ABSTRACT
Mucormycosis is a rare and opportunistic infection caused by fungi belonging to the order Mucorales. Recent reports have demonstrated an increasing incidence of mucormycosis, which is frequently lethal, especially in patients suffering from severe underlying conditions such as immunodeficiency. In addition, even though conventional mycology and histopathology assays allow for the identification of Mucorales, they often fail in offering a species-specific diagnosis. Due to the lack of other laboratory tests, a precise identification of these molds is thus notoriously difficult. In this study we aimed to develop a molecular biology tool to identify the main Mucorales involved in human pathology. A PCR strategy selectively amplifies genomic DNA from molds belonging to the genera Absidia, Mucor, Rhizopus, and Rhizomucor, excluding human DNA and DNA from other filamentous fungi and yeasts. A subsequent digestion step identified the Mucorales at genus and species level. This technique was validated using both fungal cultures and retrospective analyses of clinical samples. By enabling a rapid and precise identification of Mucorales strains in infected patients, this PCR-restriction fragment length polymorphism-based method should help clinicians to decide on the appropriate treatment, consequently decreasing the mortality of mucormycosis.
Mucorales are fungi that belong to the Zygomycetes class: they are characterized by the production of coenocytic and ribbon-like hyphae and by their sexual reproduction through zygospores. These molds are widespread in nature, as they subsist on soil, decaying vegetation, fruits, and seeds and are air dispersed (23). Until now these organisms had been rarely seen in clinical practice and were paid very little attention. However, recent reports have demonstrated an increasing incidence of these opportunistic molds causing rapidly evolving infections in severely immunocompromised patients. These infections, called mucormycosis, generally lead to the death of the patients (10, 17, 18, 25). Consequently, Mucorales are now listed by university hospitals as microorganisms responsible for frequent emerging infections (16, 20).
Members of the genera Rhizopus, Mucor, Absidia, and Rhizomucor are the Mucorales most commonly isolated from patients. Cunninghamella, Apophysomyces, and Saksenaea have occasionally been implicated in human diseases, sometimes even in immunocompetent patients (1, 9, 26). When it is not detected early and aggressively treated, with high doses of intravenous amphotericin B and surgical debridement, mucormycosis is frequently fatal; mortality rates may be as high as 80% in infected transplant recipients (5, 22). Indeed, amphotericin B is currently the only effective therapy for Mucorales, but its use is limited by severe nephrotoxic side effects. In a recent in vitro study, differences between genera and species of Mucorales in susceptibility to conventional and new antifungals were observed (3). However, Mucorales are phylogenetically heterogeneous with variable antifungal susceptibilities, and an appropriate medical treatment thus requires a specific identification of the pathogenic agent. Culture is still the predominant method to identify the molds responsible for mucormycosis. Yet, many published studies do not report any species identification. Additionally, a recent paper observed a 21% discrepancy in determination of genera between morphological and sequence-based methods (13).
Innovative diagnostic tools are required for therapeutic strategies specifically targeted at these emerging infections. Molecular techniques show enormous potential for rapidly and accurately identifying the etiological agents of mucormycosis (28). However, molecular detection assays for these fungi are still not widely available. The Microseq D2 sequencing kit has been developed to amplify the D2 domain of the large rRNA gene subunit and therefore to enable the sequencing of clinically important filamentous fungi (8). However, a recent report highlights the discordance between conventional phenotypic characterization of Mucorales and their identification using this kit (6). Several other studies focused on DNA amplification either for broad analyses or for the characterization of a particular isolate (2, 19, 29). Direct DNA sequencing of the PCR products obtained from panfungal primers remains the most reliable way to precisely identify a mucoralean species (12). In the present study we developed a molecular technique aiming at rapid and precise diagnosis of the main members of the Mucorales order encountered in human pathology. This epidemiologic tool can be used in parallel with the conventional mycology techniques in order to confirm the genera and species identified in clinical samples. The reliability of the technique was first controlled through bioinformatics and then assessed on fungal cultures. Finally, a PCR-restriction fragment length polymorphism (RFLP) method was then retrospectively evaluated with two recent clinical cases that occurred in the hospital at Nancy, France.
