Rapid identification of clinical common invasive fungi via a multi-channel real-time fluorescent polymerase chain reaction melting curve analysis

Abstract The incidence of invasive fungal infections (IFIs) has recently increased, and early and accurate diagnosis of IFIs is important for the rational selection of antifungal drugs with high efficacy. We developed a method for rapid and accurate clinical diagnosis of IFIs and provide a reference for personalized drug treatment. We designed and screened fungal internal transcribed spacer regions with universal primers and designed 8 TaqMan detection probes to establish a multi-channel real-time fluorescent polymerase chain reaction (PCR) melting curve analysis (MCA) method. The sensitivity, specificity, and reproducibility of this method were investigated using standard fungal strains and clinical isolates. Candidemia was detected using the MCA method. The limit of detection and assay cut-off (melting temperature [Tm]) for Candida albicans were 0.05 pg/μL and 66.50 °C; Candida glabrata were 0.1 pg/μL and 66.25 °C; Candida tropicalis were 0.1 pg/μL and 60.15 °C; Candida krusei were 0.1 pg/μL and 72.15 °C; Candida parapsilosis were 0.2 pg/μL and 63.10 °C; Candida guilliermondii were 0.1 pg/μL and 61.84 °C; Cryptococcus neoformans were 0.1 pg/μL and 65.50 °C; Aspergillus flavus were 0.05 pg/μL and 71.50 °C; Aspergillus terreus, Aspergillus fumigatus, and Aspergillus niger were 0.05 pg/μL and 76.80 °C. Analytical specificity was evaluated using 13 clinical pathogens including Streptococcus pneumoniae, Staphylococcus aureus, and Haemophilus influenzae, etc. No false-positive results were obtained for any of these samples. The MCA method can detect and identify different candidemia simulations. The limit detection concentration of C albicans was 44 cfu/mL, C glabrata was 73 cfu/mL, C tropicalis was 29 cfu/mL, C parapsilosis was 21 cfu/mL, C krusei was 71 cfu/mL, and C guilliermondii was 53 cfu/mL. The multi-channel real-time fluorescence PCR melting curve analysis displayed high sensitivity and specificity in detecting various clinically invasive fungi. Furthermore, it simultaneously detected the genera Candida, Cryptococcus, and Aspergillus and identified Candida at the species level. Our method can facilitate early and accurate clinical diagnosis and personalized medication regimens.


