New insights into the DNA extraction and PCR amplification of minute ascomycetes in the genus Laboulbenia (Pezizomycotina, Laboulbeniales)

Abstract

Molecular studies of fungi within the order Laboulbeniales (Ascomycota, Pezizomycotina) have been hampered for years because of their minute size, inability to grow in axenic culture, and lack of reliable and cost-efficient DNA extraction protocols. In particular, the genus Laboulbenia is notorious for low success with DNA extraction and polymerase chain reaction (PCR) amplification. This is attributed to the presence of melanin, a molecule known to inhibit PCR, in the cells. We evaluated the efficacy of a standard single cell-based DNA extraction protocol by halving the recommended amount of reagents to reduce the cost per extraction and adding bovine serum albumin (BSA) during the multiple displacement amplification step to reverse the effect of melanin. A total of 196 extractions were made, 111 of which were successful. We found that halving the reagents used in the single cell-based extraction kit did not significantly affect the probability of successful DNA extraction. Using the halved protocol reduces cost and resource consumption. Moreover, there was no significant difference in the probability of successfully extracting DNA based on whether BSA was added or not, suggesting that the amount of melanin present in cells of the thallus has no major inhibitory effect on PCR. We generated 277 sequences from five loci, but amplification of the internal transcribed spacer region, the mitochondrial small subunit rDNA, and protein-coding genes remains challenging. The probability of successfully extracting DNA from Laboulbeniales was also impacted by specimen storage methods, with material preserved in > 95% ethanol yielding higher success rates compared to material stored in 70% ethanol and dried material. We emphasize the importance of proper preservation of material and propose the design of Laboulbeniales-specific primers to overcome the problems of primer mismatches and contaminants. Our new insights apply not only to the genus Laboulbenia; Laboulbeniales generally are understudied, and the vast majority of species remain unsequenced. New and approachable molecular developments will benefit the study of Laboulbeniales, helping to elucidate the true diversity and evolutionary relationships of these peculiar microfungi.

Introduction

Traditionally, fungal species identification and delimitation relied on morphological characteristics, but phenotypic plasticity within species and cryptic species posed challenges (Bridge et al. 2005; Cao et al. 2021; Maharachchikumbura et al. 2021). Some groups used alternative traits like enzymatic activity or mating compatibility (Perkins and Raju 1986; Pincus et al. 2007), but these methods are not applicable to unculturable species, thus other unambiguous characters should be used. Early molecular methods based on GC-content and DNA hybridization had limited use, except in yeast studies (Bridge et al. 2005). The introduction of PCR enabled the discrimination of closely related taxa based on molecular data, revolutionizing molecular systematics in mycology (White et al. 1990; Bridge et al. 1998, 2005; Cao et al. 2021; Maharachchikumbura et al. 2021). The internal transcribed spacer of the nuclear ribosomal DNA (ITS) was proposed as the universal fungal DNA barcode marker (Schoch et al. 2012). While the ITS is nowadays often used to delimit and identify fungi, the interspecific variation in this region is too low in some groups, necessitating the use of additional markers (Stielow et al. 2015).

A particularly understudied group of fungi is the order Laboulbeniales (Ascomycota, Pezizomycotina). These microfungi have an obligate association with arthropod hosts. Instead of producing hyphae, Laboulbeniales develop a 3-D structure called a thallus from a two-celled ascospore, which attaches externally to an arthropod’s integument (Blackwell et al. 2020). Molecular studies of Laboulbeniales have been challenging due to the minute size of these fungi (200–300 μm on average), their melanized cells, and their inability to grow in axenic culture (Haelewaters et al. 2015; Sundberg et al. 2018a). Initial attempts using a dry ice protocol by Weir and Blackwell (2001) only had a 25% success rate. Updated versions of the same protocol were used in a few studies (Goldmann and Weir 2012, 2018; Goldmann et al. 2013). However, these protocols are time-consuming, require more than one thallus per extraction, and result in only limited success of extraction and PCR amplification (Sundberg et al. 2018a). Haelewaters et al. (2015) evaluated four DNA extraction protocols and different pre-treatments with mixed success. They had difficulties extracting DNA of Laboulbenia species, with success rates between 0 and 20%. DNA extraction and amplification of Laboulbenia species has been suggested to be particularly difficult, because many species in the genus have melanized cells, which is known to interfere with DNA polymerase during PCR (Eckhart et al. 2000; Gibson 2012; Haelewaters et al. 2015; Sundberg et al. 2018a). Several PCR inhibitor removal kits are available to remove molecules such as melanin and phenols (Hu et al. 2015; Vicente et al. 2019). Similar results can be reached by adding bovine serum albumin (BSA) to the PCR mixture (Giambernardi et al. 1998), but this has not yet been tested for Laboulbenia species nor Laboulbeniales in general.

