Introduction

Systemic fungal infections with high morbidity and mortality rates in immunocompromised patients are growing. Besides the increasing incidence, recent epidemiology of fungal infection shows the expanding variety of fungal pathogens [1].

Identification of the causative pathogen is a fundamental step for appropriate treatment of infectious diseases, and early initiation of antifungal therapy is crucial for reducing the mortality rate in infected patients. Despite efforts by many researchers, however, early and rapid diagnosis of systemic fungal infection remains limited. Conventional diagnostic procedures, such as cultivation of fungi from clinical samples, are time-consuming and suffer from low sensitivity. Furthermore, sufficient technique and experience are required at the identification step. In recent years, other methods, such as PCR and serological tests, have been established for rapid and sensitive detection of fungi from clinical samples [2, 3]. However, these methods are difficult to use to identify a variety of fungal species at a time. Although multiplex PCR can be used to identify several species in one test, its applicability is limited by the primer sets used because specific primer sets are needed for each species.

A variety of DNA array systems have been developed to identify several bacteria and/or fungi simultaneously with high sensitivity and specificity [4–8]. Fluorescent labels are widely used to detect the signal in DNA microarrays due to their high sensitivity. However, the low stability of most fluorescent dyes and the necessity of expensive scanning equipment call for the development of alternative labeling systems that are inexpensive and robust.

To facilitate the diagnosis of fungal infectious disease, we established a rapid and specific DNA microarray system for identifying a variety of causative fungal species simultaneously. We applied the chemical color reaction of biotin-peroxidase and its substrate as the signal detection for the microarray system, enabling examination of the spot pattern with the naked eyes, without the need for expensive scanning equipment. To evaluate the specificity and sensitivity of this visible DNA microarray system, we tested it on several kinds of samples, such as reference fungal strains, blood samples containing a certain number of fungal cells, serum samples with serial dilutions of fungal DNA, and blood culture samples from patients.

Conventional PCR methods have been used for labeling and amplifying DNA from pathogen in microarray identification systems, but this method could not be used for bedside analysis and therefore difficult to be widely adopted. Recently, several isothermal amplification methods that do not require expensive thermal cyclers, such as loop-mediated isothermal amplification (LAMP), helicase-dependent amplification (HAD), and isothermal and chimeric primer-initiated amplification of nucleic acids (ICAN), have been developed to replace PCR amplification [9–12]. However, the maximum amplifiable length of the products in these isothermal methods is too small (100–500 bp) for our purposes, making primer design difficult. Hence, in this study, we applied the recombinase polymerase amplification cycle (RPA) for labeling and amplification of DNA products from the pathogen for microarray detection [13]. The RPA technology is based on a combination of polymerases and DNA recombinases. These enzyme mixtures are active at low temperature (optimum around 37 °C) and recognize template target sites by oligonucleotide primers, followed by strand-displacing DNA synthesis. Thus, exponential DNA amplification of the target region is proceeded under the isothermal condition.

Materials and Methods

Microorganisms and Growth Conditions

A total of 355 strains were obtained from the Medical Mycology Research Center IFM Collection (Chiba university, Japan) (Table S1). All fungal strains were cultivated on PDA medium at appropriate temperatures. Small amount of fungal cells were picked by toothpick and suspended in distilled water to become little bit cloudy solution and used as template for PCR amplification.

Design of Capture Probes

The fungal oligonucleotide probes were designed based on the whole internal transcribed spacer (ITS) sequences regions available in the GenBank database and from our own sequencing data. The alignment was prepared by BioEdit using several objective and nonobjective fungal ITS sequences as listed in Table S2. After sequence alignment, species- or genus-specific oligonucleotide sequences were selected to be unique to each species/genus. To evaluate the specificities against other organisms, we performed additional BLASTN searches of the GenBank database. The designed probes were consisted of 14–21 species/genus-specific oligonucleotides and a poly-T anchor at the end of the oligonucleotides [14]. Detailed sequences of the capture probes are given in Table 1.

