 HPLC column, and peaks were detected by UV at 214 nm. The results shown here are representative of those obtained in eight separate preparations (infected lung preparations) or three preparations (control preparations).

not separate under these conditions. As shown in Table 1, the aL values calculated from standard sterols were very close to

published values (28). The a values calculated for the unknowns suggest that ergosterol, a common 'sterol in many fungi, is not present in detectable amounts in P. cariinii.

STEROL CONTENT OF P. CARINII

VOL. 38, 1994

2537

TABLE 1. Comparison of HPLC retention times of sterol standards with those prepared from a saponified neutral lipid fraction from P. cariniia Sterol standard

Desmosterol Ergosterol Brassicasterol Cholesterol Stigmasterol Campesterol

,B-Sitosterol

% of total sterol Measured a Expected a Peak no. fraction

steon

1.81 1.70 1.42 1.33 1.16 1.14 1.0

1.68 1.65 1.36 1.28 1.16 1.14 1.0

3.1 2.9 79.9 8.6 4.3 1.2

1 2 3 4 5 6

Measured ot for each unknown peak

1.73 1.59 1.32 1.19 1.05 0.84

a For each sterol standard, a was calculated as described previously (28) (measured a column) and was compared with published values (Expected a column). For each of the unknown peaks, a was calculated similarly. The HPLC retention time is dependent on sterol structure; thus, the P. carinii peaks have a degree of structural similarity to those of the standards. hiVhNumbered peaks correspond to those shown in Fig. 2A.

Comparison of sterol content of P. carinii with that of P. a phylogenetically related organism. Although ergosterol is the predominant sterol in many fungi, sterols other than ergosterol have been reported (36). A recent study suggested that there is a high degree of sequence homology between P. carinii and P. polycephalum on the basis of their 5S RNA sequences (35). Other studies have reported the presence in P. polycephalum of sterols other than ergosterol (4). Recent work has shown that phylogenetic relationships between organisms can also be established by comparing their sterol contents (23). Therefore, it was of interest to compare the sterol content of P. polycephalum with that of P. carinii. As shown in Fig. 3, an HPLC trace of a saponified neutral lipid fraction from P. polycephalum showed multiple sterol peaks, the most prominent of which had a retention time and an at value nearly identical to those of stigmasterol. Other peaks had retention times similar to those of campesterol and ,-sitosterol. There was no evidence for the presence of cholesterol in these preparations. GC and GC-MS analysis of P. carinii sterols. To further identify the sterols contained in P. carinii, acetate-derivatized