Sundberg et al. (2018a) developed a new DNA extraction protocol, in which a manual press system was used to disrupt the tough cell walls of Laboulbeniales. Using one thallus per DNA extraction, they successfully obtained 156 sequences: 20 nrSSU, 56 ITS, 59 nrLSU, and 21 mtSSU. Their protocol does not involve any other treatment of the thalli, which is a major advantage compared to the other described methods. However, drawbacks for their protocol include the need for custom-made components and the fact that DNA extractions are fully depleted during PCR amplification and thus cannot be stored.

Haelewaters et al. (2018a) published another method to extract DNA of Laboulbeniales. Using the REPLI-g Single Cell Kit (Qiagen), a sufficient amount of DNA can be obtained from a single thallus. This protocol is different from the previous methods, as it involves a whole-genome amplification (WGA) step. Due to the WGA, there is a higher risk to amplify contaminants. Yet, it has been successfully used in many other studies by Haelewaters and colleagues (Haelewaters et al. 2018b, 2019a, 2019b, 2022; Walker et al. 2018; Haelewaters and Pfister 2019; Haelewaters and De Kesel 2020; Liu et al. 2020; Van Caenegem et al. 2023a, 2023b). Haelewaters et al. (2019b) used a modified protocol, in which they halved the use of every component, to save products and reduce costs per extraction. There are doubts about the effectiveness of this modified protocol (D. Haelewaters and P. Mironova, pers. comm.), but no formal tests have been performed to evaluate the significance of these doubts.

Currently, the nuclear small and large subunit ribosomal RNA (nrSSU and nrLSU), the internal transcribed spacer region (ITS), the minichromosome maintenance complex component 7 protein-coding gene (MCM7), the translation elongation factor 1α protein-coding gene (TEF1), and the mitochondrial small subunit rRNA (mtSSU) have been sequenced for several species of Laboulbeniales (Goldmann and Weir 2012; Goldmann et al. 2013; Haelewaters et al. 2015, 2018a, 2019b, 2022; Sundberg et al. 2018a, 2018b; Liu et al. 2020; Van Caenegem et al. 2023a, 2023b). General fungal primers designed by White et al. (1990) have mainly been used to amplify regions of ribosomal RNA (nrSSU, ITS, nrLSU). Haelewaters et al. (2015) developed a Laboulbeniomycetes-specific nrSSU primer pair. Additionally, recent research reported low amplification of the ITS region using general fungal primers, which resulted in the design of a Hesperomyces-specific ITS primer pair (ITShespL and ITShespR) and the Laboulbeniomycetes-specific LabITS1 forward primer (Haelewaters et al. 2018a, 2019b).

Only 10 nrSSU sequences of 9 species, 12 ITS sequences of 4 species, and 34 nrLSU sequences of 12 species of Laboulbenia are available in GenBank. Given that almost 700 species are described (Haelewaters et al. 2024), there is a huge discrepancy between described and sequenced species of Laboulbenia. The main goal of this study was to generate DNA sequences of Laboulbenia species for future molecular studies, given their paucity in public sequence databases. We had the opportunity to test questions regarding DNA extraction protocols, primer pairs, and PCR protocols. During the quest to consistently extract and sequence DNA of Laboulbenia, we (1) researched how the preservation methods correlate with DNA extraction success, (2) explored the boundaries of the REPLI-g Single Cell Kit by halving the amount of reagents per extraction, (3) tested whether the addition of BSA results in more successful DNA extractions, and (4) identified the usefulness of different primer pairs and PCR protocols for multiple loci to successfully generate high-quality DNA sequences of Laboulbenia species.

Methods

Collection and processing of beetles

Beetles (Coleoptera) were collected using different entomological methods (light traps, pitfall traps, mouth-operated aspirator, and hand collection) by the authors and by entomologists who sent specimens for study of their Laboulbeniales. Specimens were collected in 70% to 99% ethanol. All specimens were transferred to 99% ethanol upon arrival in the lab at Ghent University. Beetles were screened for the presence of Laboulbeniales using an RZB-PL 65.500 stereoscope (Novex, Arnhem, The Netherlands). Infected specimens were identified by the authors (using Muilwijk et al. 2015) or their collectors. Other host specimens were sent to Dr. Menno Schilthuizen (Taxon Expeditions, Leiden, The Netherlands) for identification and subsequent vouchering. Specimens are stored in the collection of Taxon Expeditions (TXEX, Leiden, The Netherlands) or the entomology collections of the Royal Belgian Institute of Natural Sciences (KBIN, Brussels, Belgium).

