Comparison of methods for DNA extraction from Candida albicans

Department of Medical Biochemistry and Microbiology Uppsala University

Comparison of methods for DNA extraction from Candida albicans

Ashraf Dadgar

2006

Department of Clinical Microbiology, Uppsala University Hospital, SE- 751 85 Uppsala Sweden

Supervisor: Åsa Innings and Björn Herrmann

ABSTRACT

Invasive Candida infection is an increasing cause of morbidity and mortality in the immunocompromised patient. Molecular diagnosis based on genomic amplification methods, such as real time PCR, has been reported as an alternative to conventional culture for early detection of invasive candidiasis. The template DNA extraction step has been the major limitation in most reported nucleic acid based assays, due to problems in breaking fungal cell walls and incomplete purification in PCR inhibitor substances.

The aim of this study was to compare enzymatic cell wall disruption using recombinant lyticase with mechanical disruption using glass beads. The QIAamp tissue kit was compared with two automated DNA extraction robots, the BioRobot M48 and NucliSens easyMAG, to determine their sensitivity, reliability and duration for DNA release of C.

albicans. Mechanical cell wall disruption shortened and facilitated the extraction procedure, but the quantity of released DNA was significantly lower than when

enzymatic cell wall disruption was used. Use of robots did not significantly shorten the DNA extraction time, compared with manual DNA extraction. However the NucliSens easyMAG resulted in a higher yield of target DNA compared to the BioRobot M48 and the manual QIAamp tissue kit.

KEYWORDS: Candida albicans, candidemia, DNA extraction, cell wall disruption, real

time PCR

SAMMANFATTNING

Invasiva svampinfektioner är ett stort problem hos patienter med dåligt immunförsvar.

Förekomst av invasiva svampinfektioner har ökat under senare år och medför hög dödlighet. En svampinfektion som inte snabbt diagnostiseras och behandlas kan bli livshotande om patientens kondition är dålig. Candida albicans är den vanligaste orsaken till invasiva svampinfektioner. Med traditionell svampidentifiering kan det ta dagar till veckor att isolera och artbestämma svampen. En snabbare metod att detektera Candida är att använda sig av molekylärbiologiska metoder som påvisar svampens arvsmassa, DNA. Svampar har en cellvägg som är svår att bryta ner och därför är DNA extraktionssteget ett av de mest rapporterade problemen vid DNA svampdiagnostik.

Syftet med denna studie var att jämföra enzymatisk och mekanisk cellväggsnedbrytning av C. albicans med hjälp av enzymet lyticase respektive glaskulor. Vi jämförde också en manuell metod med två automatiska robotar för att bestämma deras känslighet,

tillförlitlighet och tidsåtgång för DNA-extraktion från C. albicans. De slutsatser som nåtts är att den enzymatiska cellväggsnedbrytningen var känsligare men betydligt mer tidskrävande än den mekaniska cellväggsnedbrytningen. Denna studie visade även att en

av de automatiska systemen extraherade signifikant mer DNA än den manuella metoden.

INTRODUCTION

Invasive fungal infection has become a major cause of morbidity and mortality in immunocompromised patients, for example, neuropenic patients with hematological malignancies and recipients of allogeneic bone marrow transplants [1]. The most

common fungi causing disease are Candida species [2]. The genus Candida is comprised of more than 200 species, of which approximately a dozen have been associated with human infection. Of these the most important cause of disease is Candida albicans [3].

There is an increasing incidence of bloodstream infections caused by Candida species and this genus now ranks as the fourth most common cause of nosocomial bloodstream infections [4]. A prospective epidemiological survey of candidaemia has been performed in central Sweden from January 1998 to December 1999. Out of a total of 191 reported cases, C. albicans was identified in 128 cases (67%), followed by C. glabrata in 30 (15,7%) and C. parapsilosis in 14 (7,3%) [5].

C. albicans is an opportunistic and a commensal organism that is carried by a large proportion of the population on the mucosal surfaces of the gastrointestinal and

urogenital tract without clinical symptoms. C. albicans causes both superficial infections and life threatening systemic candidiasis in immunocompromised hosts, such as AIDS patients, cancer patients and other immunosupressed individuals [6].

Candida virulence factors include the ability to adhere to host tissues, production of tissue damaging secreted enzymes, and morphological changes that may enhance tissue penetration and avoidance of immune surveillance. There is evidence of an

important role of mononuclear phagocytes, as well as protective antibodies and T-helper cells, in primary and acquired resistance to systemic and disseminated candidiasis.

