mycoses

Diagnosis,Therapy and Prophylaxis of Fungal Diseases

Review article

Candida albicans or Candida dubliniensis? Ruan Ells, Johan L. F. Kock and Carolina H. Pohl Department of Microbial, Biochemical and Food Biotechnology, University of the Free State, Bloemfontein, South Africa

Summary

Candida albicans is increasing as an opportunistic pathogen causing candidemia and candidiasis worldwide. In addition, other non-albicans Candida species are now also associated with pertinent infections. These include the closely related C. dubliniensis, which shares many phenotypic similarities with C. albicans. These similarities pose problems in the identification of isolates and have previously led to misidentification of these species. As a result, several identification techniques based on phenotypic and genotypic characteristics have been developed to differentiate between these Candida species. This review will focus on the similarities and differences between these two Candida species highlighting different identification methods and their advantages and disadvantages.

Key words: Candida albicans, Candida dubliniensis, genotypic identification, phenotypic identification.

Introduction Several members of the genus Candida exist as commensals of mucosal membranes in most healthy individuals and other warm-blooded animals, where they grow without causing any damage.1,2 However, under conditions where the hostÕs defence mechanisms are compromised, they can become pathogenic.3 Until recently Candida albicans was considered the most important opportunistic pathogen in this genus. However, other non-albicans Candida species, such as C. glabrata, C. krusei, C. parapsilosis and C. tropicalis have also emerged as causative agents of infections.2 Another non-albicans Candida species found to be an emerging opportunistic pathogen is C. dubliniensis, initially isolated from AIDS patients in Dublin, Ireland.4 Although C. dubliniensis was mainly associated with oral candidiasis in HIV infected patients worldwide, it is now emerging as a cause of superficial and systemic disease in HIV negative individuals with an estimated prevalence rate below 5%.5–8 Correspondence: Carolina H. Pohl, Department of Microbial, Biochemical and Food Biotechnology, University of the Free State, PO Box 339, Bloemfontein, 9301, South Africa. Tel.: +27 51 4019197. Fax: +27 51 4443219. E-mail: [email protected] Accepted for publication 2 June 2009

 2009 Blackwell Verlag GmbH • Mycoses 54, 1–16

Due to phenotypic similarity between C. dubliniensis and C. albicans, C. dubliniensis was previously misidentified as C. albicans.9 These similarities still make it difficult to rapidly differentiate between the two species, especially in clinical samples and may lead to an underestimation of the prevalence of this species. To assess the epidemiological role and clinical significance of C. dubliniensis, reliable methods for differentiation between these two species are necessary. Therefore, several phenotypic and genotypic characterisation and differentiation methods have been developed. This review will focus on the similarities and differences between C. albicans and C. dubliniensis. A critical evaluation of the accuracy of some differentiation techniques will be attempted.

Phenotypic characteristics for differentiation The production of germ tubes and chlamydospores was considered typical of C. albicans isolates and was routinely used for identification of this species. However, it was realised that some isolates identified using these characteristics, differ genetically and in their carbohydrate assimilation profiles from the known C. albicans strains. This led to the description of a new species, C. dubliniensis, by Sullivan et al. [4]. Confirmation studies of two yeast culture collections indicated that between 1% and 2% of C. dubliniensis isolates was

doi:10.1111/j.1439-0507.2009.01759.x

R. Ells et al.

incorrectly identified as C. albicans10,11 and in another study of oral yeasts isolated from HIV infected individuals, 16.5% of C. dubliniensis isolates was incorrectly identified as C. albicans.12 Due to the demand of clinical diagnostic laboratories for accurate, reliable, inexpensive and rapid identification techniques which can be applied to large numbers of samples, several phenotype-based tests are available to differentiate between these species. The most commonly used phenotypic tests involve the determination of colony colour and morphology using differential media. Characteristics on chromogenic media

A widely used commercial medium for the differentiation of Candida species in mixed samples is CHROMagar Candida, which contains a combination of chromogenic substrates, including a b-hexosaminidase substrate.13,14 Differentiation on CHROMagar Candida is based on colour differences produced by different Candida species after growth at 37 C for 48 h, due to the presence of species-specific enzymes. Candida albicans colonies appear light blue-green whereas C. dubliniensis isolates form dark green colonies after incubation at 37 C for 48 h, with the colour being more prominent after 72 h of incubation.15 Odds & Davidson [16] examined the influence of incubation temperature on colour formation and found that for certain species, including C. albicans and C. dubliniensis, incubation temperature plays an important role in the colour development and that the original specification of 37 C should be adhered to. Several authors found that incubation at 30 C may result in the production of colonies with different shades of green by C. dubliniensis,17 or that some C. albicans isolates also produced dark

(a)

green colonies.18,19 It should also be noted that C. dubliniensis might lose its ability to form this distinct dark green colour after subculturing and storage,20 limiting the accuracy of this medium to primary isolates. Other chromogenic media containing hexosaminidase substrates (e.g. Albicans ID2 and the improved Candida ID2) are also often used, however 100% accuracy is not guaranteed either.21 The reduction of 2,3,5-triphenyltetrazolium chloride (TTC) in agar media has been used traditionally to identify C. tropicalis.22 This medium was also found to differentiate C. albicans, which cannot reduce TTC and produces white to light pink colonies, from C. dubliniensis, which can reduce TTC and produces red to maroon colonies. Giammanco et al. [23] evaluated the use of this medium to identify C. dubliniensis in clinical samples and found that all five isolates identified as C. dubliniensis produced dark pink to red colonies, while all 16 C. albicans isolates produced white or light pink colonies. Unfortunately, larger collections of isolates have not been screened to determine the conserved nature of this character. Chlamydospore production

It is known that C. dubliniensis tends to form pairs, triplets or clusters of chlamydospores on the ends of short-branched hyphae, whereas C. albicans forms a single chlamydospore on the tip of an elongated suspensor cell.4 However, care should be taken when using chlamydospore production to differentiate between the species. In contrast to SullivanÕs work, Ellepola et al. [24] demonstrated on corn meal-Tween agar that differentiation based on chlamydospore formation is not reliable, since C. dubliniensis showed no constant pattern of chlamydospore formation (Fig. 1).