Morphological study of Laboulbeniales

Laboulbeniales microfungi were mounted in permanent slides as described by Liu et al. (2020). A 1:1 mixture of Hoyer’s medium and glycerin was used, as pure Hoyer’s medium dries too quickly. A small droplet of water was placed on a microscope slide, on which a 22 × 22 mm coverslip was put. The purpose of this was to ensure that the 22 × 22 mm coverslip was somewhat stuck to the microscope slide during further manipulation and thus could not move unexpectedly or fall. A droplet of the Hoyer’s/glycerin mixture was placed off-center on the coverslip. A hypodermic needle was used to remove Laboulbeniales thalli from the host and place them in the droplet. Thalli were arranged in one row or multiple rows in the middle of the coverslip. A smaller 18 × 18 mm coverslip with a drop of Amann’s medium was flipped upside down (drop of Amann’s medium facing down) and positioned over the thalli by gently lowering it with a dissecting pin. Next, the corners of the 18 × 18 mm coverslip were sealed to the larger coverslip with nail polish. Solakryl BMX (Ento Sphinx, Pardubice, Czech Republic) was applied to the microscope slide, and the coverslip assembly with the smaller coverslip facing downwards was slowly lowered and gently placed sideways on the microscope slide. Our permanent slides are each composed of a 22 × 22 mm coverslip on top of an 18 × 18 mm coverslip, with the thalli in between those two coverslips, and the microscope slide at the bottom.

Mounted thalli were viewed at 100–400 × magnification under an Olympus BH-2 microscope (Olympus, Center Valley, PA). Thalli were identified based on relevant literature (Thaxter 1896; Majewski 1994; De Kesel et al. 2020; Haelewaters and De Kesel 2020; Santamaria and Pedersen 2021) and supplementary papers (Santamaria et al. 1991). Permanent slides of Laboulbeniales are deposited in the Herbarium Universitatis Gandavensis (GENT).

DNA extraction, PCR amplification, and sequencing

DNA extractions were done using the REPLI-g Single Cell Kit (Qiagen, Stanford, CA). The initial steps are described in Fig. 1. Molecular work was done at the Centre for Molecular Phylogeny and Evolution (CeMoFE) at the Ledeganck Campus of Ghent University. All steps were performed wearing disposable latex gloves. To avoid contamination, hypodermic needles for micromanipulation of thalli were thoroughly cleaned with 70% ethanol and bleach before and after every prepared extraction. Thalli of Laboulbeniales were removed from their host under a dissecting microscope, using a needle inserted onto a glass syringe for holdfast. The tip of the needle was submerged in glycerin to prevent thalli from flying away during transfer. The removed thalli were placed in a droplet of glycerin on a microscope slide. Appendages were often cut off to avoid downstream contamination, as they may harbor debris including fungal propagules. The thalli were either cut into multiple smaller pieces using the sharp tip of the hypodermic needle (sensu Haelewaters et al. 2018a), crushed by pressing the tip of the hypodermic needle onto the thalli (sensu Weir and Blackwell 2001; Sundberg et al. 2018a), or a combination of both techniques was used. These pieces were then placed in a 0.2 ml PCR tube with 4 µl of phosphate-buffered saline (PBS). Next, 3 µl of prepared D2 buffer was pipetted against the inner wall of the PCR tube to prevent accidental removal of thallus fragments, and the tube was shortly centrifuged. The tube was then incubated at 65 °C for 30 min. After incubation, the tube was centrifuged for 20 min at 4000 RPM, and 3 µl of STOP solution was added. Again, to prevent accidental removal of thallus fragments, the STOP solution was pipetted against the inner wall of the PCR tube, followed by a brief centrifugation step. From here on, two different protocols were followed:

• 1. The normal REPLI-g protocol: in the tube, 29 µl Reaction Buffer; 9 µl ddH₂O; and 2 µl REPLI-g sc DNA Polymerase was added, as indicated in the manufacturer’s instructions (Qiagen).

• 2. The halved REPLI-g protocol: in the tube, 14.5 µl Reaction Buffer; 4.5 µl ddH₂O; and 1 µl REPLI-g sc DNA Polymerase was added.

Fig. 1

Illustration of the initial steps to perform DNA extractions using the REPLI-g Single Cell kit (Qiagen). A Thalli of Laboulbenia galeritae attached to the pronotum of Galerita bicolor. B The thalli are removed from their host using a hypodermic needle. C The removed thalli are placed in a droplet of glycerin. D Thalli are cut in multiple pieces using a surgical blade (left) or crushed (right) using the sharp tip of a hypodermic needle. E The pieces are gathered onto the tip of the needle. F The needle is carefully moved inside the PCR tube, while keeping a clear vision through the dissecting microscope. Make sure that there is still a bit of glycerin mixed with the pieces of thalli. This will prevent the loss of thalli due to static electricity during transfer. G The tip of the needle is inserted into the PBS buffer. This will result in the pieces of thalli releasing from the tip almost immediately. Eventually, stirring can help to make sure the pieces come off the tip of the needle. H It is recommended to visually check if the pieces of thalli are present in the PCR tube. Arrows indicate the pieces of thalli of L. galeritae

We also made 32 extractions that we supplied with 21.75 µl Reaction Buffer; 6.75 µl ddH₂O; and 1.5 µl REPLI-g sc DNA Polymerase (a total of 30 µl). Initially, we wanted to test whether the probability to successfully extract DNA differed significantly between these three protocols (normal protocol, halved protocol, and 30 µl protocol). After the first tests, we already observed that this probability did not differ significantly, and we decided to simplify and streamline our experimental setup by focusing on the two extremes (comparing the normal and the halve protocol).