Introduction
Invasive fungal infections (IFIs) have increased with the widespread use of broad-spectrum antibiotics, immunosuppressive agents, antineoplastic drugs, and in-depth development of organ transplantation, and various invasive diagnostic techniques. [1][2][3] Candida, Aspergillus, Pneumocystis, and Cryptococcus neoformans are the primary pathogens causing IFIs, with Candida responsible for the largest number of cases, [4,5] and the infection rates of Aspergillus, Pneumocystis, and C neoformans have increased. [6] IFIs primarily occur in patients with severe underlying diseases, malignant tumors, and other severe diseases compromising immune function and in those undergoing organ transplantation. [7] The clinical manifestations of IFIs are often non-specific and easily masked by primary underlying diseases. Early diagnosis is difficult, often resulting in delayed diagnosis, misdiagnosis, and delayed treatment. Furthermore, IFIs have a poor prognosis and are associated with high mortality, [8] with approximately 1.4 million deaths worldwide each year. [9] Sun et al [10] reported that the total mortality rate with IFIs is 13.4%, and Barnes [11] reported that the mortality rates associated with invasive candidiasis (IC) and invasive aspergillosis are 36% to 63% and 70%, respectively. The authors have no conflicts of interests to disclose.
Laboratory detection of IFIs primarily involves traditional detection methods including direct microscopy, culture, and histopathology. [12] Traditional methods have a low positive detection rate and poor sensitivity and do not yield a rapid positive clinical diagnosis. [13] Serological analysis primarily includes the 1,3-b-D-glucan test (G test), galactomannan test (GM test), and latex agglutination test. Although these methods potentially provide a basis for the early diagnosis of IFIs, they can be influenced by various factors, leading to false-positive findings. [14] Furthermore, they do not accurately detect fungal species. [15] Molecular biology technologies based on PCR have been widely used to detect IFIs. Fluorescence PCR melting curve analysis (MCA) is an emerging detection method for identifying fungal species without sequencing. [16] MCA has a high sensitivity, throughput, speed, and accuracy, [17,18] and it is less expensive; thus, it is applicable for detecting fungal infections. [19,20] This method is based on the principle that different double-stranded DNA molecules have different Tm values, and changes in the shape of the melting curve can be monitored using fluorescent dyes or probes to rapidly and accurately detect and identify various fungi. [21] The TaqMan probe melting curve analysis uses the TaqMan probe rather than the SYBR Green dye to specifically bind to the target DNA. By analyzing the melting curve and Tm value of specific binding products, fungal detection at the species level can be achieved. [22] Furthermore, TaqMan probes can link fluorescent reporter groups of different wavelengths, and thus completely harness the advantages of multi-channel fluorescent probe PCR for simultaneous detection of various fungi. For example, Alonso et al [23] performed real-time fluorescent probe PCR melting curve analysis to detect and identify Aspergillus, and Valero et al [24] also detected and identified Mucor, Rhizopus, and endemic diseasedrelated fungi and established a corresponding melting curve library. In this study, we used a combination of PCR amplification technology and multi-channel real-time fluorescent probe melting curve analysis for closed-tube detection and species identification of common IFI-causing pathogens (Candida, Aspergillus, and Cryptococcus). This study was conducted to develop a method for rapid and accurate clinical diagnosis of IFIs and provide a reference for personalized drug treatment.

Strain
The strains used in this study were subjected to sequencing analysis with specific primers (by Shanghai Sunny Biotech Co., Ltd., Shanghai, China). The following fungal strains were used (

Blood samples
Blood was collected from patients who visited the First People's Hospital for a health checkup, with no clinical symptoms of sepsis and inflammatory markers.

DNA extraction
Strain preparation: Each fungal strain was inoculated on potato glucose liquid medium (Shanghai Qigong Biotechnology Co., Ltd., Shanghai, China) and cultured (30°C, 4 Â g, 2-3 days). Pretreatment: 200 mL aliquots of the culture was centrifuged (4°C , 18,000 Â g, 5 minutes), and then the pellet was treated with 600 mL of 1.2 M sorbitol and 10 U Lyticase (Tiangen Biochemical Technology Co., Ltd., Beijing, China), thoroughly mixed, incubated at 30°C for 30 minutes to disrupt the fungal cell wall, and centrifuged (4°C, 18,000 Â g, 5 minutes), and the supernatant was discarded. DNA extraction: Genomic DNA of the fungi was extracted using the magnetic bead method with the fungal DNA extraction kit (Hangzhou Qianji Biological Technology Co., Ltd., Hangzhou, China) in accordance with the manufacturer's instructions. DNA purity and concentration were measured using the Nano-300 spectrophotometer (Hangzhou Aosheng Instrument Co., Ltd., Hangzhou, China). DNA samples were diluted to 1 ng/mL, 10, 1, 0.1, 0.05, and 0.01 pg/mL.

Design and synthesis of primers and probes
The internal transcribed spacer (ITS) in the fungal genome is a conserved, species-specific 18S rDNA-ITS1-5.8S rDNA-ITS2- 28S rDNA tandem repeat. The ITS sequences of C albicans and C glabrata, C tropicalis, C krusei, C parapsilosis, C guilliermondii, C neoformans, A flavus, A terreus, A niger, and A fumigatus were compared and analyzed with the National Center for Biotechnology Information (NCBI) database. Based on the highly conserved/relatively specific segment of the ITS region of the strain, primer 5.0 software was used to design universal primers (ITSF1/ITSR1)/specific probes (CaP1/CgP1/CtP1/CkP1/CpP1/ CguP1/CnP1/A. P1), and the probes were labeled with different fluorophores (Table 2).