(b)

(c)

Figure 1 Morphology of C. albicans and C. dubliniensis chlamydospores. (a) Example of a C. dubliniensis strain demonstrating occasional

triplet or pairs of chlamydospores (arrows), on the ends of short, hyperbranching hyphae. (b) Example of a C. albicans strain demonstrating a single terminal chlamydospore arising from a suspensor cell (arrow), (c) Example of a C. dubliniensis strain showing almost identical single terminal chlamydospores (arrow) to that in (b). Reproduced with permission from Ref. [24].

2

 2009 Blackwell Verlag GmbH • Mycoses 54, 1–16

Candida albicans or Candida dubliniensis

Similarly, Kurzai et al. [18] found that characteristic chlamydospore formation on rice-Tween agar was only observed in 57% of C. dubliniensis isolates and concluded that this characteristic cannot be used for accurate identification of this species. In addition, Kirkpatrick et al. [25] found that only 70% of the examined C. dubliniensis isolates produced abundant chlamydospores, while 4% of the C. albicans isolates also produced abundant chlamydospores on corn meal agar. This was ´ lvarez et al. [17] who found that 13% of confirmed by A C. albicans isolates examined also produced abundant chlamydospores in pairs, triplets or clusters on corn meal-Tween agar. Al Mosaid et al. [26] used PalÕs agar, containing sunflower seed extract, to differentiate between C. albicans and C. dubliniensis. It was found that all the isolates formed smooth creamy-grey colonies after 48–72 h incubation. However, C. dubliniensis isolates had hyphal fringes, grew as rough colonies and produced chlamydospores, whereas C. albicans did not. It was found that incubation at 30 C rather than 37 C provided better differentiation between these species. A combination of equal volumes of CHROMagar Candida and PalÕs agar, as described by Sahand et al. [27], enhances the differences between these two species and has the added benefit of allowing identification of these species in mixed clinical samples, since C. dubliniensis forms dark green, rough colonies and C. albicans forms light green, smooth colonies (Fig. 2). Another medium, on which the same effect of chlamydospore and hyphal fringe production by C. dubliniensis but not C. albicans was observed, is

Figure 2 Colonies of different species of Candida after growing

for 48 h at 37 C in CHROMagar Candida medium supplemented with PalÕs agar. Reproduced with permission from American Society for Microbiology from Ref. [27].

 2009 Blackwell Verlag GmbH • Mycoses 54, 1–16

tobacco agar.28 However, Kumar & Menon [29] found that tobacco agar is not always suitable for the differentiation of C. albicans from C. dubliniensis, since 96% of C. albicans strains tested by these authors also produced chlamydospores on tobacco agar. Staib agar (Niger seed agar), which was originally developed to differentiate Cryptococcus neoformans from other yeasts found in clinical samples, was also reported to be able to differentiate between C. albicans and C. dubliniensis based on especially the presence and absence of chlamydospores.30 These authors found that all 14 C. dubliniensis isolates tested produced rough colonies, pseudohyphae and chlamydospores (Fig. 3 and Fig. 4). Although the 11 C. albicans isolates tested produced pseudohyphae after prolonged growth, none of these isolates ever produced chlamydospores on this media. These results were confirmed by Kurzai et al. [18] with 21 C. dubliniensis and 82 C. albicans strains. However, when a larger collection of isolates (120 C. dubliniensis and 166 C. albicans isolates) was tested, it was found that, although no C. albicans isolates produced chlamydospores on this medium, only 85.4% of C. dubliniensis isolates were able to produce chlamydospores.5 These authors found that all C. dubliniensis isolates produced rough colonies and all C. albicans isolates produced smooth colonies on this media and suggested that colony morphology is a more reliable discriminating character. The solution suggested by Wabale et al. [31] of using both chlamydospore and rough colony formation on this media to differentiate between these species seems to be the most suitable in this case. It should, however, be noted that Ma¨hnß et al. [19] found that incubation time is an important factor in chlamydospore production and

Figure 3 Colony morphology of C. dubliniensis strain and C. albicans strain after 3 days of growth on Staib agar at 30 C. Note the rough appearance of the C. dubliniensis colonies and the smooth surface of the C. albicans colonies. Reproduced with permission from Ref. [30].

3

R. Ells et al.

(a)

critical evaluation of more strains will have to be performed to determine the sensitivity and specificity of these media. Fluorescence on methyl-blue agar

(b)

(c)

Methyl blue-Sabouraud agar is also reported as useful for differentiation of these species.35 Differentiation using this media is based on the ability of C. albicans colonies to fluoresce under WoodÕs light (long wave length UV) and the inability of C. dubliniensis to do so. The exact reaction between methyl blue and C. albicans is unknown, but there might be a reaction with specific cell wall polysaccharides which produces the fluorescent metabolite.36 However, this medium also has its limitations. Candida albicans isolates can lose their fluorescence and C. dubliniensis can become fluorescent after subculturing and storage of the isolate.8 This indicates that this medium is only useful after initial isolation of a Candida culture. Growth in the presence of sodium chloride

(d)

Figure 4 Pseudomycelial growth and chlamydospore production by C. dubliniensis Wu¨284 after 2 (a), 3 (b), and 10 (c) days of growth on Staib agar at 30 C. Shown are cells that were taken from the edge of a colony and were photographed using objective 40 · . (d) Cells from a C. albicans Wu¨212 colony grown for 3 days under identical conditions are shown for comparison. Reproduced with permission from Ref. [30].

indicated that three to four days produced the best results. Alves et al. [32] evaluated the use of tomato juice agar to differentiate between these species and found that all 26 C. dubliniensis isolates tested produced chlamydospores, while 92% of the C. albicans isolates did not produce any chlamydospores and those that did, produced them in very low numbers. Other new media, containing extracts of sesame seed, linseed, rosemary or oregano33,34 were proposed. Similar to well known media containing plant extracts, C. dubliniensis colonies were rough with hyphal fringes and chlamydospores, compared to the smooth colonies without chlamydospores produced by C. albicans on these media. However,