To test the effect of BSA during the whole genome amplification step of the REPLI-g Single Cell Kit, we arbitrarily added 5 µl BSA (20 mg/ml, stock concentration).

After the addition of all reagents, the samples were incubated at 30 °C for 8 h. During this incubation step, whole-genome amplification (WGA) took place: the whole genome DNA in the tubes was amplified using Multiple Displacement Amplification (MDA) (see Discussion). After the WGA, the polymerase was inactivated at 65 °C for 3 min. DNA extractions were stored at -20 °C. DNA quantification was done using the Qubit 2.0 fluorescence spectrometer (Thermo Fisher Scientific, Waltham, MA) and measurements of the A260/A280 and A260/230 absorbance ratios were taken using a NanoDrop 2000 (Thermo Fisher Scientific).

The nrSSU, ITS, and nrLSU were amplified. Additionally, we attempted to amplify MCM7, TEF1, and mtSSU. All primer pairs used are given in Table 1. PCR reactions (25 µl total) consisted of 13.3 µl of RedExtract Taq polymerase (Sigma-Aldrich), 2.5 µl of each 10 µM primer, 5.45 µl of ddH₂O, and 1 µl of DNA extract. Before pipetting 1 µl of DNA extract, the PCR tube was vortexed briefly. PCR conditions are listed in Table 2. PCR products were stored at -20 °C. We also attempted to amplify additional nrLSU, ITS, and TEF1 sequences of older, preserved extractions, which were made during former studies (Haelewaters 2018; De Weggheleire 2019; Haelewaters et al. 2019a).

Table 1 Primer pairs used in this study, including their PCR products and references

Table 2 PCR conditions for each targeted locus

Following PCR, gel electrophoresis was performed by loading the PCR products on a Tris–acetate-EDTA (TAE) 1% agarose gel at 130 V for 30 min. The gels were placed in an ethidium bromide solution for 15 min to visualize the PCR products. Purification of PCR products was done using 1.5 µl of Exo-FAP (0.5 µl Exonuclease I, 1 µl FAST Alkaline Phosphatase) (Thermo Fisher Scientific) per 10 µl of PCR product, at 37 °C for 15 min, followed by deactivation at 85 °C for 15 min. Purified PCR products were sequenced using an automated ABI 3730XL capillary sequencer at Macrogen (Amsterdam, The Netherlands). Sequencing primers were the same as the primers used to amplify the region of interest. Forward and reverse sequence reads were assembled and edited in Sequencher version 5.4.6 (Gene Codes Corporation, Ann Arbor, MI). Newly generated sequences were submitted to NCBI GenBank.

To assess the identity of the newly generated sequences, we constructed four alignments (nrSSU, ITS nrLSU, and TEF1) of these sequences, supplemented with a broad range of Laboulbeniomycetes sequences available on NCBI Genbank. We aligned nrSSU, nrLSU, and TEF1 sequences by locus with the G-INS-i strategy and ITS sequences with the E-INS-i strategy using the online version 7 of MAFFT (Kuraku et al. 2013; Katoh et al. 2019). Models for nucleotide substitution were selected for each partition with ModelFinder (Kalyaanamoorthy et al. 2017) according to the corrected Akaike Information Criterion (AICc). Maximum likelihood (ML) was inferred using IQ-TREE (Nguyen et al. 2015) under partitioned models (Chernomor et al. 2016). Ultrafast bootstrapping was performed with 1000 replicates (Hoang et al. 2018). Alignments and resulting phylogenetic trees are available on GitHub: https://github.com/dannyhaelewaters/teamlaboul/tree/main/molecular_laboulbenia_paper.

Statistical analyses

To test the difference in success/fail ratio between methods of preservation and whether the probability to successfully extract DNA differs between the halved and the normal REPLI-g protocol, we used generalized linear mixed models as implemented in the lme4 package in R (Bates et al. 2015; R Core Team 2021). We tested the assumptions for these models using the DHARMa package (Hartig 2022). Three methods of preservations were defined: ‘doubtful’ (specimens that were collected and preserved in 70% ethanol or dried and pinned for a prolonged time); ‘uncertain’ (specimens for which the preservation method was unknown); and ‘good’ (specimens that were collected and preserved in > 95% ethanol). The ‘uncertain’ group was created to include specimens from which the preservation method was unknown, to prevent assigning a specimen to a wrong group (to ‘doubtful’ or ‘good’). To compare the means of DNA concentration between the halved and the normal REPLI-g protocol and whether the addition of BSA increases the probability to successfully extract DNA, we used linear mixed models as implemented in the lme4 package in R (Bates et al. 2015; R Core Team 2021). Graphical representation of data was made using the ggplot2 package (Wickham 2016).