MATERIALS AND METHODS
Fungal strains.PCR and RFLP techniques were performed using 17 mucoralean strains maintained at the Centraalbureau voor Schimmelcultures, Utrecht, The Netherlands (CBS); the Pasteur Institute, Paris, France (UMIP); and the Scientific Institute of Public Health, Brussels, Belgium (IHEM) (reference numbers are given in parentheses): Absidia corymbifera (CBS 429.51, CBS 118994, IHEM 3809), Rhizopus microsporus (CBS 112285), Rhizopus microsporus var. oligosporus (CBS 338.62), Rhizopus microsporus var. rhizopodiformis (CBS 536.80, CBS 118987), Rhizopus microsporus var. chinensis (CBS 261.28), Rhizopus oryzae (CBS 112.07, UMIP 1443.83), Rhizopus azygosporus (CBS 357.93), Mucor circinelloides f. circinelloides (CBS 195.68, IHEM 6794), Mucor hiemalis f. hiemalis (CBS 201.65), Mucor indicus (CBS 11229), Rhizomucor miehei (CBS 182.67), Rhizomucor pusillus (IHEM 4897). The following fungi from the Service de Parasitologie-Mycologie, Nancy, France (CHU Nancy), were used as controls: Acremonium sp. (CBS 118986), Aspergillus fumigatus (CBS 118990), Scedosporium sp. (CBS 118991), Penicillium sp. (CBS 118992), Fusarium sp. (CBS 118995), Alternaria sp. (CBS 118988), Geotrichum sp. (CBS 118989). Strains were grown at 27°C on Sabouraud chloramphenicol agar (SCA).
Clinical samples.Two clinical samples from patients with suspected mucormycosis were collected from the intensive care unit in the Nancy hospital (France) between January and July 2005. Sample 1 originated from a 63-year-old female suffering from an aggressive B lymphoma diagnosed a month before. In spite of receiving antineoplastic chemotherapy against lymphoma the patient developed, 10 days after her admission, a rapidly expanding back lesion evolving towards necrosis. Mycological examination confirmed the clinical suspicion of primary cutaneous mucormycosis as respiratory and blood samples were negative for zygomycetes. Cultures of samples on SCA rapidly grew an aerial mycelium, which was microscopically characterized as Rhizopus sp. Because of the severe immunosuppression, liposomal amphotericin B (Ambisome; 10 mg/kg of body weight) was given intravenously and extensive surgery was undertaken. However, the patient went into multiorgan failure and died after 30 days of intensive care (15).
Sample 2 came from a 22-year-old male presenting with acute lymphoblastic leukemia. During his prolonged immunosuppressed state, this patient developed a lung abscess detected by chest X ray and tomodensitometric abnormalities. Mycological examinations were performed on a first biopsy specimen that led to the diagnosis of pulmonary mucormycosis with the presence of Absidia sp. localized in the right lobe. Surgical debridement was then undertaken, and histopathology and mycology confirmed the previous diagnosis. His status improved with daily intravenous liposomal amphotericin B administration.
A part of each sample was processed using mycological techniques, i.e., direct examination in black chlorazol solution, followed by culture on SCA. The remainder was tested by PCR-RFLP as described below.
Identification of Mucorales by macroscopic and microscopic examinations.Diagnosis of mucormycosis is based on the morphological detection of hyphae. First, direct mycological examination was made in black chlorazol solution. Biological samples were then cultured on SCA, and fungal identification was made on the basis of macroscopic and microscopic morphological features. Microscopic observation of the mycelium was done by the preparation of lactophenol cotton-blue-stained slides.
DNA extraction.Purified DNAs from various Mucorales were kindly provided by the CBS. DNA from other cultures or from the clinical samples was extracted using the High Pure PCR template preparation kit (Roche Diagnostics, Meylan, France). The standard protocol was slightly modified by a short pretreatment of the sample. Briefly, an aliquot of mycelium was suspended in 200 μl of tissue lysis buffer and incubated for 30 min at 37°C in the presence of lysozyme (10 U/μl, final concentration). DNA was then extracted by following the instructions of the manufacturer. Finally, the concentrations were determined with an Ultrospec 2100 spectrophotometer (Amersham). Samples were kept at −20°C until used.
Selection of primers and restriction enzymes.The design of specific primers and the selection of specific restriction enzymes were performed according to the following sequential steps.
Selection of targeted 18S fungal sequences.A set of 162 fungal small subunit sequences corresponding to 69 species (36 Mucorales, 22 other filamentous fungi, and 11 yeasts) were recovered from GenBank. When the available sequences were polymorphic or different for a given fungus, a consensus sequence was selected. When identical sequences were available for a given fungal genus or species, a unique representative sequence was selected (Table 1).