Neutrophils are still considered to be the most important effector cells. Therefore, sustained neutropenia results in a predisposition to disseminated fungal infection [6].

Candida infections are treated with antifungal agents such as azole drugs, mainly fluconazole, expensive drugs with high incidences of side effects. Treatment of

candidiasis patients is further hampered by limited choice of antifungal agents and the

appearance of clinical isolates resistant to azole drugs [7].

Non-albicans Candida (NAC) species cause 35-65% of fall candidaemia. NAC species are emerging as both colonizers and pathogens causing nosocomial fungal bloodstream infections. The most common species are C. tropicalis, C. glabrata, C. krusei and C. parapsilosis, which as a group represent about one-half of all Candida spp. isolated from blood cultures.

Two general problems are associated with the occurrence of NAC. The virulence and pathogenicity of some NAC species, mainly in the immunocompromised host, results in significant mortality. Another issue of concern is the occurrence of resistance to

currently available antifungal drugs [8].

Candida glabrata infections can be mucosal or systemic and are common in immunocompromised persons or those with diabetes mellitus. In contrast to other Candida species, C. glabrata is not dimorphic. As a consequence, it is found as

blastoconidia, both as commensal and as a pathogen. Treatment of C. glabrata infections is difficult due to frequent resistant to many azole antifungal agents, especially

fluconazole [9].

Candida krusei is often seen in leukaemic patients and bone marrow transplant recipients.

It is rare in surgical and intensive care patients and neonates. In general, C. krusei is primarily resistant to fluconazole but sensitive to itraconazole, ketoconazole and amphotericin [8].

Candida tropicalis is the four most commonly isolated NAC species after

C. glabrata , C. parapsilosis and C. krusei. The incidence of infection due to C. tropicalis does not seem to be increasing. C. tropicalis is seen more frequently in cancer patients.

Animal models indicate that the virulence and pathogenicity of C. tropicalis is at least as virulent as C. albicans. C. tropicalis was early found to have lower sensitivity to

ketokonazole and moconazole. Lately also fluconazole and amfotericin B resistance have

been reported [8].

In 1995, a new Candida species, C. dubliniensis, closely related to C. albicans, was identified in cases of oral candidiasis in HIV-infected individuals. Despite the very close phylogenetic relationship between C. albicans and C. dubliniensis and the fact that they share a large number of phenotypic traits, epidemiological and virulence model data indicate that the former is a far more successful pathogen. Resistance to fluconazole has been reported with C. dubliniensis [3].

C. parapsilosis causes 17-50% of fungaemia in children compared with 2,5-12% in adult surgical or intensive care unit populations. C. parapsilosis has the lowest mortality of all NAC, 8%, compared to 21% for C. albicans, 40% for C. glabrata and 30% C. krusei.

C. parapsilosis is sensitive to most available anti mycotic drugs on the market [8].

Candida gulliermondi rarely causes infection in human. No specific risk factors have been described for C. gulliermondi. There are no data on virulence and pathogenicity in comparison with C. albicans or other NAC species, but C. gulliermondi can undoubtedly cause invasive infections in man. C. gulliermondi is sensitive to fluconazole [8].

As candidiasis incidence continue to rise, quick laboratory identification of Candida is becoming increasingly important for a growing population of patients at-risk. Early initiation of antifungal therapy is critical in reducing the high mortality rate in these patients. The traditional reference method for detection of blood infections is blood culture. Significant improvements have been made over the last decades with this

method. For example optimization of culture media attempting to shorten the turnaround time to detection of negative or positive results, and to increase the strength and yield for the assay. Culture detection is often delayed because of slow or absent growth of fungal isolate from clinical specimens. Blood cultures are positive for fewer than 50% of patients with hepatosplenic candidiasis [10].

C. albicans can be identified by germ tube information tests. C. albicans, C. krusei and

C. tropicalis can presumptively be identified by growth on CHROMagar medium, and

other species of Candida can be identified by rapid (in 4 h) enzymatic tests. However each of these procedures requires the organism to be grown on solid medium for at least 24 h, and more often 48 h, before such tests can be performed. The “gold standard” for definitive yeast identification requires further analysis by assimilation and fermentation tests, witch can require up to 28 days to complete. Therefore, a definitive test for the rapid identification of Candida at the species level would be clinically and

epidemiologically important [11].