We included the species of Laboulbenia as a random effect, because success of DNA extraction might be correlated with species, as each species has a different degree of melanization. It is important to note that most species were only represented once or only a few times, so there was no balanced design. We assigned an extraction as ‘successful’ when there was at least a clear single band of the nrSSU amplicon on the gel or when the ITS or nrLSU sequence matched with Laboulbenia species. We used the emmeans package to obtain the Estimated Marginal Means (EMM) for each group and to compare the means between groups (Lenth et al. 2024). R code, output of the analyses, and additional exploratory figures can be found on GitHub: https://github.com/dannyhaelewaters/teamlaboul/tree/main/molecular_laboulbenia_paper.

Results

Comparison of preservation methods

A total of 196 extractions were made, of which 111 were successful. The probability of successfully extracting DNA of thalli from ‘doubtful’ specimens was significantly lower (π = 0.167, 95% confidence interval (CI): 0.0669–0.360) compared to ‘good’ specimens (π = 0.824, 95% CI: 0.6581–0.919) (Fig. 2A). There was no obvious pattern visible between DNA extraction success and the time (both in months and years) between collection and DNA extraction (Fig. 2B).

Fig. 2

Results of comparison of different preservation techniques and protocols to improve efficacy of successfully extracting DNA from thalli of Laboulbenia. A Bar plots indicating the number of extractions in each preservation group, showing successful (in blue) and unsuccessful (in red) DNA extractions. B Bar plots for each preservation method separately, indicating the percentage of successful (in blue) and unsuccessful (in red) DNA extractions in relation to the period (in years) between the collection date and the date on which the DNA extraction was performed. C Bar plots of successful (in blue) and unsuccessful (in red) DNA extractions for each combination of BSA addition (no or yes) and REPLI-g protocol (halved or normal). D Box plots showing the variation in DNA concentration of extractions made with the normal and halved REPLI-g protocol, with indication of the number of samples per group

Comparison of extraction protocols and DNA concentration

There was no significant difference in probability of ‘success’ between protocols, except between the ‘No BSA and halved’ and ‘BSA and halved’ groups (Fig. 2C). The EMM of the probabilities and their CIs are given in Table 3. There was a significant difference in DNA concentration between the halved and normal REPLI-g protocols (p < 0.001) (Fig. 2D). The DNA concentration was significantly higher with the normal REPLI-g protocol (29 µl Reaction Buffer, 9 µl ddH₂O, and 2 µl REPLI-g sc DNA Polymerase added, totaling 40 µl) (EMM = 63.4 µg/ml, 95% CI: 56.1–70.7) compared to the halved REPLI-g protocol (only 20 µl mixture added) (EMM = 37.2 µg/ml, 95% CI: 30.9–43.5). The A260/A280 and A260/A230 ratios of 45 DNA extractions (33 undiluted and 12 1/10 diluted) were measured. Only one of the undiluted DNA extractions had an A260/A280 ratio within the accepted range of ‘pure’ DNA (1.7–2.0). For the diluted DNA extractions, 10 extractions had an A260/A280 ratio of around 1.8, and 4 of these were also within the range of the ideal A260/A230 ratio (1.9–2.2).

Table 3 The Estimated Marginal Mean probability of ‘success’ for each DNA extraction protocol, and its 95% confidence interval (CI)

Evaluation of primer pairs, PCR protocols, and sequence success

Of the 111 successful extractions, 16 were contaminated and we could only generate sequences of the nrSSU region with the Laboulbeniales-specific NLS1/NSL2 primers. The nrSSU sequences of these contaminated extractions were of good quality. The contaminants were identified using ITS or nrLSU sequences and are shown in Table 4. From the successful extractions, we generated 104 nrSSU, 64 ITS, 92 nrLSU, 1 MCM7, and 16 TEF1 sequences (Tables 5 and 6). We also generated 1 nrSSU, 1 ITS, 4 nrLSU, and 5 TEF1 sequences of preserved extractions from former studies (Haelewaters 2018; De Weggheleire 2019; Haelewaters et al. 2019a). All generated sequences were submitted to NCBI Genbank and their accession numbers are presented in Table 6.

Table 4 Contaminants in the DNA extractions, with indications of the phylum, order, family, and the number of extractions that were contaminated with these species

Table 5 Primer pairs, their PCR products, and the number of sequences generated during this study using each of those primer combinations

Table 6 Sequences generated in this study, with an overview of the Laboulbenia species, its host species, country of record, and the loci with their accession numbers. Asterisks indicate isolates that were already made in former studies and for which additional sequences were generated here

The ‘CombSIL’ PCR protocol (acronym for Combination of nrSSU, ITS, nrLSU) is a combination of the PCR protocols for nrSSU, ITS, and nrLSU. These separate protocols all use the same annealing temperatures, and similar timings for each step. By combining them into one protocol, we can use the same PCR machine to amplify these different regions at the same time. When using the old protocol to amplify TEF1 (Table 2: TEF1), if bands were visible, there were often multiple bands. When increasing the annealing temperature, clear single bands were observed on the gel after staining (Table 2: TEF1 New).