4

Alves et al. [37] developed a screening media based on the inability of C. dubliniensis to grow in Sabouraud dextrose broth containing 6.5% NaCl. All the C. albicans isolates tested (250) grew in this medium, while none of the 19 C. dubliniensis isolates were able to grow. The same results were obtained in subsequent studies.38,39 Using Sabouraud dextrose agar (SDA) containing 6.5% ´ lvarez et al. [16] found that two of the seven NaCl, A tested C. dubliniensis isolates grew in the presence of NaCl, albeit to a very small extent. Growth at elevated temperatures

It was found by Sullivan et al. [4] that C. dubliniensis can grow well at 30 C and 37 C, producing creamy white colonies on solid media similar to C. albicans. However, differing from C. albicans, it grew poorly or was unable to grow at 42 C on SDA or potato dextrose agar. In contrast to this, Pinjon et al. [40] found that 9.2% of the C. dubliniensis isolates showed partial growth at 42 C after 48 h but that none of the 120 C. dubliniensis isolates tested grew at 45 C after 48 h. In the case of C. albicans, all 98 isolates grew at 42 C and 97 grew at 45 C. Similar results regarding the inability of C. dubliniensis isolates to grow at 45 C were obtained by other authors.18,23 However, Kurzai et al. [18] reported that 15.9% of C. albicans isolates also failed to grow at this temperature. In addition, Gales et al. [41] reported that 23% of the tested C. albicans isolates failed

 2009 Blackwell Verlag GmbH • Mycoses 54, 1–16

Candida albicans or Candida dubliniensis

to grow at 45 C and Kirkpatrick et al. [25] reported an even higher percentage of C. albicans isolates (36%) that failed to grow at this temperature. Pincus et al. [20] stated that these variable results may be explained by poor temperature control in incubators or differences in media composition. Carbohydrate assimilation and commercial identification systems

Candida isolates have the ability to assimilate a range of carbohydrate compounds as their sole carbon source, and differences in these assimilation profiles have been used to differentiate C. dubliniensis from C. albicans.4 These differences include the inability of C. dubliniensis to assimilate a-methyl-D-glucoside (MDG), lactate (LAT) or xylose (XYL). Several authors have evaluated and compared the use of commercially available identification systems, which rely on carbon source assimilation profiles, in the differentiation of these two species. Evaluating the API 20C AUX system, Pincus et al. [20] found that, although the final identification accuracy was low, with 95% of the isolates misidentified as C. albicans, none of the 80 C. dubliniensis isolates tested could assimilate XYL or MDG after 72 h of incubation, whereas the database percentages of these substrates for C. albicans are 88 and 85, respectively. Gales et al. [41] confirmed these results for the API 20C AUX system on 66 C. dubliniensis isolates and 100 C. albicans isolates. None of the tested C. dubliniensis isolates could assimilate XYL or MDG, while all the C. albicans could assimilate at least one of these two carbon sources. Jabra-Rizk et al. [10] also found that none of the C. dubliniensis strains tested using this system could assimilate XYL or MDG. Although the VITEK YBC system misidentified 99% of the C. dubliniensis isolates as C. albicans after 48 h of incubation, Pincus et al. [20] found that only 6% of the tested isolates could assimilate XYL, whereas the database percentage for this substrate for C. albicans is 95%. Evaluation of this system by Gales et al. [41] found that 3% of C. dubliniensis isolates could assimilate XYL. Assimilation of XYL was also detected in four of the 12 C. dublinienisis isolates studied by Ahmad et al. [38], however they found no assimilation of MDG or LAT by these isolates. Interestingly Pincus et al. [20] found that 64% of C. dubliniensis isolates could assimilate MDG after 48 h using this system, but Gales et al. [41] found only 4.5% MDG positive isolates. Again, all 100 C. albicans isolates could assimilate either XYL or MDG or both. Evaluation of the VITEK 2 ID-YST system produced accurate species identification of 98% of the

 2009 Blackwell Verlag GmbH • Mycoses 54, 1–16

C. dubliniensis isolates and Pincus et al. [20] found that assimilation of MDG and LAT could differentiate between the two species, with only 1% of C. dubliniensis isolates able to assimilate MGD and 4% able to assimilate LAT. The database percentages of these substrates for C. albicans are 99 and 98, respectively. Although 100% identification accuracy have been achieved using this system,42 other authors have been less successful,43 however the discrepancies did not occur in the assimilation of MDG, but 12% of the C. dubliniensis isolates could assimilate LAT. Ahmad et al. [39] used this identification system to characterise clinical isolates of C. dubliniensis obtained in Kuwait and also found that none of the isolates could assimilate MDG or LAT. Pincus et al. [20] also evaluated the ID 32 C system, and found that none of the C. dubliniensis isolates could assimilate LAT, MDG or XYL. The database percentages for C. albicans are 96%, 98% and 98%, respectively, for these carbon sources. The same results were obtained regarding MDG and XYL assimilation by 19 Brazilian C. dubliniensis isolates, however one of the isolates could assimilate LAT.44 Pincus et al. [20] obtained an overall identification accuracy of 97% and this level of accuracy has been confirmed43 and exceeded,45 however in another study of 53 C. dubliniensis isolates, this system could only identify 3.7% correctly, due to the high frequencies of assimilation of XYL (58%), LAT (87%) and MDG (55%).46 Candida dubliniensis isolates, able to assimilate XYL using this system, were also detected in another study, leading to incorrect identification.19 A relatively new commercial identification system, Micronaut-Candida (Micronaut-RC) has been evaluated by Kurzai et al. [18] using 21 C. dubliniensis isolates. They found that although none of the isolates could assimilate LAT, 11 could assimilate XYL, however an overall identification accuracy of 95% was obtained. Szabo´ et al. [45] studied 40 C. albicans and 9 C. dubliniensis isolates and used this system to correctly identify all these isolates, however larger isolate collections will provide a better understanding of the accuracy of this system. Pincus et al. [20] found that the rapID Yeast Plus system, which does not contain C. dubliniensis in the database, generated 32 different profiles for the 80 C. dubliniensis isolates tested, and misidentified 81% of the isolates as C. albicans. Although Sullivan et al. [4] found that neither C. dubliniensis nor C. albicans could assimilate glycerol as sole carbon source, other studies indicated that C. dubliniensis does have the ability to assimilate glycerol.47 Pincus et al. [20] found that depending on the identification system used, different percentages of C. dubliniensis isolates were able to assimilate glycerol.

