Recombinant polymerase amplification combined with lateral flow strips for the detection of deep-seated Candida krusei infections

The incidence of Candida infections in intensive care units (ICU) has significantly increased in recent years, and these infections have become one of the most serious complications threatening the lives of ICU patients. The proportion of non-Candida albicans infections, such as Candida krusei and Candida glabrata infections, which are resistant to fluconazole, is increasing each year. Early identification of the strains causing Candida infections is important for the timely implementation of targeted treatments to save patients’ lives. However, the current methods of direct microscopy, culture, and histopathology, as well as other diagnostic methods, have many shortcomings, such as their low sensitivity and long assay times; therefore, they cannot meet the needs for early clinical diagnosis. Recombinant polymerase amplification (RPA) is a promising isothermal amplification technique that can be performed without sophisticated instruments and equipment, and is suitable for use in resource-poor areas. RPA combined with lateral flow strips (LFS) can be used to rapidly amplify and visualize target genes within 20 min. In this study, RPA-LFS was used to amplify the internal transcribed spacer 2 (ITS2) region of C. krusei. The primer-probe design was optimized by introduction of base mismatches (probe modification of five bases) to obtain a specific and sensitive primer-probe combination for the detection of clinical specimens. Thirty-five common clinical pathogens were tested with RPA-LFS to determine the specificity of the detection system. The RPA-LFS system specifically detected C. krusei without cross-reaction with other fungi or bacteria. A gradient dilution of the template was tested to explore the lower limit of detection and sensitivity of the assay. The sensitivity was 10 CFU/50 µL per reaction, without interference from genomic DNA of other species. The RPA-LFS and qPCR assays were performed on 189 clinical specimens to evaluate the detection performance of the RPA-LFS system. Seventy-six specimens were identified as C. krusei, indicating a detection rate of 40.2%. The results were consistent with those of qPCR and conventional culture methods. The RPA-LFS system established in our study provides a reliable molecular diagnostic method for the detection of C. krusei, thus meeting the urgent need for rapid, specific, sensitive, and portable clinical field testing.