polycephalum,

L

0~~~~~~~~~~

~~~~~~~3 y

0

20

10

By

30

Time (minutes) FIG. 4. GC separation of saponified and acetate-derivatized P. carinii neutral lipids. Samples were separated on an Ultra-1 column (25 m by 0.22 mm by 0.33 ,um) at 280°C with flame ionization detection. The results shown here are representative of those obtained in three separate preparations. The prominent peak at 14.2 min was identified as cholesterol on the basis of a retention time identical to that of a cholesterol standard. Relative retention of numbered peaks to retention of cholesterol (RRTcholesterol) were as follows: 2, 1.10; 3, 1.21; 4, 1.33; 5, 1.44; 6, 1.47; 7, 1.58; 8, 1.65; 9, 2.04; 10, 2.11; 11, 2.26.

samples were analyzed by GC. On a nonpolar capillary column with flame ionization detection, multiple sterol peaks were observed, similar to the results observed by HPLC (Fig. 4). The predominant peak again comigrated with cholesterol. However, a greater number of peaks were apparent by GC than by HPLC. Like the HPLC results, the relative retention times of the major peaks other than cholesterol corresponded closely to the relative retention times of the phytosterols campesterol, stigmasterol, and 3-sitosterol. Acetate derivatives from the P. carinii fraction were also analyzed by GC-MS. Peaks that eluted prior to 15 min included fatty acids and fatty acid derivatives and were not analyzed further. The total ion plot of the region from 15 to 20 min showed eight major peaks (Fig. 5). The identity of the peak at 16.6 min was confirmed as cholesterol. In relative terms, the quantities of the other unknown lipids were very similar to those observed by HPLC. The major features for each of the unknown sterols are summarized in Table 2. Three of these sterols exhibited spectra identical to those of commercially prepared plant sterols, namely, campesterol (24a-meth-

yl-cholesta-5-en-3p-ol), stigmasterol(24a-ethylcholesta-5,22-

E

A

;l 20 1 4o

-

dien-3,-ol), and ,B-sitosterol(24ot-ethylcholesta-5-en-3I3-ol). Five additional unknown lipids were shown to exhibit spectra indistinguishable from previously published spectra of phytosterols (6) (Table 2). These data confirm the presence of cholesterol in the P. carinii samples but, more importantly, demonstrate conclusively that these lipids are phytosterols. DISCUSSION

0

10

20

rime (minutes) FIG. 3. HPLC separation of saponified neutral lipids from P. polycephalum. Separations were carried on a C-18 HPLC column, and peaks were detected by UV at 214 nm. The results shown here are representative of those obtained in three separate preparations. The calculated a value of the predominant sterol (arrow) was identical to that of stigmasterol.

In the study described here we examined the lipid content of rat-derived P. carinii. These data demonstrate the presence of lipids not previously identified in this organism. Specifically, HPLC and GC retention times and mass spectral data reveal the presence of sterols not present in mammalian lungs. Mass spectral data confirm the presence of cholesterol in the lungderived P. caninii preparations. Comparison of the HPLC and GC retention times with those of sterol standards and values in

ANTIMICROB. AGENTS CHEMOTHER.

FURLONG ET AL.

2538

undace

1900000

i

1800000 1700000 1600000 1500000 1400000 1300000 1200000 1100000 1000000 900000 800000

700000 600000 500000 400000

300000 200000

100000j

_imne--w

16.00

16.50

I

17.00

T

I

17.50

I

18.00

18.50

FIG. 5. GC separation of saponified P. cannii neutral lipids. Total ion plot from lipids separated on an HP-5 column (30 m by 0.25 mm by 0.25 ,um) with temperature programming (75°C for 2 min and then 20°C per minute to 310°C). The mass spectral information for each of the numbered peaks is presented in Table 2. Unknown sterols were identified by comparison with standards and published spectra (6). Unnumbered peaks have not yet been identified.

the literature and careful analysis of the mass spectra show that other P. carinii neutral lipids have many features characteristic of phytosterols. The sterol results provided here are particularly interesting when taken in the context of other recent information showing

the phylogenetic relationship of P. carinii to the protista fungi. Analysis of 5S RNA shows that P. carinii has a high degree of sequence homology to A. castellanii and P. polycephalum (35). Unlike the results shown here for P. carinii, these organisms contain no cholesterol. A. castellanii contains approximately

TABLE 2. MS information for unknown P. carinii sterols Peak no.a

Retention time (min)

1 2 3 4 5 6 7 8

16.61

16.82 17.27 17.46 17.70 18.01 18.49 18.66