5

R. Ells et al.

Similarly, several authors used differences in trehalose assimilation as a distinguishing character. However, Sullivan et al. [4] found trehalose assimilation to be variable between the originally characterised isolates of C. dubliniensis and Pincus et al. [20] reported that, depending on the identification system used, different percentages of the tested C. dubliniensis isolates could assimilate trehalose. Ahmad et al. [38,39] as well as Al-Sweih et al. [48] found that none of the tested C. dubliniensis isolates characterised using the VITEK 2 ID-YST system could assimilate trehalose. Alves et al. [44] found that one of the 19 clinical C. dubliniensis strains characterised using the ID 32C system, could assimilate trehalose and Coleman et al. [49], using the same system, reported that half of the tested isolates could assimilate trehalose. In contrast, Kurzai et al. [18] reported that all the tested C. dubliniensis isolates were able to assimilate trehalose using the MicronautCandida system. Serological characteristics

Bikandi et al. [50] demonstrated the possibility of using differences between C. albicans and C. dubliniensis cell wall localised antigens to differentiate between these species. Cross reactivity was, however, observed for C. krusei. Later it was demonstrated that antigens extracted from the cell walls of C. dubliniensis lead to the production of species specific antibodies in a rabbit model of systemic infection.51 The detection of these C. dubliniensis specific antibodies made it possible to diagnose invasive candidiasis caused by this species. In addition, Bliss et al. [52] reported the use of single-chain antibody fragments that bind specifically to C. albicans and C. dubliniensis hyphae. By using these fragments in combination with another fragment described by Haidaris et al. [53], which specifically recognises C. albicans blastoconidia, it became possible to differentiate between these two species. These findings, together with the fact that serological tests based on monoclonal antibodies (MAbs) exist for the identification of C. albicans and C. krusei,54 have lead to the development of serological tests based on MAbs for the identification of C. dubliniensis. An immunochromatographic membrane test (ICM albi-dubli test) uses two MAbs. The one binds to a glycoprotein expressed by both C. albicans and C. dubliniensis cell walls and the second binds to a glycoprotein exclusively expressed by the surface of C. albicans hyphae.54 This system was evaluated for 58 C. albicans and 60 C. dubliniensis isolates and it was found that three C. albicans isolates and two C. dubliniensis isolates were misidentified. It was also reported

6

that the incubation media played a role in the accuracy of this test with the most accurate results obtained with SDA, CHROMagar Candida and CandiSelect. The most inaccurate results were obtained using Candida ID media. A commercial latex agglutination test (Bichro-dubli) consists of latex beads coated with an antibody specific for C. dubliniensis, suspended in a dissociating dye.55 Evaluation of this assay revealed that none of the tested C. albicans strains produced a positive result and that it was therefore 100% specific for C. dubliniensis. However, one C. dubliniensis isolate, grown on Candida ID medium, produced a negative result. The sensitivity could be increased to 100% when C. dubliniensis was grown on SDA, CandiSelect and CHROMagar Candida. Two other studies have so far evaluated the performance of this rapid test. Sahand et al. [56] confirmed that none of the tested C. albicans isolates and 97.1% of the tested C. dubliniensis isolates produced a positive result. However, they found that the media in which the isolates were grown did not influence the results. Chryssanthou et al. [57] confirmed the sensitivity and specificity of this test and it was used to correctly identify a clinical C. dubliniensis isolate in a study of denture stomatitis.58 Protein fingerprinting

Analyses of the cellular proteins of microorganisms have proven useful in the identification and resolution of bacterial strains. One of the most popular and inexpensive techniques is SDS–PAGE.59 Cellular proteins are extracted, denatured and separated on a polyacrylamide gel, producing a characteristic protein banding pattern. Although Cunningham & Noble [60] stated that the reproducibility of SDS–PAGE results of different Candida species is low and that the technique is not suitable for differentiation of these species, other authors have shown that this technique is reproducible and able to identify different Candida species.59,61 In addition, Rosa et al. [62] confirmed the suitability of using this technique to aid in the differentiation between C. albicans and C. dubliniensis. However, Neppelenbroek et al. [59] stated that more research on larger collections of isolates is needed to determine the efficacy of this technique for the differentiation of these two species. Another protein-based identification technique is multilocus enzyme electrophoresis (MLEE) which gives an indication of the polymorphism of isoenzymes or allozymes of the isolates.59 Cellular proteins are extracted and separated in starch or polyacrylamide gels under native conditions. After separation the

 2009 Blackwell Verlag GmbH • Mycoses 54, 1–16

Candida albicans or Candida dubliniensis

activity of the enzymes is revealed by means of specific staining protocols.63 The enzymes examined are usually house keeping enzymes and are assumed to be less prone to selective pressure. Data obtained by MLEE are therefore considered to be indirectly representative of the genome of the organism.64 Although several authors have used MLEE to study different Candida species, very few studies have attempted differentiation between C. albicans and C. dubliniensis using this technique.59,61 In one such study, Boerlin et al. [65] found that C. albicans isolates from HIV positive patients produced two distinct groups. One group was characterised by the absence of b-glucosidase activity and atypical assimilation patterns and was later recognised as C. dubliniensis. In a subsequent study it was confirmed that MLEE analyses of C. dubliniensis and C. albicans isolates produces two species-specific taxa and that this technique may be useful in the differentiation of these two species.62 Whole organism fingerprinting