GenBank accession numbers of the 18S sequences selected for the alignments
Sequence alignment and selection of specific primers.The selected sequences were aligned using software available at the website http://www.genebee.msu.su/genebee.html. Short oligonucleotide sequences were subsequently designed for the amplification of the main Mucorales recovered from human infections. The primer set was composed of a mix of specific sense primers corresponding to the sequences of Rhizopus sp., Rhizomucor sp., Mucor sp., and Absidia corymbifera (RpL1, 5′ TGATCTACGTGACAAATTCT 3′; RmL1, 5′ TGATCTACGCGAGCGAACAA 3′; MucL1, 5′ TGATCTACGTGACATATTCT 3′; and AbsL1, 5′ TGATCTACACGGCATCAAAT 3′, respectively) and a degenerate antisense primer (MR1, 5′ AGTAGTTTGTCTTCGGKCAA 3′). Sense and antisense primers annealed to regions of the template starting at positions 75 and 901, respectively, according to a reference sequence of Absidia corymbifera AF113407. The region selected for the design of primers excluded the amplification of human DNA and other filamentous fungi. The expected band size was 830 bp on average.
As a positive control, a set of primers was designed to amplify all the fungi and the human DNA: Lap (5′ GAAACTGCGAATGGCTCATTA 3′) and Rap (5′ CAATCCAAGAATTTCACCTCT 3′) corresponding to nucleotides 46 to 881 relative to the same reference sequence, Absidia corymbifera AF113407. The expected amplicon size was 840 bp.
Identification of endonuclease sites within the amplified fragment of Mucorales.By using the sequence alignment obtained as described above and information available at the website http://www.infobiogen.fr/services/analyseq/cgi-bin/carteres_in.pl, we selected eight restriction enzymes to specifically identify the main Mucorales genera and species encountered in human pathology (Table 2).
Restriction enzymes used in the study
PCR.PCR was performed in a reaction mixture of 50 μl containing 2 mM MgCl2, 200 μM of each deoxynucleoside triphosphate, 20 pmol of each primer, 2 U of Fast Start Taq polymerase (Roche, Meylan, France), and 5 ng of the previously extracted DNA.
The PCR conditions consisted of denaturation for 3 min at 94°C, followed by 30 amplification cycles at 94°C for 1 min, 60°C for 1 min, and 72°C for 1 min and one final extension cycle at 72°C for 5 min. Following amplification, 5 μl of amplicons was then electrophoresed in 2% agarose gels in the presence of ethidium bromide and visualized under UV light. The PCR sensitivity was evaluated on serial dilutions of purified DNA of Mucor sp., Rhizopus sp., Rhizomucor sp., and Absidia corymbifera. Tenfold dilutions from 100 ng to 0.001 ng were prepared in distilled water and used as templates in the PCR experiments. Each test was repeated five times in five different runs.
RFLP.The restriction enzymes were used to digest the 18S amplified fragments of the Mucorales. The amplicons were digested for 1 h at 37°C in a total volume of 20 μl using 10 μl of the specific PCR product and 5 U of each of the selected restriction enzymes. Table 2 summarizes the specificity, restriction sites, and expected pattern for all selected enzymes. The digested samples were analyzed on 1.5% BET-agarose electrophoresis gels (QBiogen, France).
Nucleotide sequence accession numbers.The partial 18S sequences obtained for Rhizopus microsporus var. rhizopodiformis and Absidia corymbifera isolated from both clinical cases appear in GenBank under the accession numbers DQ013302 and DQ340176, and DQ340177, respectively.
RESULTS
Mycological diagnosis of clinical specimens.The key feature associated with the presence of zygomycetes on direct mycological examination in black chlorazol solution is the presence of wide, ribbon-like, nonseptate, hyaline hyphal elements. The suspicion of Mucorales was therefore supported by a rapid growth of the cultured samples on SCA, with mycelial elements expanding to cover the entire plate in 2 to 3 days. Case 1 was identified as Rhizopus sp. on the basis of microscopy reporting the presence of stolons, brown pigmented rhizoids, sporangiophores, and globose sporangia, both apophysate and columellate (4, 15).
Case 2 grew a culture characterized as Absidia sp. by the presence of small rhizoids and flask-shaped apophysis with a large columella producing sporangiophores (4).
PCR specificity and sensitivity.A mix of four specific sense primers and a degenerate antisense primer were designed in order to amplify the main opportunistic Mucorales (Fig. 1). The specificity of this mix was controlled by using in parallel the panfungal primer set Lap and Rap that amplified all fungi and yeasts potentially recovered from clinical samples as contaminants or pathogens and human DNA with an expected size of 840 bp (Fig. 2).