Table 1 Oligonucleotide sequence of probes
Full size table

Preparation of DNA Microarray Slides

The synthetic oligonucleotides were diluted to 20 pmol/μl in TE buffer and mixed with an equal volume of 6× SSC [20× SSC is 3 M NaCl, 0.3 M sodium citrate (pH7.0)] to make a final concentration of 10 pmol/μl oligonucleotides in 3× SSC. The probe solutions were spotted on NGK plastic slides (NGK insulators LTD, Aichi, Japan) using a KCS-mini microarray printer (Kubota Comps Corporation, Hyogo, Japan). After spotting, the slides were irradiated with UV at 0.6 J/cm2 using a UV cross-linker (model CL-1000; UVP, San Gabriel, Ca) to fix the probes on plastic slides. The slides were then gently shaken in blocking buffer [3 % BSA, 0.2 M NaCl, 0.1 M Tris–HCl (pH 8.0), and 0.05 % Triton X-100] for 5 min and washed with TE buffer for 10 min. The array slides were air-dried and stably stored at room temperature at least 3 years.

Infectious Mouse Model

As an infection model, male ICR mice (Charles River Laboratories) were infected intravenously with Aspergillus fumigatus Af293 or Fusarium solani complex IFM40718 (FSSC) conidia (1 × 106 conidia/mouse) in a 200 µl volume of saline. Three mice were used in each fungal species. One hour after infection, mice were killed, and blood was collected from the heart tissues under sterilized conditions. The CFU was determined by inoculating 100 µl of collected blood on a PDA with Chloramphenicol plate, and colonies were counted after 24 h of cultivation at 30 °C. Blood samples were used directly as template for PCR.

Blood Culture

As the routine diagnosis of blood infection, blood samples were taken from patients and cultivated using the BD BACTEC FX system (BD, Tokyo, Japan) for 7 days, following the ethics of Chiba University Hospital. After cultivation, growth positive samples were inoculated onto several kinds of agar to identify bacteria and/or fungi. For the microarray identification, one growth positive and one growth negative blood culture samples were used directly as a template for PCR.

DNA Extraction

Fungal DNA was extracted as normal phenol–chloroform method. The conidia of A. fumigatus were inoculated in PDB medium and cultivated for 2 days at 37 °C. The mycelium was collected by filtration and ground by mortar using liquid N2. The ground cells were suspended in DNA extraction buffer (200 mM Tris, 25 mM NaCl, 25 mM EDTA, 0.5 % SDS, pH 8.5) and extracted with phenol/chloroform/isoamyl alcohol. After that, RNase treatment and ethanol precipitation were conducted. The cells of Candida albicans were cultivated in PDB medium 1 day at 30 °C. The cells were collected by centrifugation, and the DNA was extracted using GenTorukun (TaKaRa Bio Inc., Shiga, Japan).

PCR Amplification and Labeling

The 5′-biotin-labeled fungus-specific universal primers ITS1-bio (5′-TCCGTAGGTGAACCTGCGG-3′) and ITS4-bio (5′-TCCTCCGCTTATTGATATGC-3′) [15] were used for amplifying the entire ITS region and biotin labeling. The amplified fragment ranged from 426 to 930 bp depending on the fungal species. PCR was performed using MightyAmp DNA polymerase ver. 2 (TaKaRa Bio Inc.) in a total reaction volume of 10 μl containing 1 μl of template (fungal cell suspension, whole blood, serum, blood culture, and extracted DNAs). Amplification was carried out as follows, 2 min of initial denaturation at 98 °C, 40 cycles of DNA denaturation at 98 °C for 10 s, primer annealing at 55 °C for 15 s and elongation at 68 °C for 45 s, and a final elongation step at 68 °C for 5 min. After the PCR reaction, amplification was verified by electrophoresis. In case of low content of fungal cells or DNA, nested PCR was performed using ITS1-n (5′-GAGGAAGTAAAAGTCG-3′) and ITS4-n (5′-TTCACTCGCCGTTACT-3′) as the first-round PCR primer set. One µl of first-round PCR sample was then used as template in the second round of PCR performed in a total reaction volume of 10 μl.