Introduction
Research on C. krusei has been dominated by drug resistance studies, and the early and rapid diagnosis of infections remains a clinical challenge (He et al., 2015). Currently, the common clinical diagnostic methods for pathogenic fungi include microscopic observation of morphological features, culture in specific media, serological antigen antibody testing, and histopathological analysis (Zarrinfar et al., 2012;Taghizadeh-Armaki et al., 2017). A variety of clinical specimens are commonly used for testing, such as sputum, urine, stool, nails, hair, pus, cerebrospinal fluid, blood, and bodily fluids (Esmailzadeh et al., 2018;Nourizadeh et al., 2019). Microscopy, culture, and pathology remain the gold standard for diagnosing fungal infections (Graus et al., 2017). However, culture is time-consuming (24-72 hours), does not allow for rapid diagnosis (Stone et al., 2013), is susceptible to environmental contamination, and is dependent on the proficiency of the operator.
Candidaemia is the most common clinically invasive Candida infection, and only 30% of patients have positive blood cultures, according to the literature (Quindoś, 2014). CHROMagar Candida medium (Nanjing Yiji Biochemical Technology Co., Ltd., Nanjing, China) is commonly used in clinical practice to identify C. albicans, C. krusei, C. tropicalis, and C. glabrata, but is limited in the identification of other Candida species (Barmar et al., 2021). The diagnostic accuracy of CHROMagar Candida medium is 85.6%, and the misdiagnosis rate is 14.4% (Marr, 2004).
With advances in molecular biology understanding and techniques, various molecular biology methods have been applied to the clinical diagnosis of pathogenic fungal infections to overcome the shortcomings of traditional methods (Yong et al., 2008;Najafzadeh et al., 2021;Kashefi et al., 2021). Proteinbased detection methods, such as matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS, Bruker), and nucleic acid-based diagnostic methods, including DNA fingerprinting, PCR techniques, and DNA hybridization, are currently used as supplements to traditional diagnostic methods, because no standardized guidelines for those methods are available (Alam et al., 2014;Hedayati et al., 2019). Whereas isothermal amplification of nucleic acids amplifies specific DNA or RNA at specific temperatures, traditional PCR techniques require special instruments for automatic temperature control during the denaturation, annealing, and extension steps. Nucleic acid isothermal amplification technology greatly simplifies the instrumentation required for nucleic acid amplification. A constant temperature water bath can be used for the reaction. Moreover, the reaction time is substantially shorter than that of traditional PCR. This method thus meets the clinical needs of simple operation and rapid diagnosis (Jacobsen et al., 2007). The main isothermal amplification techniques reported to date are loop-mediated isothermal amplification (LAMP) (Fallahi et al., 2020), nuclear acid sequence-based amplification (NASBA) (Huang et al., 2019), rolling circle DNA amplification (RCA) (Carinelli et al., 2017), single primer isothermal amplification (SPIA) (Yang et al., 2021), helicase-dependent isothermal DNA amplification (HAD) (Jeong et al., 2009), and strand displaced amplification (SDA) (Zhang et al., 2017), which have been used in clinical practice.
Recombinase polymerase amplification (RPA) is a recently developed thermostatic amplification technology that combines the advantages of the above methods and addresses their shortcomings, thus enabling rapid, on-site, sensitive, and portable diagnosis (Piepenburg et al., 2006). RPA uses recombinase to open the double-stranded DNA and facilitates primer binding to the target fragment, whereas Bsu DNA polymerase, which has chain-switching activity, recognizes the 3' end of the primer and enables stable amplification (Rosser et al., 2015;Dai et al., 2019;Cossio et al., 2021).
In our study, RPA and qPCR primers were designed for the diagnosis of C. krusei infections on the basis of the ITS2 sequence. A total of 24 strains of C. krusei strains (20 clinical strains and four standard strains) were collected and used as positive control strains. Thirty-five strains of common clinical pathogens were collected as negative control strains. All strains were validated through amplification of the ITS2 sequence. The differences in sensitivity and specificity between the RPA and qPCR techniques for the diagnosis of C. krusei infection were investigated and compared.

Ethics statement
The study protocol was approved by the Medical Ethics Committee of the Second People's Hospital of Lianyungang City (Lianyungang, Jiangsu, China; permit number 2020013), and informed consent was obtained from patients before the collection of clinical specimens.

Genomic DNA extraction
Unless otherwise indicated, all bacterial strains (1 mL of 10 6 CFU/mL) were boiled at 100°C for 10 min before being used as templates for nucleic acid amplification. For C. krusei and other fungi, genomic DNA was extracted and purified from cultures or clinical specimens with a GeneJET Genomic DNA Purification Kit (Tiangen Biotechnology Co., Ltd., Beijing, China) according to the manufacturer's instructions. Extracted Genomic DNA was quantified with a Qubit 4 fluorometer (Thermo Fisher Scientific), according to the manufacturer's instructions.

Primer and probe design and screening
Two pairs of RPA primers based on ITS2 were designed in Primer Premier 5.0 software (Premier Biosoft International, CA, USA). After the sequence of the specific target region was entered, the following primer parameters were set. The product size was set to 80-150 bp. The primer size was set to 30-35 bp, with no more than three consecutive bases with complementary pairing at the 3′ end, a maximum hairpin score of nine, a maximum primer-dimer score of nine, and a maximum poly-X of five. The primers were confirmed with Primer-BLAST on the NCBI website (https://www.ncbi.nlm.nih. gov/tools/primer-blast), which was also used to confirm the species specificity of the sequences of the primers and probes. Forward primer 5, which extended 16 bp upstream of the target, was evaluated for probe and reverse primer performance in Primer Premier 5 software, to avoid the formation of dimers and hairpin structures. Between the probe and reverse primer, we aimed to achieve a probe size of 46−51 bp, a GC content of 30 −80%, and a Tm of 55−80°C. The 5' end of the probe was labeled with FITC, the 3' end was closed with a C3 spacer, the 5' end of the forward primer was in the middle of the probe, the first base substitution extending backward was replaced with tetrahydrofuran (THF), and the 5' end of the reverse primer was labeled with biotin.