Major ions for corresponding peak (m/z)

368, 353, 260, 255, 247, 213, 380, 365, 337, 282, 255, 213, 382, 367, 274, 261, 255, 213, 394, 379, 351, 282, 255, 228, 442, 427, 382, 367, 315, 255, 396, 381, 288, 275, 255, 213, 456, 441, 396, 381, 315, 255, 454, 356, 313, 298, 281, 255,

159, 169 147, 213, 81 147, 229, 253,

147, 145 81 81, 55

81 213, 81, 55 55

Identification

Cholesterol 24-Methylcholesta-5,22-dienol 24-Methylcholesterol 24-Ethylcholesta-5,22-dienol 24-Methylcholesta-7-enol 24-Ethylcholesterol 24-Ethyllathosterol 24-z-Ethylidenelathosterol

a The peaks correspond to those shown in Fig. 5. Identifications were based on similarities with those of standards and previously published spectra (6). All of these peaks show features typical of steroids. Furthermore, with the exception of cholesterol, the mass spectra confirm the similarities of these sterols to those commonly associated with plants and fungi. b The identifications are consistent with data from GC and HPLC.

VOL. 38, 1994

60% ergosterol and 40% dihydroporiferasterol (32). Neither P. polycephalum nor Physarum flavicomum contains cholesterol or ergosterol; they contain rather, 39 and 57% poriferasterol [243-ethylcholesta-5,22(trans)-dienol], 45 and 33% dihydroporiferasterol (2413-ethylcholesta-5-en-3p-ol), and 15 and 10% 24-methylcholesta-5-en-3p-ol, respectively. Other minor sterols present are likely A5-ergostenol, ergostanol, and poriferastanol (4). At present we cannot determine the chirality of the C-24 groups on the P. carinii sterols. In P. polycephalum and P. flavicomum it was necessary to measure the melting points of the purified sterols to distinguish the a from the ,B configuration; neither chromatography nor MS could be used to determine chirality. It would be difficult to purify enough of the P. cannii sterols to carry out such a determination. However, on the basis of the phylogenetic relationship described above, it is likely that the configuration is i in P. carinii also. Recent studies have clearly demonstrated the relationship between sterol structure and phylogeny (23). Although our chromatography data and the results of MS analysis strongly support the hypothesis that P. carinii contains phytosterols in addition to cholesterol, some of the cholesterol measured may reflect contaminating host material. In control lung preparations which did not contain P. carinii, cholesterol, but no other sterol, was observed. As a result, in the present study as well as in previous work (19) the estimates of the cholesterol content of P. carinii are likely to be exaggerated. However, it is also possible that a significant amount of the observed cholesterol is incorporated into the organism's membranes. In a recent study in which attempts were made to minimize host lipid contamination of P. carinii preparations (21), the cholesterol observed as a percentage of total sterols was only slightly less than that described in the present work. In other organisms the cholesterol supplied exogenously is incorporated into cellular membranes (17). It is unlikely that the sterols measured in our preparations were due to cholesterol metabolism in the inflamed lung. As described above, all of the sterols measured in these preparations were phytosterols which could not arise by any known mammalian synthetic pathways. Similarly, it is unlikely that these sterols could have arisen from coinfecting organisms because culture of these preparations confirmed the absence of other infecting organisms. Furthermore, the reproducibility of the sterol contents of these preparations and the absence of ergosterol expected for common fungal pathogens demonstrate that these lipids arise from the infecting P. carinii. The present study cannot address questions about the need for cholesterol derived from the host environment to satisfy "bulk" sterol requirements (e.g., for membrane biogenesis) in P. cainni. Studies with protozoa and plant cells have shown that cholesterol alone is not sufficient to sustain these organisms; a 24-ethyl sterol such as stigmasterol is essential for growth and viability (17). A similar requirement may be expected for P. carinii. Many lower fungi are incapable of de novo sterol synthesis (36). If P. carinii is incapable of sterol synthesis, the concentration of sterols in the lung microenvironment could conceivably affect the growth of the organism in at least two ways: (i) cholesterol could fulfill bulk sterol requirements and the concentrations of cholesterol in the lung microenvironment may regulate the growth of the organism, or (ii) the sterols found in coinfecting fungi or other microorganisms could have a regulatory effect on the growth of the organism (36). It is clear that the regulation of the sterol content of P. carinii is important for understanding the growth of the organism. It may be relevant in this regard that other studies from this laboratory have shown significant increases in