Discussion

Comparison of preservation methods

The preservation method had a significant effect on the probability to successfully extract DNA of Laboulbeniales. Thalli stored in a doubtful way (in 70% ethanol or dried and pinned for a prolonged time) had a significantly lower probability of successful DNA extraction than thalli collected and stored in > 95% ethanol. This was already reported in a few studies on Laboulbeniales (Weir and Blackwell 2001; Haelewaters et al. 2015, 2019a). Anecdotally, we observed lower extraction success for thalli that were stored for more than 4–6 months (between collection and DNA extraction) in 70% ethanol. A more detailed and standardized experiment should be conducted to study the effect of different preservations methods on DNA extractions success of Laboulbeniales, including short-term versus longer-term preservation on 70% ethanol, 96% ethanol (expensive molecular grade and cheap denatured bio-ethanol), RNAlater, CTAB, isopropanol, and on -20 °C.

The REPLI-g Kit utilizes a Multiple Displacement Amplification (MDA) to amplify DNA during the WGA step, which involves random hexamer primers and phi29 polymerase (Long et al. 2020). MDA makes use of primers that randomly link to multiple sites of the DNA template and thus no target-specific primers are needed. Disadvantages of MDA include incomplete coverage and over-representation of certain fragments in the resulting DNA extract by chance (e.g., multi-copy regions). If DNA is fragmented, amplicons may be short of even absent, leading to incomplete amplification of the whole-genomic DNA. Subsequent PCR of a specific region is likely to fail due to fragmentation in the primer annealing sites or in the target amplicon. DNA fragmentation is expected in thalli when hosts were stored dried or in 70% ethanol for a prolonged time (> 4 months), as observed in other organisms (Bruns et al. 1990; Kates et al. 2021). For successful molecular work on Laboulbeniales, it is essential to perform the first steps of DNA extraction protocols immediately after collection, or to directly transfer of host specimens to > 95% ethanol, according to our current knowledge.

While this result was expected, it is crucial to emphasize the value of collections made by collaborators and entomologists as a critical resource for Laboulbeniales research (Haelewaters et al. 2015, 2021). Unfortunately, these collections are often inadequately preserved in 70% ethanol or dried and pinned for extended periods, as it is standard procedure in entomological research. For instance, in a 2022 collection of carabid beetles, only 12 out of 41 DNA extractions were successful, likely due to preservation in 70% ethanol for 6–12 months, even though the specimens were collected and processed within the past year (Fig. 2B). For other organisms it is known that their DNA will degrade in 70% ethanol after three months (Flournoy et al. 1996) to one year (Nagy 2010).

For the purpose of molecular work involving Sanger sequencing, we encourage collectors and collaborators to store their collections in > 95% ethanol, with a single ethanol refreshment (1 to 4 weeks after collection) to maintain the required concentration, as ethanol can extract water from host specimens and Laboulbeniales (Nagy 2010; Marquina et al. 2021). Taking proactive measures and collaborating with potential partners can maximize the utility of future collections for various entomological and mycological research purposes. Laboulbeniologists should seize these opportunities to investigate diverse aspects of these understudied insect-associated fungi, encompassing alpha taxonomy, ecology, evolution, and molecular research.

However, preliminary data suggest that preservation on > 95% ethanol and long-term preservation of DNA extracts (post-MDA) at -20 °C may not not useful for other applications such as whole genome sequencing (D. Haelewaters and C.A. Quandt, unpubl. data). Additionally, preservation in ethanol destroys the biofilms present on the thallus surfaces of Laboulbeniales, impeding the study of these unknown communities (Lubbers et al. 2022).

Comparison of extraction protocols and DNA concentrations

We had a high probability to successfully extract DNA of Laboulbenia species (0.824 < π < 0.873) if material was preserved correctly, compared to the 0% and 20% reported previously (Haelewaters et al. 2015). There was no difference in the probability of obtaining a successful DNA extraction between the different protocols (halved and normal) and with or without the addition of BSA. There was one exception: this probability is significantly lower for the ‘No BSA and halved’ group compared to the and ‘BSA and halved’ group, which is likely the result of the small sample size of the former group and an artifact of the number of samples used that were stored in a doubtful way. First of all, this means that we can use the halved protocol to make DNA extractions. This way, twice the number of extractions can be made with the same kit. A single DNA extraction using the REPLI-g Single Cell Kit costs between 24 and 30 euros. When halving the amount of resources per extraction, the costs per extraction are also halved. Yet, the extracts using the halved protocol are still more expensive than other widely used extraction protocols (Lickfeldt et al. 2002; Romanelli et al. 2014; Lahuf et al. 2019). The search for time- and cost-efficient and reliable DNA extraction methods for Laboulbeniales continues.