RPA reaction
RPA reactions were performed with a TwistAmp ® Liquid DNA Amplification Kit (TwistDx Inc., Maidenhead, UK) according to the manufacturer's instructions. The 25 µL reaction system contained 12.5 µL of 2× reaction buffer, 2.5 µL of 10× Basic e-mix, 1.25 µL of 20× core mix, 1.2 µL of 10 µM forward primer, 1.2 µL of 10 µM reverse primer, and 4.6 µL of distilled water. A volume of 1.25 µL of 280 mM magnesium acetate and 0.5 µL of template were added to the lid of each reaction tube. After a short centrifugation, the reaction mixture was incubated at 37°C for 20 min. The RPA amplification products were purified with a PCR cleanup kit (Meiji Biotechnology, Shanghai, China) and separated on a 1.5% agarose gel.

RPA-LFS assay
RPA reactions were as described above, except that each 25 mL reaction contained 1.05 mL of each primer (10 mM), 0.3 mL of probe (10 mM), 1.0 mL of template, and other standard reaction components. Primers and probes were synthesized by Anhui General Biotechnology Co., Ltd. To initiate the reaction, 1.25 mL of magnesium acetate (280 mM) was added, and the reaction mixture was incubated at 37°C for 20 min. Then 5 mL of the amplification product was diluted 20 fold, and the LFS was inserted into 100 mL of the diluted solvent for approximately 2 minutes. The test and control lines were then visually inspected.

Sensitivity of the RPA-LFS assay
A 10-fold gradient dilution was tested from 10 0 to 10 6 CFU/µL (reaction volume of 50 µL containing 1 µL of C. krusei inactivation solution). A 10 6 CFU/µL inactivating solution of another common pathogen (C. glabrata) was also prepared for the RPA-LFS reaction.

Quantitative PCR analysis
The primers and probes are listed in Table 1. Primers used for qPCR were targeted to ITS2 of C. krusei. The qPCR reaction mixture consisted of 12.5 mL of MonAmp ™ TaqMan qPCR mixture (Mona Biologicals Co., Ltd., Suzhou, China), 0.5 mM of forward and reverse primers, 1 mL of genomic DNA, and distilled water to 25 mL. qPCR was performed with a Roche LightCycler 480 qPCR machine, with a program of 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 55°C for 60 s.

Results
Primer validation screening strategy ITS2 was selected from the C. krusei genome as the target for RPA-LFS detection. Two potential primer pairs were obtained by searching NCBI primer-BLAST for primers specific to the ITS2 sequence ( Table 1). The primers were initially screened through target gene fragment amplification and a no-template control. The amplification products were electrophoresed on agarose gels to compare the amplification performance of the target gene and primer dimer formation in the no-template control. The primer pair F2/R2 was selected because it showed the best amplification performance and no cross-dimer formation ( Figure 1A). Candidate probes were obtained by extending the 3' end of the forward primer F2 by 16 bp. All possible dimers generated by the probe and reverse primer were predicted, and the bases were modified (Table 1) until no dimer was formed ( Figure 1B). Finally, five forward primers were designed, screened, and tested upstream of the probe. The LFS results showed that F3/P/R2B, F4/P/R2B, and F5/P/R2B amplified the target gene fragment efficiently, whereas F6/P/R2B and F7/P/ R2B amplified the target less efficiently. The no-template controls of F3/P/R2B, F4/P/R2B, F6/P/R2B, and F7/P/R2B all showed positive results, and only F5/P/R2B met the assay requirements ( Figure 1C). Therefore, F5/P/R2B was used in subsequent experiments.