One molecular method that is faster in generating results is polymerase chain reaction (PCR). A number of PCR assays have been developed for detection and identification of Candida species using rRNA genes or the internal transcribed spacer (ITS) of the ribosomal DNA as target sequences [12]. The real time PCR assay that has been used in this study is able to detect seven Candida species: C. albicans, C.

dubliniensis, C. glabrata, C. guillermondi, C. krusei, C. parapsilosis and C. tropicalis with high sensitivity and specificity using the RPR1 gene as a target. The two most common species, C. albicans and C. glabrata, as well as the fluconazole resistant C.

krusei are identified using species specific TaqMan probes. The remaining species are detected by a broad range TaqMan probe.

Fig 1. Primers and probes for the seven Candida species

A major limitation for the use of molecular methods in Candida detection is the DNA extraction step. One problem is the difficultly associated with breaking cell walls, and extraction of Candida cells from blood require more than half a working day. It is important to find a good solution to this problem. Today’s extraction kits and protocols are, in addition to being time consuming and labor demanding in extraction of DNA, also introducing some problems: (i) they accumulate significant amounts of PCR inhibitor compounds in the extraction steps, (ii) human leukocyte DNA is extracted and may disturb PCR-reaction or give false positive results [15].

The ultimate sensitivity of any PCR assay for the detection of fungal pathogens depends on the efficient lysis of fungal cells in the tissue sample and the purification of DNA that is free of PCR inhibitors. Fungi have cell walls that impede lysis and recovery of nucleic acid [13]. Therefore to release DNA of fungi, thorough lysis steps are necessary. There are different fungal cell lysis methods. The cell walls can be disrupted mechanically with glass beads or enzymatically using recombinant lyticase. Thereafter further DNA

purification is necessary which can be performed using a number of DNA extraction methods. Traditional DNA extraction methods use toxic chemicals such as HLGT (heat lysis-guanidine thiocynate) and PKPC (proteinase K phenol-chloroform) but commercial methods such as QIAamp silica-gel-based membrane or automated extraction with magnetic beads based systems such as BioRobot M48 (Qiagen, www.qiagen.com) or NucliSens easyMAG (bioMérieux, www.biomerieux.com) are more suitable for routine diagnostics [14].

It has been suggested that PCR detection of free template DNA in serum is preferred over the use whole blood for the diagnosis of systematic candidiasis. In contrast to blood samples, DNA in serum is not present in a cellular component and its isolation can therefore be easily achieved. Bougnoux et al. are upholding this idea. They studied candidemia over two periods, in the experimental model of infection in rabbits [15].

Their results suggest that DNA could be released in serum and extracted for PCR assays

to detect candidemia. However, in this experiment a rabbit model used with a fully

alerted immune system competent to retaliate an injection of Candida in the blood

system. A full immune system will lead to release of Candida DNA into serum. A relevant question is if this expected in a clinical case with a patient that has a suppressed immune system. It is therefore doubtful if any conclusions can be drawn from

comparison between serum and whole blood as the most suitable sampling material.

The aim of this study was to reduce the time necessary for fungal DNA extraction significantly and to simplify the working process. We compared enzymatic cell wall disruption using recombinant lyticase with mechanical disruption using glass beads. For the following extraction steps of DNA the QIAamp tissue kit was compared with two automated extraction robots: BioRobot M48 and NucliSens easyMAG, to determine their sensitivity, reliability and duration for DNA release of the most common pathogenic yeast, Candida, a mold with a very complex cell wall. We sought to compare DNA extraction methods by using real time PCR to measure the amount of Candida DNA.

MATERIALS AND METHODS Blood samples

For evaluation of the detection limit, EDTA anticoagulated whole blood samples from healthy volunteers were spiked with C. albicans cells.

DNA extraction from blood samples

Method 1: enzymatic disruption using recombinant lyticase

The DNA extraction of candida was performed essentially as described by Löffler et al [16], using the QIAamp Tissue protocol (Qiagen). Four ml of EDTA-anticoagulated blood samples were mixed with 20 ml red-cell lysis buffer, RCLB (10 mM Tris pH 7,6 ; 5 mM MgCl

; 10 mM NaCl) and incubated at room temperature for 15 min. The mixture was centrifuged at 3600 rpm for 10 min. The supernatant was then discarded and the procedure was repeated. Lysis of leucocytes was performed by incubation with 1 ml RCLB with the addition of 200µg proteinase K (QIAgen) at 65°C? for 45 min.