Recent developments in analytical instruments have made the rapid, automated and accurate identification of clinical yeast possible without the need for extensive sample preparation.66 Physiochemical spectroscopic methods such as pyrolysis-mass spectrometry, Fourier transform-infrared spectroscopy and UV resonance Raman spectroscopy allows for the direct analyses of samples and provides so called whole organism fingerprints which is a phenotypic measure of the physiochemical composition of the organism. It must, however, be noted that other sources of chemical variation in biological samples, such as media composition and colony size should be minimised to obtain standardised spectra for each organism.67 Pyrolysis-mass spectrometry has been used successfully in the identification of clinically relevant bacteria and involves the thermal degradation of the sample and the production of volatile molecular fragments.66 These fragments are analysed using a mass-spectrometer, yielding a mass spectrum which measures the strengths of covalent bonds between the molecules of the organism. It was found that the mass spectra of C. dubliniensis and C. albicans showed little qualitative differences, but quantitative differences between the spectra could be observed. After statistical analyses of the data, it was possible to group the C. albicans strains together and separate them from the group of C. dubliniensis strains. Fourier transform-infrared (FT-IR) spectroscopy measures vibrations of functional groups and polar bonds.66 It has been reported that the complex biochemical

 2009 Blackwell Verlag GmbH • Mycoses 54, 1–16

fingerprints obtained using this technique can be used to identify bacteria and fungi at strain level. When the FT-IR spectra of C. albicans and C. dubliniensis were compared, no qualitative differences were again observed. However, after statistical analyses of the quantitative differences, C. albicans and C. dubliniensis could again be grouped into separate groups. These results were confirmed by Tietz et al. [68] when FT-IR spectroscopy was used to describe a new Candida species (C. africana) which is closely related to C. albicans and C. dubliniensis. In another FT-IR spectroscopy study by Maquelin et al. [67] one C. albicans reference strain was misidentified as C. dubliniensis, but all clinical isolates were correctly identified as were the C. dubliniensis reference strains.

Genotypic characteristics used in differentiation As can be seen from the previous discussion, most phenotypic characteristics are at best not confirmed for large numbers of strains and at worst are variable and may lead to incorrect classification of the species. A more reliable approach in discriminating between organisms is based on molecular techniques. Restriction fragment length polymorphism

A popular technique used in the generation and analyses of DNA fingerprints is restriction fragment length polymorphism (RFLP). This involves the isolation and digestion of genomic DNA. The fragments obtained are separated by gel electrophoresis and can either be detected by hybridizing with species-specific DNA probes, followed by Southern blot analyses, or be stained with ethidium bromide and directly visualised by UV transillumination.59 Candida dubliniensis was initially identified by digesting genomic DNA of atypical C. albicans isolates with EcoRI, followed by hybridisation with the C. albicans-specific mid repeat sequence probe, 27A (Fig. 5).4 This probe gave a fingerprinting pattern of 10–15 strongly hybridising bands, varying from 500 bp to 20 kb in the case of C. albicans, whereas the atypical isolates (C. dubliniensis) produced only four to seven faint bands varying between five and 20 kb in size. Similar results were obtained by Diaz-Guerra et al. [69] when EcoRI digested genomic DNA of C. dubliniensis isolates yielded weakly hybridising bands ranging between 4.4 kb and 23.1 kb in size. Another C. albicansspecific probe often used is the moderately repetitive sequence, Ca3. Schoofs et al. [35] reported that this probe yields 15–25 bands ranging from 2.25 kb to

7

R. Ells et al.

(a)

(b) Figure 5 Southern blot analysis of EcoR1 digested total genomic DNA from C. albicans and C. dubliniensis isolates probed with (a) 32P-labelled C. albicans-specific probe 27A and (b) 32[P]-labelled oligonucleotide (GGAT)4. The fingerprints shown correspond to C. albicans isolates 132A (lane 1) and CM 3 (lane 2), C. stellatoidea type II strain ATCC 20408 (lane 3), C. stellatoidea type I strain ATCC 11006 (lane 4), C. dubliniensis isolates CD33 (lane 5), CD36 (lane 6), CD38 (lane 7), CM2 (lane 8), CM5 (lane 9), CM7 (lane 10) and NCPF 3108 (lane 11). Size reference markers are indicated in kb on the left of the figure. Reproduced with permission from Ref. [4].

7.9 kb in the case of C. albicans EcoRI digested genomic DNA, but produced only four weakly hybridising bands, between 6.5 kb and 7.9 kb and one or two heavier bands, for C. dubliniensis. Kurzai et al. [70] demonstrated that it was possible to use an 800-bp ClaI fragment of the PHR1 gene or a 1.6 kb HindIII fragment of the PHR2 gene of C. albicans as hybridisation probes for differentiation of these two species. After digestion of genomic DNA by EcoRI all C. albicans isolates produced a 3.7 kb fragment which hybridised with the PHR1 probe, while all C. dubliniensis produced a 6.5 kb fragment. After digestion of genomic DNA by HindIII, C. albicans isolates produced an 8 kb fragment which hybridised with the PHR1 probe, while all C. dubliniensis isolates produced a 7 kb fragment. The PHR2 probe was less useful with several isolates of both C. albicans and C. dubliniensis producing no hybridising fragments. However, in cases where fragments were detected, clear differences were again observed, especially after digestion with EcoRI. A probe specific for C. dubliniensis has been developed and consists of a moderately repetitive element, Cd25.71 This probe has been successfully used, in combination with Ca3, to differentiate between these two species.72,73 Although RFLP with hybridisation is considered very informative, it is seen as time consuming since blots are required.59 Direct observation of the ethidium bromide stained bands by UV transillumination may be one solution to this draw back. Using this technique, Sullivan et al. [4] demonstrated that digestion of the genomic DNA of atypical isolates (C. dubliniensis) with HinfI results in a high molecular mass fragment (15 kb) unique to these isolates. Fujita & Hashimoto [74]