PCR conditions and MCA
PCR amplification and MCA were performed using SLAN-96S (Shanghai Hongshi Medical Technology Co., Ltd., Shanghai, China). The PCR was carried out with a reaction mixture of volume 25 mL comprising 2.5 mL of 10Â PCR buffer, 5 mL of 5Â probe buffer, 0.2 mM dN(U)TPs, 1 U of Hot Start Taq polymerase, 0.2 U of UDG, 0.4 mM ITSF1, 0.04 mM ITSR1, 0.16 mM primers, and sterile water. The PCR program comprised a UDG enzyme reaction for 2 minutes at 50°C, followed by initial denaturation for 3 minutes at 95°C, 60 cycles for 30 seconds at 95°C, 30 seconds at 56°C, and 30 seconds at 72°C, and termination for 2 minutes at 95°C. The amplification products for MCA were cooled for 2 minutes at 40°C and heated to 85°C to monitor fluorescence at 0.06°C/s. Thereafter, 1 ng/mL of DNA each reference fungal strain was analyzed in 10 independent runs and the mean melting temperature (Tm) and standard deviation (SD) values were determined for each melting domain.

Sensitivity and specificity
The genomic DNA samples of each fungal strain were diluted to 10, 1, 0.1, 0.05, and 0.01 pg/mL, and then 4 mL of each sample was analyzed in 20 independent runs and the limit of detection was determined. The genomic DNA of the clinical isolates was extracted using the DNA Extraction kit (Hangzhou Qianji Biological Technology Co., Ltd.) in accordance with the manufacturer's instructions, and 100 pg/mL of each sample was analyzed to evaluate the specificity of the analysis.

Detection of clinical sample
Candidemia' clinical samples were detected using the MCA method. Clinical sample preparation: blood samples were drawn into a 2-mL anticoagulation tube (K2EDTA, BD Medical Equipment Co., Ltd. Shanghai, China). The 6 model clinical isolates of fungal strains were counted, and inoculated blood samples later, in order to obtain the fungal cell number of 10 3 , 10 2 , 10 1 cfu/mL for each of simulated candidemia sample. The diluted sample (400 mL) was added into the pathogen lysis tubes (L) (QIAGEN, Germany) and treated with 10 U Lyticase, thoroughly mixed, and incubated at 30°C for 30 minutes. DNA extraction: simulated candidemia samples' DNA was extracted according to the manufacturer's instructions of the QIAmp UCP Pathogen kit (QIAGEN, Germany). Furthermore, 10 6 cfu/mL of each fungal suspension was used as the positive control and healthy human blood as the negative control. The MCA was performed with DNA samples from the positive control, negative control, and different concentrations of simulated candidemia samples. The research was approved by the local Bioethics Committee of Hangzhou First People's Hospital.

MCA for reference strains
MCA using 6 Candida reference strains, 1 Cryptococcus reference strain, and 4 Aspergillus reference strains revealed reproducible melting peaks for each species. Most strains displayed a single peak (Fig. 1). All Candida, Cryptococcus, and Aspergillus species were directly identified by MCA. The mean Tm ± SD value(s) together with the shape of the melt curve distinguished the 6 Candida spp., 1 Cryptococcus sp., and 1 Aspergillus sp. (Fig. 1, Table 3). However, the other 3 Aspergillus spp. could not be reliably distinguished. The mean Tm ± SD values of A terreus, A niger, and A fumigatus were similar and overlapping.

Sensitivity and specificity
The limit of detection for C albicans, A flavus, A terreus, A niger, and A fumigatus was 0.05 pg/mL; for C glabrata, C tropicalis, C krusei, C guilliermondii, and C neoformans was 0.1 pg/mL; and for C parapsilosis was 0.2 pg/mL. The detection rate of each fungal strain was >95% (Fig. 2). None of the 13 clinical bacterial isolates displayed specific melting curves and Tm values, and no cross-reaction with fungal detection was observed (Fig. 3).

