INTRODUCTION

The human oral cavity is a unique ecosystem with a great variety of ecological niches that can be colonized by microorganisms, including yeasts ¹ . The advent of techniques based on DNA sequencing has enabled phylogenetic recognition of yeast species considered cryptic phylogenetically close. They were Candida albicans, Candida dubliniensisand Candida africana ^{2, 3} .

Correct identification of these cryptic species in a clinical setting is relevant from an epidemiological and medical standpoint, and to better understand evolution of antifungal resistance. Moreover, rapid identification is crucial to clinical treatment of local or systemic candidiasis. There are different phenotypic methods for distinguishing species, but they are laborious, time-consuming and do not provide definitive confirmation. In contrast, molecular methods are currently considered the gold standard since, unlike phenotypic methods, they provide appropriate sensitivity and specificity. ^{4,5, 6} There are still limitations associated with rupture of yeast walls to release gDNA for amplification. Due to the complex structure of fungal cell wall, it is difficult to produce cell lysis, which limits the sensitivity of PCR assays. In addition, the fungal load of Candida spp. in colonized sites is relatively low ⁷ . With the advent of commercial extraction kits, quality and quantity of DNA obtained has greatly improved. However, not all low-complexity laboratories can afford to purchase them for routine diagnosis.

In our experience, C. dubliniensis has been the species most frequently recovered from oral samples, probably due to the selective pressure caused by the administration of antifungal agents, since C dubliniensis has higher susceptibility profiles than C albicans. In order to identify these species correctly, we needed to find a simple method for rupturing the cell wall ^{4, 8} .

The aim of this study was to compare the performance of different methods for extracting gDNA from Candida albicans and Candida dubliniensis, subsequently amplifying DNA by PCR and Real Time PCR for correct molecular identification.

MATERIALS AND METHODS

This study analyzed 52 Candida albicans and Candida dubliniensis isolates obtained from subgingival biofilm from patients living with HIV with clinical signs of periodontal disease ¹⁰ , who were under high-activity antiretroviral treatment (HAART) but were not receiving antibiotic or antifungal treatment ^{4, 8} . Informed consent was obtained according to the Declaration of Helsinki. The project and informed consent were approved by the Facultad de Odontología, Universidad de Buenos Aires (FOUBA) Ethics Committee, 023/2019-CETICA-FOUBA.

All diagnoses and treatments were conducted at the periodontics service of the High-Risk Patients Oral Care Unit (CLAPAR I), Facultad de Odontología, Universidad de Buenos Aires; Hospital general de agudos Dr. Juan A. Fernández”, and Hospital de Infecciosas Francisco Javier Muñiz. Patients voluntarily signed an Informed Consent after receiving an explanation of dental practices and benefits of participating in the protocol. Subgingival biofilm samples were collected using the protocol described by Gliosca et al. ⁸ .

Phenotypic methods were compared with molecular ones to identify the isolations at species level.

RESULTS

In this study, 52 yeast isolates were identified by PCR and qPCR, of which 16 were C. albicans and 36 were C. dubliniensis.

Five replicates of each gDNA extraction method were performed per isolate. Tukey’s post hoc, mean concentrations, their respective standard deviation (std) and the standard error for all methods are shown in Tables 1 and 2.

Results of the comparisons are statistically significant (p<0.001) except for C. albicans in METI with MET3. Commercial methods were more efficient for concentrating gDNA than in-house methods, although MET 2 and MET 4 provided acceptable yields. Regarding the subsequent amplification, for C. albicans, the use of zymolase in ddH₂O and LiOAc - SDS were the most effective in-house methods in PCR (100%). In qPCR, the least effective methods were heating (0%) and heating - freezing (0%) with mean melting temperature 84 °C and std 1.5, followed by LiOAC - SDS with mean melting temperature 85 °C and std 0.7.