Alignment of partial 18S rRNA gene sequences for Mucorales, other filamentous fungi, yeasts, and human. Crosses and dots indicate sequence homologies or polymorphisms in this domain, respectively. The binding regions of the sense and antisense primers are boxed. The sequence polymorphisms between the four main pathogenic Mucorales (in boldface) or the human sequence are highlighted in gray. Sequences were obtained from GenBank and represent parts of the sequences given under accession numbers listed in Table 1.
Electrophoresis patterns of amplicons obtained from fungal and human DNA with the panfungal primers Lap and Rap. Lanes M, 100-bp ladder; lanes T−, negative control (sterile water); lane 1, Mucor sp.; lane 2, Acremonium sp.; lane 3, Aspergillus fumigatus; lane 4, Candida albicans; lane 5, Fusarium sp.; lane 6, Geotrichum sp.; lane 7, Alternaria sp.; lane 8, Scopulariopsis brevicaulis; lane 9, Scedosporium sp.; lane 10, Penicillium sp.; lane 11, DNA from Absidia corymbifera; lane 12, human DNA.
PCR sensitivity was tested in five different experiments on serial dilutions of DNA amplified from Mucor sp., Rhizopus sp., Rhizomucor sp., and Absidia corymbifera. The detection limit was 100 pg/μl of purified DNA per reaction.
RFLP and molecular identification of Mucorales.Restriction enzymes were validated for genus identification of the four main opportunistic Mucorales: Absidia corymbifera (AclI), Rhizopus sp. (BmgBI), Rhizomucor sp. (PpuMI), and Mucor sp. (AflII) (Fig. 3; Table 1). In order to further confirm our choices for species discrimination, the enzymes XhoII, CspCI, and AseI were shown by testing to be specific for Rhizomucor pusillus, Rhizopus oryzae, and the group Rhizopus microsporus and Rhizopus azygosporus, respectively. XmnI is specific for the group comprising Mucor circinelloides, Mucor racemosus, Mucor ramosissimus, and Mucor plumbeus but does not cut amplicons obtained from Mucor hiemalis or Mucor indicus.
Amplification and restriction patterns obtained from PCR-RFLP on fungal cultures (A) and on clinical samples (B and C). Lanes M, 100-bp ladder; lane 1, nondigested PCR product (Mucor sp.); lane 2, digestion with AclI (Absidia corymbifera); lane 3, digestion with AflII (Mucor sp.); lane 4, digestion with BmgBI (Rhizopus sp.); lane 5, digestion with PpuMI (Rhizomucor sp.); lanes 6 to 8, PCR-RFLP on a cutaneous biopsy specimen of a back lesion. The first 100-bp band of the ladder is not distinguishable on this photograph. Lanes 9 and 10, PCR-RFLP on a biopsy specimen from pulmonary abscess; lanes 6 and 9, nondigested amplicons; lane 7, digestion with BmgBI (Rhizopus sp.); lane 8, digestion with AseI (Rhizopus microsporus or Rhizopus azygosporus); lane 10, digestion with AclI (Absidia corymbifera).
The reproducibility of the RFLP was examined, and subsequent experiments yielded similar digestion patterns. In addition, the restriction patterns agree with bioinformatic expectations for each organism.
Identification of clinical isolates by PCR-RFLP.An 830-bp band corresponding to the presence of Mucorales was amplified from clinical isolates 1 and 2. In case 1, digestions were performed by using enzymes PpuMI, AflII, BmgBI, and AclI, specific for Rhizomucor sp., Mucor sp., Rhizopus sp., and Absidia corymbifera, respectively. Two fragments (600 bp and 230 bp) were visualized after BmgBI restriction. The pattern was then consistent with the presence of a Rhizopus sp., digestions with PpuMI, AflII, and AclI being ineffective since the amplicon does not possess any restriction site for these enzymes. Subsequently the specific pattern of Rhizopus microsporus or Rhizopus azygosporus was established using AseI (Fig. 3). On the basis of epidemiological and clinical data, Rhizopus microsporus var. rhizopodiformis was found to be likely implicated in this infection.
In case 2, we obtained a specific feature of Absidia corymbifera with use of the restriction enzyme AclI. The amplicon was not digested by PpuMI, AflII, and BmgBI and revealed the absence of Rhizomucor sp., Mucor sp., and Rhizopus sp. (Fig. 3). This result was consistent with the mycological diagnosis.
DISCUSSION
Mucorales are widespread environmental fungi with minimal intrinsic pathogenicity for healthy individuals. These increasingly emerging opportunistic infections are related to severe trauma or disorders of innate immunity and may rapidly evolve (21, 24). Most antifungals are ineffective against Mucorales, though new and promising compounds such as posaconazole are being developed (2, 27). Until then, an effective strategy for increasing survival rates requires improvement of diagnostic accuracy for these life-threatening infections.