Spheroplasts were directly generated by incubation of the pellets with 500 µl lyticase

buffer (200 U/ml lyticase L-2524, sigma; 50 mM Tris pH 7,5; 10 mM EDTA; 28 mM ß-

mercaptoethanol) for 30 min at 37°. The spheroplasts were collected by centrifugation at

13000 ×g for 10 min. The pellet was resuspended in 180 µl buffer ATL and 20 µl proteinase K provided in the QIAamp DNA mini kit and incubated at 55° for 15 min.

Finally spheroplast lysis and DNA extraction were accomplished with the QIAamp DNA Tissue protocol and eluted once with 50µl water.

Method 2: mechanical disruption using glass beads

DNA extraction was performed using a modification of the method described above using glass beads for cell disruption. The lysis of leucocytes and the lyticase step were replaced with a vortexing step using different amounts of 0.5 mm glass beads (Sigma, www.sigma.com) and different vortexing times.

Method 3: automated extraction using the BioRobot M48 and NucliSens easyMAG

DNA was extracted as described previously in method one expect after the spheroplasts were collected, the pellet was resuspended in 200 µl PBS. We used MagAttract DNA blood Mini M48 kit (QIAgen) in combination with the BioRobot M48 workstation (200 µl sample input, 100 µl output) and the NucliSens magnetic extraction reagents

(bioMérieux) according to the manufacturer’s instructions in combination with the NucliSens easyMAG (200 µl sample input and 60 µl output).

Real-time PCR

Table 1. Sequences and modifications of the primers and probes used in the multiplex real time PCR .

Primers and Probes

Sequence 5’-3’ modifications

5’, 3’

Target Primers

cand-CR1 (forward) CGGGTGGGAAATTCGGT broad-range

cand-CR5 (reverse) CAATGATCGGTATCGGGT broad-range

gla-CR3 (forward) RGCAACGGCTGGGAAT C. glabrata

krus-CR5 (reverse) TAGTGATCGGTATCGAGTT C. krusei Probes

alb-FAM (reverse) CAGCTTGTAGTAAAGAATTACTCAC 6-FAM, BHQ1 C. albicans

gla-JOE (reverse) TAAAGCCTCACCACGATTTTGACAC JOE, BHQ1 C. glabrata

cand-ROX (reverse) TTCGCATATTgCAcTAAaYaG

ROX, BHQ2 broad-range

krus-Cy5 (reverse) CCAAAGTTGTACAAGCAAGTACCA Cy5, BHQ2 C. krusei

Real time PCR using Taqman probes was carried out in 0.1 ml micro reaction tube using the Rotor-Gene 3000 instrument (Corbett Research, Australia).

10 µl of extracted DNA was amplified in a 50µl reaction mixture with 1X PCR buffer, 2 mM MgCl

, each dNTP at a concentration of 0.2 mM, 0.1 µM of the four primer and four probes in Table 1 and 3 U of TaqDNA polymerase. Thermal cycling conditions consisted of heating at 94°C for 10 minutes, which preceded a two-stage temperature profile of 15 seconds at 95°C and 60 seconds at 58°C for 55 cycles. The recording of the fluorescence at four different wavelengths at 58°C in the Rotor-Gene instrument took approximately 20 seconds. Setting the instrument to 40 seconds at 58°C made an actual 60 seconds stage.

To minimize the risk of contamination, sample preparation and PCR analysis were carried out in separate rooms. Sterile gloves and barrier tips were used for all steps in the protocol. To monitor for contamination, a negative control (sterile water) was used in each amplification.

RESULTS

Comparison of enzymatic and mechanical disruption methods

We compared mechanical cell disruption using glass beads with enzymatic cell wall

disruption using recombinant lyticase. For the following extraction steps of DNA the

QIAamp tissue kit was used. For evaluation of the detection limit, whole blood samples

were spiked with increasing amounts of C. albicans cells. DNA extractions with and

without blood followed by real time PCR were repeated several times and duplicates of

each sample were run in the PCR. Different vortexing times and different amounts of 0.5

mm glass beads were tested. 15 min vortexing and 0.2 g glass beads were the optimal

time and amount, because longer vortexing and larger amounts of glass beads did not

increase the amount of DNA recovered. Table 2 shows the mean Ct-values of duplicate

samples for different amounts of glass beads with different vortexing times.

Table 2. Ct-values in PCR using mechanical cell disruption with comparison to enzymatic cell disruption.