8

evaluated the use of RFLP analyses of genomic DNA of different Candida species using HinfI and found that C. albicans isolates produced between two and six major fragments, ranging in size from 2.03 kb to 7.51 kb, while C. dubliniensis isolates produced four major fragments ranging between 2.35 kb and 2.88 kb in size, but made no mention of a high molecular mass fragment. One of the strengths of RFLP analyses lies in the fact that it may be combined with other molecular techniques such as PCR. Several authors have demonstrated that it is possible to use restriction patterns generated from amplified ribosomal DNA sequences to differentiate between closely related species such as C. albicans and C. dubliniensis. Irobi et al. [75] amplified the internal transcribed spacer (ITS) regions of several Candida species, including C. albicans and C. dubliniensis and found that both these species produced a PCR product of approximately 1200 kb. Further analyses of restriction patterns after digestion with BfaI, DdeI or HaeIII revealed that there are distinct differences between these two species, enabling their differentiation. McCullough et al. [76] amplified the ITS regions together with the 5.8S rDNA of both species and digested the product with DdeI. The results clearly differentiated between C. albicans, producing only one fragment, and C. dubliniensis, producing two fragments. These authors also amplified the V3 region of the 25S rDNA of both species and digested the products with HaeIII. Although both species produced only one fragment there was a size difference, with the fragment from C. albicans being slightly larger (314 kb) than the C. dubliniensis fragment (234 kb). Park et al. [77] amplified a conserved part of the 5.8S rDNA, the adjacent ITS2 region and a

 2009 Blackwell Verlag GmbH • Mycoses 54, 1–16

Candida albicans or Candida dubliniensis

portion of the 28S rDNA. The products were digested with BsmAI which recognises a specific sequence in this region from C. dubliniensis, yielding two fragments (0.25 and 0.9 kb). The undigested products were restricted with NspBIII, which recognised a site in the ITS region of C. albicans, yielding two fragments (0.18 and 0.16 kb). Using a semi-nested PCR technique Ahmad et al. [38] amplified the ITS2 regions of these species and found that digestion of these products with MspAII yielded three fragments for C. albicans (35, 143 and 167 bp) and only two for C. dubliniensis (35 and 515 bp). Graf et al. [78] used primers that bind to conserved regions of the 18S and 28S rDNA to amplify the ITS regions together with the 5.8S rDNA of both species. The PCR products were digested with HpyF10VI yielding three fragments (141, 184 and 261 bp) for C. albicans isolates and only two fragments (264 and 325 bp) for C. dubliniensis. Sullivan et al. [4] also performed RFLP analyses using short oligonucleotide primers which target the microsatellite regions and found sufficient differences in these patterns to differentiate the two species. The generation of simple RFLP patterns (such as reported by Ahmad et al. [38] and Graf et al. [78]) allowing unequivocal differentiation between these species, has an advantage over other more complex RFLP patterns which are often ambiguous, difficult to interpret and require computer software, including databases. Randomly amplified polymorphic DNA analyses

Randomly amplified polymorphic DNA (RAPD) analyses involve the PCR amplification of target genomic DNA sequences using short primers at a low annealing temperature.59 The PCR products are separated according to size on an agarose gel and stained with ethidium bromide to yield RAPD profiles. Five oligonucleotide primers were used by Sullivan et al. [4] to produce RAPD profiles of atypical C. albicans isolates (C. dubliniensis) in the original description of the species. Differentiation between C. albicans and C. dubliniensis on the basis of RAPD profiles generated using the minisatellite probe derived from the phage M13 core sequence as a single primer was also possible.79 In a study by Mariano et al. [80] RAPD profiles of isolates previously identified as C. albicans were generated using three different primers, making it possible to ascertain that approximately 2% of these isolates were incorrectly identified and were in fact C. dubliniensis. Several other authors could also generate species specific RAPD fingerprints, enabling the differentiation of C. albicans and C. dubliniensis.81–83

 2009 Blackwell Verlag GmbH • Mycoses 54, 1–16

Amplified fragment length polymorphism

A more recently developed fingerprinting technique is amplified fragment length polymorphism (AFLP). This technique involves the isolation and restriction of genomic DNA, followed by ligation of restriction half-sitespecific adaptors to the sticky ends of the fragments.59,84 These fragments are then amplified using primers with sequences corresponding to the adaptor and restriction site sequences and is visualised on an agarose gel. This method is considered more reproducible than RAPD or RFLP. Borst et al. [85] applied this technique to the identification of reference strains of clinically important Candida species, including C. albicans and C. dubliniensis. These authors digested the genomic DNA with EcoRI and MseI and used adaptors and primers recognising these restriction sequences in subsequent PCRs. After phylogenetic analyses of the AFLP fingerprinting patterns it was possible to distinguish C. albicans from C. dubliniensis. The same results were obtained from clinical samples from transplant patients.86 Electrophoretic karyotyping

Electrophoretic karyotyping involves the isolation of chromosomes by creating protoplasts, embedding them in agarose plugs and removing the cell membranes and proteins. The agarose plugs are then subjected to gel electrophoresis and the resolved chromosomal bands are visualised by staining with ethidium bromide and UV transillumination. There are several variations on this technique including pulsed field gel electrophoresis (PFGE) which is able to separate larger DNA molecules by changing the direction of the electric field. Larger DNA molecules reorientate slower than smaller molecules, allowing for greater resolution.59 Results obtained by Sullivan et al. [4] using PFGE indicated that the atypical isolates (C. dubliniensis) produced karyotype profiles with nine or 10 chromosome-sized bands (Fig. 6). These differed significantly from profiles obtained for C. albicans. In addition, only the C. dubliniensis isolates produced at least one band smaller than 1 mb. This characteristic band was also observed for all C. dubliniensis isolates studied by Diaz-Guerra et al. [69] and used by Jabra-Rizk et al. [87] to differentiate between C. albicans and C. dubliniensis clinical isolates. It must however be kept in mind that the genomes of both C. albicans and C. dubliniensis contain the Major Repeat Sequence, which is the site for most chromosome reorganisations in both species.88 In fact, these authors reported that C. dubliniensis has a hypervariable karyotype and that no two isolates studied by

9

R. Ells et al.

Figure 6 Electrophoretic karyotypes of Candida isolates. Lanes: 1, C. albicans isolate 132A; 2, C. albicans isolate CM3; 3, C. stellatoidea type II strain ATCC 20408; 4, C. stellatoidea type I strain ATCC 11006; 5-14, C. dubliniensis isolates CM1, CM2, CM4, CM5, CM6, CD33, CD36, CD38, NCPF 3108 and CD41, respectively. Chromosomes from Saccharomyces cerevisiae, used as molecular mass standards (Bio-Rad), are on the left of the figure and the molecular masses indicated are in Mb. Reproduced with permission from Ref. [4].