For C. dubliniensis, statistical differences were observed in all comparisons. All in-house methods except heating followed by freezing enabled identification, though clear bands in 1.3% agarose gel were better with zymolase in ddH₂O, and LiOAC - SDS in PCR. For qPCR, all methods enabled identification, but the least effective regarding melting temperature were heating followed by freezing and LiOAC - SDS, with mean 81.8 and std 0.9 and 0.3, respectively. Two commercial kits were equally effective for both species in both PCRs (Figs. 1 and 2; Table 3).

Fig. 1 Melting point of C. albicans86°C (+0.5) and C. dubliniens 82° C (+0.5) using as extraction method heating at 100°C in qPCR.  

Fig. 2 1.3% agarose gel in TAE buffer (Tris, Acetic Acid, EDTA). Block 8: Ladder 100 pb. Lane 1 positive strain for Candida albicans, lane 2 positive strain for Candida dubliniensis lanes 3, 4, 5, 6 negative strains, lane 7 negative control. 

DISCUSSION

C. albicans and C. dubliniensis share many phenotypic and biochemical characteristics. Presumptive identification by these methods used routinely in low-complexity laboratories continues to be a problem because they do not provide definitive identification data. Distinction between these two species is important in terms of treatment, in order to understand the clinical and epidemiological significance of the role played by C. dubliniensis in human infections ^{4,5,22, 24-27} .

Although C. dubliniensis is often isolated from oral samples in Argentina ^{4,8,24, 26} , phenotypic characterization underestimates it, and it is reported mostly as C. albicans. In agreement with Livério et al. 2017, none of the phenotypic tests alone, proved to be highly effective for conclusive identification of these species ²⁷ .

Molecular techniques are more sensitive and specific, but cell wall lysis is the main obstacle to efficient gDNA recovery. Conventional methods using enzymes for chemical rupture or glass beads for physical rupture, generally followed by lysis with detergents, are time-consuming and costly for application to many samples. Moreover, methods that use phenol-chloroform are hazardous to health unless used in suitable conditions ²⁰ .

Most phenotypic methods provide unsatisfactory results (false positives and negatives), so the most appropriate methods are polymerase chain reaction (PCR), mass spectrometry (MALDI-TOF) and sequencing genomics ¹⁸ .

It is important to bear in mind that one of the main limitations in the use of PCR techniques to identify Candida spp. is that there is no consensus on the methods for cell wall rupture ²⁹ . However, molecular methods provide conclusive identification, and are fast and accurate, though they are more expensive and require specific equipment ²⁷ .

Routine identifications, such as different phenotypic methods and amplification of the ITS regions, do not discriminate properly among the 3 species, underestimating C. africana and C. dubliniensis. However, due to polymorphism of the HWP1 gene, its amplification by PCR enables these species to be distinguished with greater certainty ³⁰ .

The qPCR technique based on melting curves analyzed with SYBR Green is a simple, fast method to distinguish C. albicans from C. dubliniensisthrough the ITS1 and ITS2 regions, but not for C. africana ²¹ . Our study did not isolate C. africanafrom subgingival samples.

In 2009, a second allele for the HWP1 gene with 850 bp was described in C. albicans ²³ . Shan et al. reported that C. albicans produced two DNA fragments, demonstrating that C. albicans isolates were heterozygous at the HWP1 locus ³¹ . Fontecha et al. 2019 found that most of their isolates were heterozygous, concluding that the HWP1 gene could be considered a good marker for identifying cryptic species in this complex ⁵ , while a study in 2017 reported found 5 different genotypes ³⁰ , which provides a pattern of polymorphism presented by this gene.

CONCLUSIONS

This study aims to provide simple, inexpensive strategies for phenotypic identification and molecular confirmation of Candida albicans and Candida dubliniensis for non-reference laboratories with low complexity and/or low budgets.

The advantages of in-house extraction methods used in this study are based on their simplicity, use of minimal amounts of reagents, shorter identification time and avoiding the use of enzymes, phenol and glass beads.

The simplest, cheapest and most effective methods were heating at 100 °C for qPCR and the use of LiOAc - SDS for PCR.