Number of C. albicans cells

Vortexing times Amount of glass beads

Mean Ct- values for mechanical method

Mean cycle difference compared with enzymatic cell lysis 1000

1000 1000 1000 1000 1000

3 min 5 min 15 min 20 min 20 min 30 min

0.2 g 0.5 g 0.2 g 0.2 g 0.5 g 0.2 g

44.2 46.7 39.7 43.1 42.8 41.3

3.0 5.5 4.6 3.9 3.6 2.0

The mechanical method was not sensitive enough for reliable detection of Candida DNA when extracted from blood samples and gave significantly higher Ct-values also when extracted from cell culture. The Ct-values were in average 3.5 cycles higher for mechanical disruption method compared with enzymatic disruption method with the same number of cells (Table 2). The use of mechanical cell disruption method compared with enzymatic cell disruption method did shorten the extraction procedure, but the quantity of released DNA was significantly lower.

Comparison of the QIAamp tissue kit with BioRobot M48 and NucliSens easyMAG The QIAamp tissue kit was compared with two automated robots: BioRobot M48 and NucliSens easyMAG, to determine their sensitivity, reliability and duration for DNA release of Candida. DNA extractions with and without blood were repeated two times and triplicates of each sample were run in the PCR. For evaluation of the difference in DNA yield and quality, samples were spiked with 3,000 and 30,000 of C. albicans cells that were divided into three equal parts after cell wall lysis using lyticase. For the

following extraction steps of DNA the QIAamp tissue kit, BioRobot M48 and NucliSens easyMAG were used. Table 3 shows the sensitivity and the duration of the DNA isolation by various isolation methods. Several observations can be made.

(i) The NucliSens easyMAG robot proved to be more sensitive than BioRobot M48 and

QIAamp tissue kit, both with and without blood. (ii) The BioRobot M48 proved to be at

least as sensitive as the QIAamp tissue kit. (iii) The use of automated robots did not significantly shorten duration of the DNA isolation procedure.

Table 3. Quantitative analysis of different extraction methods

Method Mean Ct-values Std. Dev Duration for 6

samples QIAmp with blood

BioRobot M48 with blood NucliSens easyMAG with blood QIAamp without blood

BioRobot M48 without Blood NucliSens easyMAG without blood

37.18 36.48 35.31 36.71 37.23 35.04

0.24 0.37 0.25 0.45 0.78 0.61

60 min 50 min 37 min 60 min 50 min 37 min

DISCUSSION

Many different PCR assays have been developed to improve the diagnosis of Candida infection [16]. In most reported PCR assays the template DNA extraction step has been explained as the major limitation due to problems in breaking fungal cell walls, and contamination of PCR inhibitors substances. An acceptable DNA extraction method for clinical material must be able to recover minute amounts of DNA in a rapid and efficient manner. In a study comparing five commercially available extraction kits and an in-house method, sensitivity varied from 1 to 1,000 fungal cells/ml of blood [16, 17 ]. Most

published DNA extraction assays from fungi have used a small volume of blood

[15,14,17]. In our study we used a large volume of blood (4 ml) spiked with around 1,000 cells and a large sample volume might help to increase the sensitivity of the assay

because of the higher yield of Candida DNA. The aim of this study was to reduce the

time necessary for fungal DNA extraction and to simplify the working process. The cell

wall can be disrupted mechanically or enzymatically and because mechanical disruption

is faster we replaced the lysis of leucocytes and the lyticase step (which are enzymatic

steps) with glass bead cell disruption. We compared mechanical cell wall disruption

using glass beads with enzymatic disruption using recombinant lyticase. Using glass

beads would simplify the working process and shorten duration of DNA isolation with

1.5 h but the mechanical cell wall disruption was not sensitive enough for reliable

detection of Candida. It may depend on an insufficient disruption of the Candida cells because of wrong size and composition of glass beads, composition of the lysis buffer or configuration of the vortexing process (speed and duration).

We also compared the DNA extraction using QIAamp tissue kit with two automated extraction robots: BioRobot M48 and NucliSens easyMAG, to determine their sensitivity, reliability and duration for DNA release of C. albicans. The automated methods gave statistically significant lower Ct-values than the manual method indicating that the robots recovered a larger quantity and/or a higher quality of DNA. Both robots use magnetic beads for DNA purification and have thorough cleaning steps and less inhibition may explain the lower Ct-values. The BioRobot M48 and NucliSens easyMAG are fully enclosed robots, which not only increase the sample throughput but also reduce human interaction with samples, decreasing the overall risk of contamination. Automated DNA extraction also allows performing several samples on the same run. The NucliSens easyMAG robot proved to be slightly more sensitive than BioRobot M48, which maybe dependent on that NucliSens easyMAG have a more enclosed system. The BioRobot M48 has a larger elute volume than NucliSens easyMAG and QIAamp tissue kit, which could change the Ct-values with a few tenths.