Isothermal Amplification

Isothermal amplification was performed with a TwistAmp basic kit (TwistDx Limited, Cambridge, UK). Amplification was carried out at 37 °C for 40 min according to the manufacturer’s protocol using 1.5 μl of fungal cell suspension or DNA as template.

Microarray Hybridization and Signal Detection

Before microarray hybridization, amplified and labeled PCR samples were denatured at 95 °C for 2 min and chilled on ice for 2 min. Four μl of a denatured sample was then mixed with 16 μl of hybridization buffer (0.2 g tetramethylammonium chloride, 0.5 % SDS, and 1.9 mg EDTA in 1 ml of 6 × SSC). Samples were applied to the array slide, covered with a cover-film to prevent sample evaporation, and incubated at 37 °C for 1 h in a moist-chamber. Array slides were then washed with PBS buffer at 37 °C for 5 min. A color development reaction was performed on the slide in accordance with the avidin-biotinylated peroxidase complex (ABC) method using a 3, 3′, 5, 5′-tetramethylbenzidine (TMB) solution for visualization [16]. First, the conjugation reaction was performed with streptavidin and biotin-HRP for 30 min, and the array was washed twice with PBST buffer (PBS buffer with 0.1 % Tween 20) for 5 min. After washing, color development was performed with 0.02 % TMB, 0.015 % H2O2, and 0.5 mg/ml alginic acid in 0.2 M acetate buffer (pH 3.3). Color development was terminated after 30 min by washing the array slides with distilled water. The results were evaluated by visual observation.

Results

Design of Probes and DNA Microarray Slides

Because the ribosomal RNA gene, especially the 28S rRNA gene, is highly conserved among species, sequences of the ITS region are widely used for identification of fungal species. We designed species- and/or genus-specific probes within the ITS region and tested the specificity of selected sequences using 355 reference strains (Table S1). Genus-specific probes were designed for some fungi (Alternaria sp., Rhizomucor sp., Mucor sp., Trichosporon sp.); because they have highly conserved ITS region sequences within the genus, we could not design species-specific probes. We successfully designed 319 probes of species/genus-specific oligonucleotides ranging from 13 to 21 bp with a poly-T anchor at the 5′ or 3′ end for the identification of 42 species from 24 genera of fungal pathogens (Table 1). Three to twelve different specific capture probes were designed and spotted on the array slides for each fungal species/genus to ensure hybridization reaction for proper identification. Among the 319 probes, six universal probes for fungi were designed, so that the array would give a positive signal at universal probe even if fungal species in the tested sample was not listed on the Table 1. In other words, 6 universal probes could detect any fungi other than listed objective fungi without specific signal.

All designed probes and the positive control marker (biotinylated-poly-T) were spotted on one plastic slide. Figure 1a shows an example of the spot pattern of the microarray slide.

Fig. 1
figure 1

a Example layout of capture probes on microarray slide. Probe names correspond to probe names listed in Table 1. The black column labeled “biotin” indicates the spots for positive control (biotinylated-poly-T) and positional marker. b Typical hybridization patterns using fungal suspension of different fungal species as PCR template. Species-specific signals are enclosed in solid line frames, while universal signals for fungi are enclosed in dotted line frames. These figures show representative results for A. fumigatus, Trichosporon asahii, C. tropicalis, and C. albicans. c Simultaneous hybridization of several species in one array slide. Fungal cell mixtures of C. albicans, Cryptococcus neoformans, and T. asahii or of A. fumigatus and Trichophyton rubrum were directly used as template for PCR amplification and detected on the array slide

Full size image

Evaluation of the Specificity of DNA Microarray Probes

To evaluate the specificity of the designed capture probes, 66 fungal strains were used (Table S1). All fungal samples tested showed the expected species/genus-specific hybridization patterns as shown in Fig. 1b. Although some probes showed cross-hybridization within the same genus because of their highly conserved sequence (e.g., Rhizopus stlonifer was cross-hybridized to Rhizopus microsporus probes), listed organisms are the major fungi causing infection (Table 1). Moreover, the array system enabled us to identify all the mixed fungi in one test even when several fungal mixtures were used as template (Fig. 1c). Resulted spot number is sometimes varied depending on the sample (e.g., the spot of fungal common probes in Fig. 1b), because the affinity of each designed probes is different.