them had the same karyotype. This reduces the efficacy of karyotyping as a tool to differentiate between these two species.84 Identification based on sequence differences in the ITS regions

Several authors have used the sequence differences in the ITS1 and ITS2 regions of C. albicans and C. dubliniensis to differentiate between these two species. Kurzai et al. [18] amplified and sequenced both these regions and after phylogenetic analyses, two clusters were observed corresponding to the two species. The differences in the ITS2 sequence were exploited by Park et al. [77] to design species-specific fluorescent molecular beacons, which are small nucleic acid stem-loop structures which recognise a 22 nucleotide sequence in the ITS2 region. Each probe contains a quencher and a fluorophore bound to the ends. Upon binding to the target sequence the quencher is separated from the fluorophore and a fluorescent signal is produced. These beacons were used in real-time PCRs with DNA from these two species and were found to be 100% selective and sensitive. Other fluorescent probes, designed to recognise the ITS2 regions of C. albicans and C. dubliniensis, were used by Ellepola et al. [24] in an enzyme immunoassay format (PCR-EIA). In addition, the authors also amplified and sequenced a portion of the 18S rDNA, the entire ITS1 region and a portion of the 5.8S rDNA of C. albicans and C. dubliniensis. These results confirmed those obtained by PCR-EIA. A relatively new technique in identification of clinically

10

relevant Candida species is reverse line hybridisation. This technique involves the use of species-specific DNA probes covalently linked to a membrane.89 For identification of Candida species the probes are designed based on differences in the ITS2 region.90 Target DNA is amplified, labelled and hybridised to the probes. The membrane is incubated in peroxidase-labelled streptavidin conjugate and chemiluminescence is detected. Using this technique, it was possible to identify Candida species including C. albicans and C. dubliniensis simultaneously. Real-time PCR has also been used to differentiate between C. albicans and C. dubliniensis. In one application Candida-specific primers were used to amplify the ITS2 region of the two species after which speciesspecific bioprobes, labelled with Cy5 and biotin, were hybridised to the PCR products.91 Melting peaks were obtained for isolates belonging to the two species. Ten clinical isolates and one reference strain of C. albicans were tested and most isolates had a Tm value of 55 C, however one had a Tm value of 66 C. All the C. dubliniensis isolates and the reference strain had a Tm value of 62 C. According to Somogyvari et al. [92] the use of probes with real-time PCR is very specific, but also expensive. These authors developed a real-time PCR method that does not use probes but performs melting point analyses with the non-specific fluorescent dye, SYBRGreen. These authors amplified the ITS2 region of 249 C. albicans isolates and two reference strains as well as 15 C. dubliniensis isolates and two reference strains and found that C. dubliniensis had a Tm value of 86.2 C (± 0.5 C) while C. albicans had a Tm value of 87.4 C (± 0.3 C) which is sufficiently different for accurate differentiation of the two species. The major advantage of these real-time PCR assays is the relatively short time (2 h) needed to complete them.91,92 A multiplex tandem real-time PCR assay has also been developed for the identification of several pathogenic fungi and yeasts (including C. albicans and C. dubliniensis) directly from blood culture specimens.93 Primers were designed specific to the ITS2 regions of C. albicans and C. dubliniensis and melting point analyses were again used to differentiate between these species. Although the Tm values reported by Lau et al. [93] differ slightly from those reported by Somogyvari et al. [92], the same tendencies were observed with C. dubliniensis having a slightly lower Tm value than C. albicans. Identification based on sequence differences in other rDNA regions

Another region of the 25S rDNA that has been shown to be highly conserved within C. albicans isolates, is the V3

 2009 Blackwell Verlag GmbH • Mycoses 54, 1–16

Candida albicans or Candida dubliniensis

region. Sullivan et al. [4] amplified approximately 600 bp of the V3 region and after sequencing it was found that the sequences from the atypical isolates (C. dubliniensis) differ by approximately 14 nucleotides from those of C. albicans. Gilfillan et al. [94] and Kurzai et al. [18] also amplified and sequenced the V3 region of both species and confirmed that it was possible to differentiate between these two species based on the sequences of this region. Another region of the 25S rDNA used to differentiate between the two species is a region spanning the site of the transposable intron.76 Amplification of this region yielded two different products for C. albicans (either a 450 bp or an 840 bp fragment or both), depending on the genotype of the isolate. All C. dubliniensis isolates produced one product (1080 bp) allowing differentiation between the species. Although determining the sequence of the D1 ⁄ D2 domain has become the most commonly adopted molecular yeast identification method,95 sequencing equipment may not be readily available in all laboratories. Ramos et al. [95], therefore, developed a heteroduplex mobility assay (HMA) using the amplified D1 ⁄ D2 domains. This assay is based on the denaturing and re-annealing of mixtures of the PCR products, followed by polyacrylamide gel electrophoresis. Isolates belonging to the same species have identical or very similar D1 ⁄ D2 sequences. This will result in the formation of re-annealed homoduplexis, which migrate to the bottom of the gel, as well as self-annealed single strands, which produce faint bands at the top of the gel. In cases where PCR products of different species are mixed, imperfect annealing takes place and heteroduplexes are formed. These migrate slower through the gel than homoduplexes. Using this principle it was possible to differentiate between C. albicans and C. dubliniensis. Fluorescent peptide nucleic acid probes (which mimic nucleic acid probes) have been designed by Oliveira et al. [96] for C. albicans and C. dubliniensis. These probes have the advantage of being more hydrophobic and therefore easily diffuse through cell membranes allowing them to be used, without disrupting the cells, in fluorescent in situ hybridisation assays. The C. albicans probe binds to a region of the 26S rDNA and the C. dubliniensis probe binds to a region of the 18S rDNA. Evaluation of these probes showed that they were 100% selective and specific. Identification based on other gene sequence differences

Differences in sequences of other genes have also been exploited. Donnely et al. [97] amplified and sequenced