STEROL CONTENT OF P. CARlNII

2539

the amount of cholesterol of bronchoalveolar lavage from human immunodeficiency virus-positive patients (29). Results from the present study are in general agreement with those from previous studies from other laboratories (19, 21, 26). Our analyses showed that the major lipids in P. carinii include phosphatidylcholine, phosphatidylethanolamine (26), and cholesterol (19). We found no evidence for the presence of either phosphonolipids or DGTS (data not shown). Cholesterol was the major sterol present in all of our preparations examined as described previously (21), and the percentage of cholesterol observed (approximately 80%) agrees in general terms with that reported previously (21). Thus, while this is the first report that identifies sterols other than cholesterol in P. carinii preparations, other studies have reported that cholesterol does not constitute the entire sterol content of the organism. The high concentration of cholesterol and the lack of ergosterol in these preparations provide an explanation for why P. carinii is not susceptible to polyene antibiotics such as amphotericin B. In summary, the identification of phytosterols in P. carinii is significant because it provides further evidence for classifying the organism as a fungus and could be used diagnostically to indicate the presence of the organism. Furthermore, the plant-like character of the sterols suggests possible new targets for drug development on the basis of differences in sterol biosynthetic pathways between plants and fungi (5). Finally, given the difficulties associated with growing this organism in vitro, this information may provide insight into nutritional supplements that could enhance growth in culture. ACKNOWLEDGMENTS We gratefully acknowledge the technical assistance of Kathleen Thibault and Laetitia Morbelli. We also thank Karen Klomparens for

the kind gift of P. polycephalum cultures and Thomas Trainor for performing the GC-MS. This work was supported by NIH grant HL34510.

ADDENDUM

Since the manuscript was submitted for publication another report of a study (20) that used a somewhat different methodology from that used in the present study provided similar but not identical results to those described in the present report. REFERENCES 1. Andrews, D., and D. L. Nelson. 1979. Biochemical studies of the excitable membrane of Paramecium tetraurelia. II. Phospholipids of ciliary and other membranes. Biochim. Biophys. Acta 550:174187. 2. Bartlett, M. S., J. A. Fishman, M. M. Durkin, S. F. Queener, and J. W. Smith. 1990. Pneumocystis carinii: improved models to study efficacy of drugs for treatment or prophylaxis of Pneumocystis pneumonia in rats. Exp. Parasitol. 70:100-106. 3. Bartlett, M. S., J. A. Fishman, S. F. Queener, M. M. Durkin, M. A. Jat, and J. W. Smith. 1988. New rat model of Pneumocystis carinii infection. J. Clin. Microbiol. 26:1100-1102. 4. Bullock, E., and C. J. Dawson. 1976. Sterol content of the myxomycetes Physarum polycephalum and P. flavicomum. J. Lipid Res. 17:565-571. 5. Burden, R. S., D. T. Cooke, and G. A. Carter. 1989. Inhibitors of sterol biosynthesis and growth in plants and fungi. Phytochemistry 28:1791-1804. 6. Chitwood, D. R, and W. R. Lusby. 1991. Metabolism of plant sterols by nematodes. Lipids 26:619-627. 7. De Stefano, D., J. A., M. T. Cushion, R. G. Sleight, and P. D. Walzer. 1990. Analysis of Pneumocystis carinii cyst wall. I. Evidence for an outer surface membrane. J. Protozool. 37:428-435. 8. Durkin, M. M., M. M. Shaw, M. S. Bartlett, and J. W. Smith. 1991. Culture and filtration methods for obtaining Pneumocystis carinii

2540

FURLONG ET AL.

trophozoites and cysts. J. Protozool. 38:210S-212S. 9. Edman, J. C., J. A. Kovacs, H. Masur, D. Santi, H. J. Elwood, and M. L. Sogin. 1988. Ribosomal RNA sequence shows Pneumocystis cannii to be a member of the fungi. Nature (London) 334:519-522. 10. Edman, U., J. C. Edman, B. Lundgren, and D. V. Santi. 1989. Isolation and expression of Pneumocystis carinii thymidylate sytnase gene. Proc. Natl. Acad. Sci. USA 86:6503-6507. 11. Ferguson, M. A. J., A. K. Allen, and D. Snary. 1982. The detection of phosphonolipids in the protozoan Trypanosoma cruzi. Biochem. J. 207:171-174. 12. Folch, J., M. Lees, and G. H. Sloan-Stanley. 1957. A simple method for isolation and purification of total lipids from animal tissues. J. Biol. Chem. 226:497-509. 