Sensitivity of the RPA-LFS assay
To determine the limit of detection (LOD) of the RPA-LFS system for the detection of C. krusei, we tested 10-fold gradient dilutions of C. krusei from 10 0 to 10 6 CFU/µL (reaction volume: 50 µL, with 1 µL of C. krusei genomic DNA added to each reaction). A red band appeared on the test line at 10 CFU/µL and became progressively darker as the concentration of C. krusei increased (Figure 2A). To test whether other fungal DNA might interfere with the assay, we added 10 6 CFU/µL of genomic DNA of another common pathogen, C. glabrata, to the RPA reaction. The genomic DNA of C. glabrata did not interfere with the detection of C. krusei ( Figure 2B). Therefore, we concluded that the LOD for the RPA-LFS system developed in this study was 10 CFU/50 µL per reaction, and genomic DNA from other fungi did not interfere with assay sensitivity.

Specificity and inclusivity of the RPA-LFS assay
To confirm the inclusivity and specificity of F5/P/R2B, we performed RPA-LFS on four reference strains (C. krusei ATCC  Table 2). The four reference strains and 20 clinical isolates showed positive results (Figure 3), whereas all other pathogenic cultures showed negative results ( Figure 4). Thus, the primerprobe set showed good inclusion and specificity for C. krusei, effectively detecting this species without cross-reaction with other pathogens.

Clinical specimen testing
To verify the practical application of RPA-LFS, we collected 189 clinical specimens for RPA-LFS and qPCR assays. The results are shown in Table 3. Seventy-six specimens contained C. krusei, and had a detection rate of 40.2%, whereas 113 tests indicated non-C. krusei, in agreement with the detection results with qPCR and traditional culture methods. Neither method produced false positive or false negative results. RPA-LFS had the same accuracy as qPCR, and the results were consistent with those of the traditional culture method.