Sensitivity of the DNA Microarray System

To evaluate microarray detection sensitivity, we used blood samples containing a known number of fungal cells and serum with fungal DNA in place of actual clinical samples. Serial ten-fold dilutions of fungal cells in blood (106–100 CFU/ml) were prepared by adding A. fumigatus conidia or C. albicans cells to rabbit whole blood. The PCR reaction was performed directly using 1 μl of rabbit whole blood with or without fungal cells as template (see Materials and Methods). After the PCR reaction, amplification was verified by electrophoresis, and the samples were used for microarray analysis (Fig. 2a). For both A. fumigatus and C. albicans, 103 CFU/ml was the minimum concentration needed for PCR amplification followed by the microarray detection. Although 102 CFU/ml could be considered as the limit of detection, amplification at this concentration is not reproducible. To increase the sensitivity, we conducted nested PCR. However, it did not enhance the sensitivity.

Fig. 2
figure 2

Agarose gel electrophoresis of PCR products. a PCR amplification of ITS region using blood sample spiked with A. fumigatus cells as template. Upper panel shows the results of normal PCR; lower panel shows the results of nested PCR. Lanes: M, molecular marker (Gene Ladder Wide 1: NIPPON GENE co., Tokyo, Japan); 1–7, blood sample spiked with conidia. b PCR amplification of ITS region using serum sample with C. albicans DNA as template. Upper panel shows the results of normal PCR; lower panel shows the results of nested PCR. Lanes: M, molecular marker; 1–8, DNA in 1 ml of serum

Full size image

We also evaluated detection limits of A. fumigatus and C. albicans DNA in serum. Extracted fungal DNA ranging from 10 ng to 10 fg was separately added to 200 μl of rabbit serum. When 1 μl of the serum sample was used directly as template for PCR, the detection limit was 5 fg per 1 μl of serum. After nested PCR, the minimal amount of DNA required for fungal identification decreased to 0.05 fg per 1 μl of serum (Fig. 2b).

Identification of the Infected Fungi from Mice

Blood samples from infected mice were tested in place of actual human clinical samples. After 1 h of fungal infection, blood was collected and used directly for PCR amplification. Because the first-round PCR did not yield enough amplicon, nested PCR was performed to increase the labeled amplicon, making it possible to detect inoculated fungi in the blood from the infected mouse by microarray. At this moment, the CFU of A. fumigatus and FSSC remained in blood stream were 500 ± 50 (colonies/ml) and 230 ± 30 (colonies/ml), respectively. This detection level is consistent with the result of sensitivity test using rabbit whole blood spiked with fungal cells.

Identification of the Fungi from Blood Culture Sample

Blood from a patient was cultivated in a blood culture bottle for 7 days; one culture positive and one negative sample were subjected to microarray analysis. Samples from blood culture bottle were used directly as a template for nested PCR amplification. The microarray result was consistent with the identification made by the Chiba University Hospital clinical laboratory.

Isothermal Sample DNA Amplification

Because PCR amplification requires a thermal cycler, which is not always available, in small hospitals, or in less developed regions, we attempted to carry out isothermal amplification for biotin labeling of sample DNA using RPA technique [13]. The RPA cycle was performed using 1.5 μl cell suspensions of various fungi as template (Fig. 3). If amplification reaction was successful, the microarray system we developed gave correct identification results in all of the tested samples, even though several amplification bands were observed in some samples.