13. Frenkel, J. K., M. S. Bartlett, and J. W. Smith. 1990. RNA homology and the reclassification of Pneumocystis. Diagn. Microbiol. Infect. Dis. 13:1-2. 14. Furlong, S. T., and J. P. Caulfield. 1988. Schistosoma mansoni: sterol and phospholipid composition of cercariae, schistosomula, and adults. Exp. Parasitol. 65:222-231. 15. Furlong, S. T., J. A. Leary, C. E. Costello, and E. A. Dawidowicz. 1986. Isolation and identification of 1(3),2-diacylglyceryl-(3)-O4'-(N,N,N-trimethyl-homoserine from the soil amoeba, Acanthamoeba castellanii. J. Lipid Res. 27:1182-1189. 16. Furlong, S. T., A. Mednis, and H. G. Remold. 1992. Interferon -y stimulates lipid metabolism in human monocytes. Cell. Immunol. 143:108-117. 17. Goad, L. J. 1990. Application of sterol synthesis inhibitors to investigate the sterol requirements of protozoa and plants. Biochem. Soc. Trans. 18:63-65. 18. Golan, D. E., S. T. Furlong, C. S. Brown, and J. P. Caulfield. 1988. Monopalmitoylphosphatidylcholine incorporation into human erythrocyte ghost membranes causes protein and lipid immobilization and cholesterol depletion. Biochemistry 27:2661-2667. 19. Kaneshiro, E. S., M. T. Cushion, P. D. Walzer, and K. Jayasimhulu. 1989. Analyses of Pneumocystis fatty acids. J. Protozool. 36: 69S-72S. 20. Kaneshiro, E. S., J. E. Ellis, K. Jayasimhulu, and D. H. Beach. 1994. Evidence for the presence of "metabolic sterols" in Pneumocystis carinii: identification and initial characterization of Pneumocystis carinii sterols. J. Eukaryotic Microbiol. 41:78-85. 21. Kaneshiro, E. S., M. A. Wyder, L. H. Zhou, J. E. Ellis, D. R. Voelker, and S. G. Langreth. 1993. Characterization of Pneumocystis cannii preparations developed for lipid analysis. J. Eukaryotic Microbiol. 40:805-815. 22. Mirovsky, P., and J. A. Fishman. 1993. An improved method for the prolonged maintenance of Pneumocystis carinii in vitro. J. Infect. Dis. 167:1470-1473. 23. Nes, W. D., R. A. Norton, F. G. Crumley, S. J. Madigan, and E. R.

ANTIMICROB. AGENTS CHEMOTHER.

24. 25. 26.

27. 28.

29.

30. 31.

32. 33. 34.

35.

36. 37. 38.

Katz. 1990. Sterol phylogenesis and algal evolution. Proc. Natl. Acad. Sci. USA 87:7565-7569. Nes, W. R. 1985. A comparison of methods for identification of sterols. Methods Enzymol. 111:3-37. Nozawa, Y., and G. A. Thompson. 1971. Studies of membrane formation in Tetrahymena pyriformis. J. Cell Biol. 49:712-721. Pesanti, E. L. 1987. Phospholipid profile of Pneumocystis carinii and its interaction with alveolar type II epithelial cells. Infect. Immun. 55:736-741. Pixley, F. J., A. E. Wakefield, S. Banerji, and J. M. Hopkin. 1991. Mitochondrial gene sequences show fungal homology for Pneumocystis carinii. Mol. Microbiol. 5:1347-1351. Rodriguez, R. J., and L. W. Parks. 1985. High-performance liquid chromatography of sterols: yeast sterols. Methods Enzymol. 111: 37-50. Rose, R. M., P. J. Catalano, H. Koziel, and S. T. Furlong. 1994. Abnormal lipid composition of bronchoalveolar lavage fluid obtained from individuals with AIDS related lung disease. Am. J. Respir. Crit. Care Med. 149:332-338. Smulian, A. G., and P. D. Walzer. 1992. The biology of Pneumocystis carinii. Crit. Rev. Microbiol. 18:191-216. Stringer, S. L., J. R. Stringer, M. A. Blase, P. D. Walzer, and M. T. Cushion. 1989. Pneumocystis carinii: sequence from ribosomal RNA implies a close relationship with fungi. Exp. Parasitol. 68: 450-461. Ulsamer, A. G., F. R. Smith, and E. D. Korn. 1969. Lipids of Acanthamoeba castellanii: composition and effects of phagocytosis on incorporation of radioactive precursors. J. Cell Biol. 43:108. Vanden Bossche, H., J. Gorrens, and M. C. Coene. 1990. Biochemical basis for the activity and selectivity of oral antifungal drugs. Br. J. Clin. Pract. 71:41-46. Wakefield, A. E., S. E. Peters, S. Banerji, P. D. Bridge, G. S. Hall, D. L. Hawksworth, L. A. Guiver, A. G. Allen, and J. M. Hopkin. 1992. Pneumocystis carinii shows DNA homology with the ustomycetous red yeast fungi. Mol. Microbiol. 6:1903-1911. Watanabe, J., H. Hori, K. Tanabe, and Y. Nakamura. 1989. Phylogenetic association of Pneumocystis carinii with the 'Rhizopoda/Myxomycota/Zygomycota group' indicated by comparison of SS ribosomal RNA sequences. Mol. Biochem. Parasitol. 32:163168. Weete, J. D. 1989. Structure and function of sterols in fungi. Adv. Lipid Res. 23:115-167. Yoshikawa, H., H. Morioka, and Y. Yoshida. 1987. Freeze-fracture localization of filipin-sterol complexes in plasma- and cyto-membranes of Pneumocystis carinii. J. Protozool. 34:131-137. Yoshikawa, H., and Y. Yoshida. 1988. Distribution of digitoninsterol complexes in Pneumocystis carinii revealed by freeze-fracture method. Int. J. Parasitol. 18:39-45.