Discussion
Candida, Aspergillus, and Cryptococcus are the three most common clinically invasive fungi. Among these, Candida species are the most common clinically opportunistic fungi, with the highest Determination of the limit of detection (LOD) of the C. krusei RPA-LFS assay. (A) The LOD for the RPA-LFS assay for C. krusei was established with the primer-probe set F5/P/R2B on genomic DNA extracted from an inactivated bacterial broth of C. krusei at serial dilutions from 10 0 to 10 6 CFU for each reaction. The image shows the results of the RPA-LFS. The number of templates is indicated at the top of the bar graph. (B) Picture showing the results of the RPA-LFS assay with primer-probe set F5/P/R2B and 10 6 CFU of C. glabrata for interference.
reported carriage rate for the digestive tract (50%), vagina (20%-30%), skin surface (2%), and pharynx (1-4%). Candida can cause diseases when the body's immunity is low (Zarrinfar et al., 2016). Two main factors influence susceptibility to Candida infections in humans: (1) the human body itself, including factors such as organ transplantation, low neutrophil counts, or impaired cellular immunity, and (2) external factors, such as broadspectrum antibiotics and surgical procedures resulting in nosocomial infections. Candida infections have been found to affect 80% of patients hospitalized in the ICU for more than 1 week; moreover, 10% of these patients may develop invasive infections, thus threatening their prognosis (Arendrup et al., 2011;Gökahmetoglu et al., 2016). At least 17 species of Candida cause human disease, among which C. albicans causes the most common infections (Minooeianhaghighi et al., 2019). However, in recent years, with organ transplantation, and the use of immunosuppressive agents, broad-spectrum antibiotics, and the increasing application of glucocorticoids, the spectrum of common clinical Candida species has changed. C. krusei, C. glabrata, C. tropicalis, and C. parapsilosis have increasingly caused non-Candida albicans infections, accounting for 35%-65% of cases (Krcmery and Barnes, 2002;Mohammadi et al., 2013;Guinea et al., 2014). C. krusei is a common clinical opportunistic pathogenic that can cause limited or systemic infections in humans, and has attracted research attention because of its natural resistance to fluconazole and its increasing prevalence.
C. krusei results in a high mortality rate after infection, and early and rapid diagnosis is important for patient prognosis. The traditional diagnostic methods, based on the morphological and physiological characteristics of C. krusei, are time-consuming and complicated to perform, thus often delaying patient treatment. Automatic or manual rapid diagnostic kits have become available for clinical applications, thereby overcoming several shortcomings of traditional diagnostic methods. With the development of molecular biology diagnostic technologies, related diagnostic techniques have been applied in clinical settings for early Validation of the specificity of primer pair F5/P/R2B toward C. krusei. #1-#20 refer to 20 isolates of C. krusei from clinical specimens; NTC indicates no-template control. The positions of the test and control lines are marked on the right side of the image. Reactions were performed at 37°C for 20 min. The image is representative of three independent experiments.  and rapid diagnosis of pathogenic fungi. In recent years, many nucleic acid amplification-based methods have been reported for the diagnosis of infections with common clinical Candida species, including C. krusei. Brito et al. (2009) initially reported the use of ITS primers to amplify Candida DNA and identify the species class according to the length of electrophoretic bands. The authors found that the ITS3 and ITS4 primer pairs amplified fragments of similar length (at 331 bp) for C. tropicalis and C. krusei, thus failing to distinguish these species. Therefore, the ITS2 primer pair, which allowed for good differentiation between C. krusei and other Candida species, was chosen to perform the initial strain validation test. Khlif et al. (2009) have used nested PCR for the diagnosis of common clinical Candida infections and have performed two PCR reactions, thereby improving the sensitivity and specificity. However, that method is more complicated than RPA and requires an expensive PCR instrument to perform the reactions.

B A
A qPCR assay has been described by Decat et al. (2013); however, in that study, the most closely related strains of C. krusei were not included as controls. The RPA test herein used primers targeting the ITS2 sequence, and was able to achieve specific amplification within 2.5 hours and provide advantages over qPCR. A comparison of the sensitivity, specificity, practicality, and speed of the RPA and qPCR techniques, indicated that both RPA and qPCR techniques had 100% specificity in the diagnosis of C. krusei infection. However, RPA has the benefits of being a thermostatic amplification technique that does not require temperature changes to achieve specific amplification of DNA and can be combined with LFS to meet visualization needs. Although RPA does not require high experimental conditions, non-specific amplification due to aerosol contamination can be problematic. This drawback can be prevented through strict spatial partitioning to isolate the areas in which reaction and detection are performed. qPCR requires specific fluorescence PCR instruments and is expensive; moreover, it requires 2.5 hours to complete, whereas RPA can be completed in 20 min (Wu et al., 2017;Tian et al., 2018;Wang et al., 2021).
Overall, the RPA technique was easy to perform, the reaction results were readily determined, the time required for the reaction was short, and no specialized instrumentation was necessary. Thus, this method may enable early diagnosis in resource-poor areas. Many technologies have been reported, such as those used to detect Listeria monocytogenes, C. albicans, or C. neoformans (Wang et al., 2019;Wang et al., 2021;Wang et al., 2022). However, the traditional diagnostic methods for pathogenic fungi, which are commonly used in clinical practice, are time-consuming, prone to falsepositive and false-negative results, and susceptible to researchers. Molecular biology diagnostic techniques have been gradually applied in clinical practice because of their time-consuming and high specificity. In this study, we diagnosed fluconazole-resistant C. krusei by using the RPA-LFS technique, which is rapid and specific, and therefore may aid in the early diagnosis of C. krusei infections in clinical settings.

Data availability statement
The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding authors.