Based on the studies of Dannaoui et al. (3), the precise identification of Mucorales down to species level would also have great importance for further research on antifungal effectiveness. Indeed, these results show, for example, that Rhizopus spp. were significantly less susceptible to itraconazole, posaconazole, terbinafine, and amphotericin B than Absidia spp. were and less susceptible than Mucor spp. were to amphotericin B.
Currently mucormycosis diagnosis and fungal identification are mostly established by culture. However, recent studies assessing the sensitivity of fungal cultures found that they were positive in only 52% of the autopsy cases presumed to have fungal infections and in only 30% of surgical specimens in which fungi were identified on direct examination (6).
In this study we therefore described the development of a molecular technique by direct PCR-RFLP on histopathological samples to rapidly identify the main Mucorales reported to cause infections in humans. The assay targets the 18S ribosomal gene, which is characterized by a relatively low rate of molecular evolution and therefore represents a suitable marker for taxonomic identification (29). This gene has the advantage of having also some particular variations that can be exploited to distinguish between the various fungal genera and species belonging to the same order.
A wide bioinformatic analysis was necessary to design the best strategy to follow. Sequence alignments revealed high degrees of homology between Mucorales and other fungal pathogens or contaminants. Therefore, a technique based on two sequential steps was chosen. A mix of specific (sense) and degenerate (antisense) primers was first selected to amplify the DNA of the most frequently isolated genera of Mucorales in humans: Mucor, Rhizopus, Rhizomucor, and Absidia. In our case, an equimolar mixture of forward primers allowed a better specificity of the PCR and avoided the amplification of other fungal and human DNA. Even though this strategy does not have many precedents in the literature, it seems a good alternative for taking advantage of divergent sequences (7). The evaluation of the PCR on several strains of Mucorales, filamentous fungi, and yeasts led to differentiation between the presence and the absence of Mucorales. In a second stage, restriction enzymes were selected through bioinformatic software enabling specific identification of Mucorales.
PCR-RFLP was first successfully applied to fungal cultures and then retrospectively on two clinical samples. The use of PCR-RFLP on invasive clinical specimens has the advantage of saving time to confirm the presence of Mucorales in cases where direct examination is positive while mycological cultures need 24 or 48 h before growth of characteristic structures; this extra time can be deleterious to the patient if the fungus is indeed present.
In cases where there is clinical suspicion of Mucorales with absence of fungal hyphae on direct examination, molecular detection greatly speeds up the process by rapidly estimating the likelihood of the diagnosis. As a precaution, a second analysis has to be made in order to rule out false positives due to contamination of the sample.
The digestion step enables the species-level identification, which may be important for further antifungal susceptibility testing, as reported in the studies of Dannaoui et al. (3).
In the hospital at Nancy, France, we could confirm the presence of Rhizopus microsporus var. rhizopodiformis in cutaneous mucormycosis and an Absidia corymbifera in a lung abscess. These two fungi grew in immunocompromised patients between January 2005 and July 2005 and were not related to any antifungal prophylaxis. Eight Mucorales had been isolated previously from patients hospitalized in Nancy since January 2003, but none between 1981 and 2002 (unpublished data). Other previous reviews also report an increase in the incidence of mucormycosis with or without any context of antifungal prophylaxis (14). An environmental common source was unlikely to explain this increase, all the reports coming from several geographically divergent hospitals. Thus, the hypothesis of a multifactorial emergence of these fungi has been raised (11).
More studies are now needed to widely explore the accuracy of PCR-RFLP as a diagnostic tool on various biological samples from patients suspected of mucormycosis. One of the main potential difficulties could be the lack of sensitivity of the amplification step, particularly with blood or serum samples. Nevertheless, it was recently shown, in a retrospective analysis, that panfungal amplification on patient serum combined with direct sequencing of the products might be a potential approach in the diagnosis of rare mold infections such as pulmonary mucormycosis, even allowing identification down to the subspecies level (12). In order to improve PCR sensitivity, a potential adaptation of this technique using real-time PCR is currently being developed in our laboratory.
ACKNOWLEDGMENTS
We thank Quentin Vicens (Howard Hughes Medical Institute, Department of Chemistry and Biochemistry, University of Colorado, Boulder) for his valuable help in checking the English.
Financial support came from Pfizer.
None of the authors had a conflict of interest.
FOOTNOTES
- Received 9 September 2005.
- Returned for modification 20 October 2005.
- Accepted 3 January 2006.
- American Society for Microbiology