Clinical sample
The positive control (10 6 cfu/mL) displayed specific melting curve and Tm value, whereas the negative control did not show a specific response. The MCA method can detect simulated candidemia with different concentrations of fungal strain. The limit detection concentration of C albicans was 44 cfu/mL, C glabrata was 73 cfu/mL, C tropicalis was 29 cfu/mL, C parapsilosis was 21 cfu/mL, C krusei was 71 cfu/mL, and C guilliermondii was 53 cfu/mL (Fig. 4).

Discussion
As an emerging detection method, fluorescence PCR melting curve analysis is useful for diagnosing IFIs. [25] In this study, for multi-channel real-time fluorescent PCR melting curve analysis, general ITS primers were first used to identify Candida, Aspergillus, and Cryptococcus, and then used specific TaqMan probes to obtain and analyze specific melting curves and Tm values for all fungi, which simultaneously distinguished C albicans, C glabrata, C tropicalis, C krusei, C parapsilosis, C neoformans, and Aspergillus spp. in the same reaction mixture. This method harnesses the advantages of multi-channel fluorescent probe PCR. Based on the reference Tm value of each fungal primer and probe sequence, the CaP1, CgP1, and CnP1 probes with similar Tm values were labeled with CY5, FAM, and ROX fluorophores, respectively; CtP1 and CguP1 with FAM and CY5; CkP1 and A.P1 with ROX and CY5; CpP1 with FAM. These probe-fluorophore pairs yielded specific, significantly different melting curves, and Tm values among fungi in their detection channels. This method involves an asymmetric PCR, that is, the amounts of the 2 primers are different, and when the restriction primers at relatively low concentrations are completely utilized, the other unrestricted primers are amplified, yielding the target single-stranded DNA, which pairs and binds with specific probes. As the temperature of the PCR system increases, target genes can be identified based on the melting curve probe peak. Asymmetric PCRs are influenced by primer concentration and ratio, probe concentration, and Mg 2+ concentration, [26] and the annealing temperature is critical for successful asymmetric PCR. The PCR system for multi-channel real-time fluorescence PCR melting curve analysis was 25 mL, and the primer concentration was gradually increased from 0.04 to 0.4 mM, the ratio of upstream and downstream primers from 5:1 to 20:1, the probe concentration from 0.04 to 0.24 mM, and dN(U)TPs from 0.20 to 0.24 mM, and the annealing temperature from 50 to 60°C to optimize the system and finally determine the optimal PCR system and conditions. The multi-channel real-time fluorescence PCR melting curve analysis is a closed system, thus effectively preventing falsepositive results caused by external influences. This method has high sensitivity and specificity and displays a lower limit of detection of 0.5 to 1 pg/mL during MCA than the method devised by Nemcova et al, [27] with the former showing higher sensitivity. During fungal identification, based on specific melting curves and Tm values, this method distinguished C albicans, C glabrata, C tropicalis, C krusei, C parapsilosis, C neoformans, and A flavus. However, it was difficult to distinguish and identify A fumigatus, A terreus, and A niger using this method because of the similarity of their melting curves and Tm values, which was likely related to the specificity of the A.P1 probe. With respect to the entire probe for >25 nucleotides, a certain degree of conservation among fungal genera and species specificity is required, which would adjust the probe position and Tm values for interaction with the appropriate probe sequence, such that it can be completely combined with the corresponding sequences and simultaneously distinguish fungi. Furthermore, the development of a single probe with greater specificity to identify four Aspergillus spp. is even more challenging. The A.P1 probe detects Aspergillus at the genus level; however, it cannot identify A fumigatus, A terreus, Table 3 Tm ± SD values of fungal strains.   