Secondly, the addition of BSA did not result in a significantly higher probability to obtain a successful extraction. There are multiple questions and considerations that arise with this result. It is harder to extract DNA of fungi that contain high concentrations of melanin(-like) molecules. These molecules confer rigidity and protection to the cells (Butler and Day 1998). In Laboulbeniales, melanin has been suggested to be the reason for the low DNA extraction successes (Haelewaters et al. 2015; Sundberg et al. 2018a). Thalli of Laboulbenia contain variable amounts of melanin(-like) molecules in their cell walls, but their concentration is unknown. Giambernardi et al. (1998) found that the addition of more than 0.5 µg melanin to a 25-µl assay results in the inhibition of Taq polymerase. Assuming that a thallus of an average species of Laboulbenia is a cylinder with a length of 350 µm, a diameter of 60 µm, and a density between 0.1 and 1.3 g/cm³ (Bakken and Olsen 1983), results in an estimated thallus weight between 0.099 and 1.29 µg. This would mean that the amount of melanin needed to inhibit PCR (0.5 µg/25 µl) is either higher than the lowest estimated weight of a thallus (0.099 µg) or more than a third of the highest estimated weight (1.29 µg), which seems unlikely. In other words, we think that the melanin content in the thallus cells is insufficient to hinder WGA or PCR. Consequently, the addition of BSA would have little impact on the probability of successful DNA extraction. It is possible, though, that other molecules such as phenols or proteins, potentially in combination with melanin(-like) molecules, could impede WGA or PCR.

The difference in DNA concentration between the halved and normal REPLI-g protocol was expected, as reagents used to amplify the genome are halved, only half of the amount of DNA can be amplified. There is still some variation in the DNA concentration within each group, which is probably because the amount of the REPLI-g mix pipetted into the PCR tubes slightly differed between samples due to small pipetting errors. The number of thalli used did not have a significant effect on the obtained DNA concentrations. This means that using one thallus is sufficient to make a successful DNA extraction using the REPLI-g Single Cell Kit. However, some species are very small, and using multiple thalli ensures that at least a few pieces of thalli will end up in the PBS solution when transferring them into the PCR tube.

We measured the A260/A280 and A260/A230 ratios to assess the purity of the extracted and amplified DNA. The measurements of all except one undiluted DNA extractions were not within the ideal absorbance ratio range for ‘pure’ DNA of 1.7–2.0 (for A260/A280) and 1.9–2.2 (for A260/A230). This is not surprising, as BSA was added to all these extractions, and other contaminants (other proteins, melanin, and phenols) are potentially present. Most diluted DNA extractions were, based on these absorbance ratios, more purified. As the DNA is diluted, the amount of contaminants is also diluted. There is an interest to sequence whole genomes of Laboulbeniomycetes to study the evolution of fungal genomes, population genetics, speciation patterns, parasite–host interactions, and the loss of hyphae for which good quality and ‘pure’ DNA extractions are needed. Haelewaters et al. (2020) sequenced the first Laboulbeniomycetes genome, of Herpomyces periplanetae. To further purify DNA extractions of Laboulbeniales, PCR inhibitor removal kits like the OneStep PCR Inhibitor Removal Kit (Zymo Research, Irvine, CA) (Hu et al. 2015; Vicente et al. 2019; Lubbers 2021) and SPRI bead cleaning (Beckman Coulter, Brea, CA) (B. Young and W. Van Caenegem, unpubl. data) can be used. This approach seems promising for future applications.

Evaluation of primer pairs, PCR protocols, and sequence success

Generating sequences of Laboulbenia species has generally been regarded as difficult (Haelewaters et al. 2015; Sundberg et al. 2018a). Here we generated the highest number of sequences of Laboulbenia in a single study, using the REPLI-g Single Cell Kit.

The NSL1/NSL2 primer pair works very well for species in the genus Laboulbenia. This pair specifically amplifies Laboulbeniomycetes DNA, thus it can also be used even when DNA extractions are contaminated (Haelewaters et al. 2015). While the nrSSU marker is generally not useful for species delimitation as it is a very conserved region (but see Paloi et al. 2022 for a discussion on group I introns), it can be used to distinguish higher taxa (genera, family, order).

The general fungal primers for ITS do not sufficiently amplify the whole ITS region of Laboulbenia species. Although we did have some success using ITS1f/ITS4 and ITS5/ITS4, there is a considerable difference with the amplification success of nrSSU and nrLSU region. At the end of the trial, we used the ITS3/ITS4 pair to amplify the ITS2 region. Surprisingly, this generated positive results for most taxa. It is likely that there is a primer mismatch at the primer annealing sites of ITS1f and ITS5, and, for some species, of ITS4 (e.g., Laboulbenia cristata, L. galeritae, and L. gyrinicola), as previously suggested (Liu et al. 2020). Bellemain et al. (2010) reported a bias for Ascomycota in eDNA studies when using ITS2, ITS3, and ITS4, while the forward primers ITS1, ITS1f, and ITS5 show a bias towards Basidiomycota. This might explain why we successfully amplified the ITS2 region, but not the ITS1 nor the whole ITS region for most species.