There is a need for rapid, standard method for DNA extraction Candida to decrease the

work burden and to allow better comparisons between laboratories. The data generated in

this study suggest that the NucliSens easyMAG is superior to the BioRobot M48 and

QIAamp tissue kit for the extraction of Candida DNA. The NucliSens easyMAG was

shown to be a rapid, sensitive and reliable assay for automated extraction of DNA and

could hopefully be a valuable tool for DNA extraction from a wide range of fungal

species. More research in this area is needed in order to hopefully reach a stable and

reliable method to rapidly detect clinical cases of candidiasis, and give an early diagnosis

and prompt initiation of antifungal therapy.

REFERENCES

1. Loeffler J, Norbert H, Hebart H, Schmidt D, et al. Quantification of fungal DNA by using fluorescence resonance energy transfer and the light cycler system.

J Clin Microbiol, 2000. 38 (2): p. 586-590.

2. Erjavec Z, Verweij P.E. Recent progress in the diagnosis of fungal infections in the immunocompromised. Drug Res Update 5, 2002. p. 3-10.

3. Sullivan D.J, Moran G.P, Coleman D.C. Candida dubliniensis: Ten years on. FEMS Microbiol Lett, 253, 2005. p. 9-17.

4. Louie A, Deziel M, Liu W, Drusano M.F, et al. Pharmacodynamics of caspofungin in a murine model of systemic candidiasis: importance of persistence of caspogungin in tissues to understanding drug activity. Antimicrob Agents Chemother, 2005.

49(12): p. 5058-5068.

5. Klingspor L, Törnqvist E, Johansson A, et al. A prospective epidemiological survey of candidaemia in Sweden. Scand J Infect Dis, 2004. 36: p. 52-55,

6. Rupp S. Proteomics on its way to study host-pathogen interaction in Candida albicans. Curr Opin Microbiol, 2004. 7: p. 330-335.

7. Niimi M, Cannon R.D, Monk B.C. Candida albicans pathogenicity: A proteomic perspective. Electrophoresis 1999. 20: p. 2299-2308.

8. Krcmery V, Barnes A.J. Non-albicans candida spp. causing fungaemia:

pathogenicity and antifungal resistance. J Hos Infect, 2002. 50(4): p. 171-177.

9. Fidel P, Vazquez J.A, Sobel J.D. candida glabrata: review of epidemiology, pathogenesis, and clinical disease with comparison to C. albicans.

Clin Microbiol Rev, 1999. 12 (1): p. 80-96.

10. Hermann E, Hebart H, Roller G, et al. Detection and identification of fungal pathogens in blood by using molecular probes. J Clin Microbiol, 1997.

35 (6): p. 1353-1360

11. Shin J.H, Nolte F, Holloway B, et al. Rapid identification of up to three Candida

species in a single reaction tube by a 5’ exonuclease assay using fluorecent DNA

probes. J Clin Microbiol, 1999. 37(1): p.165-170.

12. Maaroufi Y, Heymans C, De Bruyne JM, et al. Rapid detection of Candida albicans in clinical blood samples by using TaqMan-based PCR assay. J Clin Microbiol, 2003.

41 (7): p. 2393-3298

13. Fredricks D, Smith C, Meier A. Comparison of six DNA extraction methods for recovery of fungal DNA as assessed by quantitative PCR. J Clin Microbiol, 2005.

43(10): p. 5122-5128.

14. Maarouf Y, Ahariz N, Husson M and Crokaert F. Comparison of different methods of isolation of DNA of commonly encountered Candida species and its quantitation by using a real-time PCR-based assay. J Clin Microbiol, 2004.

42 (7): p. 3159-3163.

15. Bougnoux M, Dupont C, Mateo J, et al. Serum is more suitable than whole blood for diagnosis of systemic candidiasis by nested PCR. J Clin Microbiol, 1999.

37: p. 925-930

16. Löffler J, Hebart H, Schumacher U, et al. Comparison of different methods for extraction of DNA of fungal pathogens from cultures and blood. J Clin Microbiol, 1997. 35 (12): p. 3311-3312

17. Löffler J, Schmidt K, Hebart H, et al. Automated extraction of genomic DNA from

medically important yeast species and filamentous fungi by using the MagNa Pure

LC system. J Clin Microbiol, 2002. 40 (6): p. 2240-2243