and A niger. This issue can be resolved by designing and synthesizing a single A fumigatus probe, A flavus probe, A terreus probe, and A niger probe to rapidly identify clinically common Aspergillus spp. Invasive candidiasis (IC) is an important type of IFIs, and it has become a major threat to public health. [28] To date, more than 17 different Candida species have been identified as pathogens of IC, and C albicans is the leading and most widely studied pathogen. [29] However, in recent years, the proportion of non-C albicans has increased rapidly. [30] The antibacterial activity of antibacterial drugs varies with strains, and the choice of treatment drugs is also different [31] ; early identification of the pathogenic species of IC is essential for early clinical medication.
In this study, we detected IC using the MCA method; 6 clinically isolated fungal strains were used to simulate candidemia samples, which included 6 positive controls, 1 negative control, and 3 concentration gradients of candidemia samples, and then the MCA was performed. All positive controls were positive, whereas the negative control showed no reaction. The results showed that there was no positive effect of healthy human blood on the established MCA detection system, and each clinical isolate of Candida strain (10 6 cfu/mL) could be detected successfully. Furthermore, the MCA method can detect simulated candidemia with different concentrations of fungal strains, and the lowest detection concentration of C glabrata and C krusei was 70 to 80 cfu/mL, C albicans and C guilliermondii was 40 to 50 Figure 2. Sensitive detection of fungal strains. C albicans: Candida albicans; C glabrata: Candida glabrata; C tropicalis: Candida tropicalis; C parapsilosis: Candida parapsilosis; C krusei: Candida krusei; C guilliermondii: Candida guilliermondii; C neoformans: Cryptococcus neoformans; A flavus: Aspergillus flavus; A terreus: Aspergillus terreus; A niger: Aspergillus niger; A fumigatus: Aspergillus fumigatus. A: The limit of detection for C albicans was 0.05 pg/mL; B: the limit of detection for A flavus was 0.05 pg/mL; C: the limit of detection for A terreus was 0.05 pg/mL; D: the limit of detection for A niger was 0.05 pg/mL; E: the limit of detection for A fumigatus was 0.05 pg/mL; F: the limit of detection for C glabrata was 0.1 pg/mL; G: the limit of detection for C tropicalis was 0.1 pg/mL; H: the limit of detection for C krusei was 0.1 pg/mL; I: the limit of detection for C guilliermondii was 0.1 pg/mL; J: the limit of detection for C neoformans was 0.1 pg/mL; K: the limit of detection for C parapsilosis was 0.2 pg/mL. Wen et al. Medicine (2020) 99:7 www.md-journal.com cfu/mL, C tropicalis and C parapsilosis was 20 to 30 cfu/mL (the concentrations were calculated from the results of the diluted coated plate count). However, there was a melting curve peak of C parapsilosis blood sample at Tm of around 76°C, which was that of Aspergillus. That is, there was contamination by Aspergillus in the detection process. Therefore, the test results of C parapsilosis should be deliberated.
In summary, compared with the gold standard culture method, the MCA method can detect candidemia samples and yield results within 4 to 6 hours, and its sensitivity was 20 to 80 cfu/mL, indicating that the method is fast, accurate, and sensitive. However, our study had some limitations. The clinical sample is a simulated candidemia sample, and not a clinical sample, and we plan to collect samples from patients with invasive candidiasis for verification. Multi-channel real-time fluorescence PCR melting curve analysis is rapid and accurate compared with the traditional culture methods. This method has a high sensitivity and specificity and can simultaneously detect various common  invasive fungi, invasive candidiasis especially, yielding detection results within 4 to 6 hours and facilitating early diagnosis of clinical IFIs. In the future, multi-channel real-time fluorescent PCR melting curve analysis may be combined with other detection methods (G and GM test), to potentially improve the early diagnosis rate of IFIs and providing a reference for clinical rationalization of antifungal treatment, reducing misdiagnosis, improving prognosis, and reducing mortality during clinical diagnosis and treatment.