For the nrLSU region, the primer pair NL1/NL4 works well. The amplicon is around 300 base pairs shorter than the one amplified by LR0R/LR5. However, LR0R/LR5 did not work for most taxa of Laboulbenia and if it worked, multiple faint bands were often observed on the gel after visualization. Similar observations were made with extractions of Gloeandromyces and Hesperomyces (W. Van Caenegem and D. Haelewaters, unpubl. data). Multiple bands on the gel might indicate that the annealing temperature used (50 °C; Table 2) was suboptimal for this primer pair (Rychlik et al. 1990). Increasing the annealing temperature might overcome this problem, but there is also likely a primer mismatch in most species of Laboulbenia for one or both primers.

Although NL1/NL4 works very well for Laboulbenia species, it also amplifies the DNA of contaminants. Species of two genera (Malassezia and Priceomyces) were observed multiple times. Malassezia species (Ustilaginomycotina, Malasseziales) are basidiomycetous yeasts that live on the skin of humans and other vertebrates (Theelen et al. 2018). These species are probably present in our DNA extractions due to the many handlings needed during the extraction process (Fig. 1). Priceomyces vitoshaensis (Saccharomycotina, Saccharomycetales) is an ascomycetous yeast described from the carabid beetle Pterostichus melas (Crous et al. 2016). There is not much known about the ecology of this species, but other species of Priceomyces are found in eDNA studies of soil, on beetles, in beetle guts, and in insect frass (Kurtzman 2011; Groenewald et al. 2018; Kudo et al. 2019). The micromanipulation needed to remove thalli from the host, may increase contamination.

We could not successfully amplify the three other markers (MCM7, TEF1, and mtSSU). Some sequences of MCM7 and TEF1 were generated for Laboulbenia species, but we decided to not explore these markers and their primers further to save time, money, and resources. Recently, MCM7 was introduced as a secondary marker in Hesperomyces (Haelewaters et al. 2022; Van Caenegem et al. 2023b) and TEF1 is easily amplified for Gloeandromyces species (Liu et al. 2020; Van Caenegem et al. 2023a). Amplifying these regions is more challenging for species of Laboulbenia compared to our experiences with Gloeandromyces, Hesperomyces, and Nycteromyces (W. Van Caenegem and D. Haelewaters, unpubl. data). The first amplification trial of mtSSU yielded no sequences and we therefore made no further use of the MS1/MS2 primer pair. Sundberg et al. (2018b, 2018a) generated mtSSU sequences of Laboulbeniales and reported this region to be “the easiest to amplify and as well as sequence.” However, they used the primer pair mrSSU1/mrSSU3R and they mostly sequenced Coreomyces species, which in part may explain the difference in amplification success.

To overcome these challenges of contaminants, primer mismatches, and low amplification success of several loci, we propose to design Laboulbeniales- and more specifically Laboulbenia-specific primers, especially for the ITS and nrLSU regions, similar to the Laboulbeniomycetes-specific nrSSU primers (NSL1/NSL2) (Haelewaters et al. 2015) and the Hesperomyces-specific ITS primers (ITSHespL/ITSHespR) (Haelewaters et al. 2018a). These developments will not only help to overcome these aforementioned challenges within the genus Laboulbenia, but they will also contribute to the study of Laboulbeniales in general, as most species remain unsequenced, resulting in the underestimation of the true diversity of these understudied microfungi as well as a poor understanding of evolutionary relationships due to under-sampling.

Conclusions

Using the REPLI-g Single Cell Kit, we made 111 successful DNA extractions from Laboulbenia species. The preservation method had a major effect on the success of DNA extraction; specimens stored for extended periods in 70% ethanol or dried and pinned were found unsuitable for molecular work. For future research, we encourage entomologists, collaborators, and collectors to preserve infected host specimens in > 95% ethanol. Our findings revealed no significant differences in DNA extraction protocols, indicating that utilizing half the recommended amount is sufficient to successfully extract DNA—saving costs. Further, the addition of BSA did not significantly impact the probability of obtaining successful DNA extractions, suggesting that the melanin content in Laboulbenia species is negligible. We generated 104 nrSSU, 64 ITS, 92 nrLSU, 1 MCM7, and 16 TEF1 sequences. The limited success in amplifying protein-coding genes can likely be attributed to primer mismatches. Furthermore, a disparity in amplification success between ITS and the nrSSU and nrLSU regions was observed, likely stemming from primer mismatches. It is imperative to develop Laboulbeniales-specific ITS and nrLSU primers to tackle contaminations and improve amplification efficiency. These new insights do not only apply to the genus Laboulbenia; the vast majority Laboulbeniales species remain unsequenced. The study of Laboulbeniales will benefit from the molecular developments reported here. And while we keep critically evaluating and improving our methods, we hope that other research groups will be inspired to start molecular work with these peculiar microfungi.