Fig. 3
figure 3

Isothermal amplification. Agarose gel electrophoresis of fungal DNA obtained by isothermal amplification using TwistAmp Basic kit. Lanes: M, molecular marker; Af, A. fumigatus; Ca, C. albicans; Cn, Cryptococcus neoformans; Fs, FSSC; Mf, Malassezia furfur; Ta, T. asahii; Ro, Rhizopus oryzae

Full size image

Discussion

Fungal infections cause severe morbidity and mortality in immunocompromised patients. Early start of proper treatment is crucial point to achieve better outcome in these patients. Because sensitivity to antifungal drugs differs among different fungal species, identification of the causative fungal agent is important for proper treatment. Rapid detection and identification of pathogens are therefore key points for diagnosis.

Recently, microarray methods have been developed for the identification of a variety of pathogens, including viruses [17], bacteria [4, 8], and fungi [5–7]. These methods are powerful tools to simultaneously detect multiple pathogens. In this study, we developed an easy-to-use, rapid and inexpensive microarray method utilizing biotin-conjugated HRP and color development of the substrate for signal detection. In addition to making the detection system straightforward, we immobilize DNA probes to ordinary plastic slides without any surface modification using UV irradiation via the poly-T anchor of the capture probes [14]. This immobilization system allowed us to use inexpensive, mass-produced, and commercially available ordinary plastic slides as the DNA microarray substrate.

For our DNA microarray, we selected ITS regions of fungal rRNA genes as target because the ITS sequence have been widely used to identify fungi. Although it was difficult to design species-specific probes for some of the fungal genera (Alternaria sp., Rhizomucor sp., Mucor sp., Trichosporon sp.), pathogenic species of these genera have similar MIC values against several antifungal drugs, making designing genus-specific probes useful [18–21]. To our knowledge, the number of fungal species/genera that could be identified by our array system, 42 species from 24 genera, is the largest among microarray identification systems reported to date [5–7]. And the number of identifiable fungal species can be further increased depending on demand of clinical use. Considering the increasing incidence of infectious diseases caused by fungi, this microarray system, which can be used to identify a variety of fungi simultaneously, has great potential.

The sensitivity of our microarray system was evaluated using whole blood spiked with a certain number of fungal cells and serum with fungal DNA. When 1 μl of either sample was used directly as the PCR template, the limit of the detection was 103 cells/ml for blood, and 5 pg/ml of DNA for serum. Nested PCR increased the sensitivity to 50 fg/ml of DNA in serum, but no change in sensitivity was observed in the blood sample. According to calculations, in the 103 cells/ml blood sample, 1 μl of template contains 1 cell, so in case of lower concentration sample, 1 μl of template does not contain any cells. That is why the amplification of 102 cells/ml sample was not constant and sensitivity could not increase by nested PCR. However, DNA extraction or concentration of cells from larger volume of blood or serum sample will have a possibility to decrease detection limit.

When we tested the sample from an infected mouse model and blood culture, it was difficult to get enough intensity in microarray detection with only one step of PCR amplification. This indicated that the amount of DNA in the actual clinical samples is smaller than the detection limit of our microarray system. Nested PCR, however, increased the sensitivity of amplification, and the nested PCR samples successfully gave the expected diagnostic results.

The PCR step in conventional microarray systems has remained as a crucial barrier for wide use in laboratories or hospitals not equipped with a PCR machine. In the present work, we adopted isothermal amplification of DNA samples using the RPA technique as an alternative to PCR to solve this problem. Although, the RPA technique has been used for rapid identification of viruses and bacteria [13], this is the first report of fungal DNA amplification by the RPA technique directly from a fungal cell suspension within 1 h. However, in the present study, the RPA method was found to be less sensitive than the conventional PCR technique. Further optimization of sample preparation and RPA conditions are expected to yield improved results.

In conclusion, we were able to establish a rapid microarray system that can specifically identify a variety of fungal pathogens. Furthermore, we demonstrated that the ABC method could yield enough sensitivity to detect signals from clinical samples, providing an alternative to expensive fluorescence-scanning methods. We also demonstrated that the isothermal amplification technique in combination with this array system has high potential for future applications, such as for bedside diagnosis. This type of assay technique enables simultaneous identification of several agents in a few, relatively simple steps and therefore will become a useful tool in the identification of a wide range of both pathogenic and nonpathogenic microorganisms.