 2009 Blackwell Verlag GmbH • Mycoses 54, 1–16

the entire ACT1 gene of C. dubliniensis, which encodes Actin 1, and compared it to the ACT1 gene of C. albicans. It was found that although the exons of the two genes are 97.9% identical, the introns are only 83.4% identical. Species specific primers for the ACT1 intron of C. dubliniensis were designed and it was found that only C. dubliniensis isolates yielded the expected 288 bp product. Ma¨hnß et al. [19] used universal primers to amplify a 614 bp fragment, which served as a positive control, as well as the species-specific primers for the ACT1 intron of C. dubliniensis (Fig. 7). Two PCR products (the 614 bp control as well as a 288 bp species-specific product) were amplified in samples containing C. dubliniensis. The same species-specific primers for the ACT1 intron of C. dubliniensis were used by Jewtuchowicz et al. [7], resulting in the amplification of a 288 bp fragment for all C. dubliniensis isolates. Kanbe et al. [98] designed C. dubliniensis speciesspecific primers in the region amplified by degenerate primers. These species-specific primers amplified a single 515 bp product for C. albicans and an 816 bp product for C. dubliniensis, which made it possible to differentiate between them in a mixed DNA sample. Marcos-Arias et al. [58] also used species-specific primers for C. dubliniensis to selectively amplify the 816 bp fragment of the topoisomerase II gene of this species. In addition, differences in the sequence of the pHregulated C. albicans gene, PHR1, and its homologue in C. dubliniensis was used to develop a simple PCR based technique to differentiate between the two species.18,70 When DNA from these two species was amplified with primers based on the coding sequence of the PHR1 gene, the C. albicans isolates generated the expected 1.6 kb fragment corresponding to the entire open reading frame of the PHR1 gene. The PCR

Figure 7 Polymerase chain reaction (PCR) products of Candida dubliniensis and C. albicans. Lane 1, bp marker; lanes 2–12, clinical isolates; lane 14, amplification result for a C. dubliniensis reference strain with amplification restricted to the smaller fragment, lane 13, the 614 bp product of the C. albicans reference strain. Lanes 4 and 8 are positive for the C. dubliniensis indicating 288 bp Actin 1 intron (ACT1) fragment. Reproduced with permission from Ref. [19].

11

R. Ells et al.

with C. dubliniensis DNA as template produced no product. Bautista-Mun˜oz et al. [81] designed several PCR procedures for differentiation between C. albicans and C. dubliniensis based on the secreted aspartyl proteinase genes, SAP1–SAP6, as well as the dipeptyl aminopeptidase (DAP2) gene. The SAP2 and DAP2 based procedures only amplified fragments from C. albicans and not from C. dubliniensis, while the SAP3 based procedure produced products from both species, with C. albicans producing a larger product (172 bp) than C. dubliniensis (134 bp). These differences allowed for the discrimination between these two species. Differences between another C. albicans gene, hwp1, encoding an adhesin and its homologue in C. dubliniensis have also been used to differentiate between these two species.99 These genes were amplified and after separation in an agarose gel, differences in the size of the products were observed. Candida albicans produced a PCR product of 941 bp while C. dubliniensis produced a smaller product of 569 bp. A multiplex real-time PCR assay for the identification of pathogenic Candida species was also developed, using the gene encoding for the RNA subunit of RNase P, the RPR1 gene.100 For the real-time PCR both C. albicans and C. dubliniensis RPR1 genes were amplified and the fluorescent probe, cand-ROX, was incorporated. To differentiate C. albicans from C. dubliniensis this was combined with the fluorescent probe, alb-FAM, which binds only to C. albicans RPR1. All these molecular techniques are capable of discriminating between C. albicans and C. dubliniensis, however certain factors need to be kept in mind. Manarelli & Kurtzman [101] cautioned that even single base mutations in target sequences may dramatically influence the specificity of the reactions and the accuracy of the identification, especially when probes are used. In addition, Soll [84] cautioned that virtually every aspect of a PCR, including the source of DNA polymerase, primer to template ratio, magnesium concentration, and PCR machine manufacturer, might cause variation in results.

Conclusions Although several reliable molecular differentiation techniques are available, the differentiation of C. dubliniensis from C. albicans still poses significant difficulties from both diagnostic and epidemiological perspectives, since most of these studies rely on phenotypic tests to identify yeast species, due to the need for specialised equipment

12

Figure 8 Hierarchical flow chart for phenotypic differentiation between C. albicans and C. dubliniensis.

and knowledge or the high costs involved in performing molecular techniques.18 Figure 8 represents a suggested hierarchical flowchart for phenotypic differentiation between C. albicans and C. dubliniensis. The production of both chlamydospores and rough colonies on Staib agar can be considered diagnostic of C. dubliniensis. The fact that this medium is often used in clinical laboratories for the identification of Cryptococcus neoformans also makes it a good choice as initial identification step. Candida dubliniensis isolates identified using this characteristic should, however, be confirmed using the absence of XYL and MDG assimilation with the API 20C AUX system and ⁄ or PCR techniques. Isolates that produce none or only one of the characteristics (chlamydospore production or rough colonies) on Staib agar should be subjected to additional tests to confirm their identity. Growth at 45 C and ⁄ or growth on hypertonic SDA (6.5% NaCl) are diagnostic of C. albicans and the inability to grow under both these conditions is diagnostic of C. dubliniensis. The Candida dubliniensis genome sequencing project is nearing completion and it is expected that the comparison of the genomes of C. dubliniensis and C. albicans will enhance research into the differences between these two species and may direct research in the development of future diagnostic tests.

 2009 Blackwell Verlag GmbH • Mycoses 54, 1–16

Candida albicans or Candida dubliniensis

References 1 Ramage G, Vande Walle K, Wickes BL, Lopez-Ribot JL. Standardised method for in vitro antifungal susceptibility testing of Candida albicans biofilms. Antimicrob Agents Chemother 2001; 45: 2475–9. 2 Sullivan DJ, Moran GP, Pinjon E et al. Comparison of the epidemiology, drug resistance mechanisms and virulence of Candida dubliniensis and Candida albicans. FEMS Yeast Res 2004; 4: 369–376. 3 Niewerth M, Korting HC. Candida albicans and the principle of opportunism. An essay. Mycoses 2002; 45: 253–8. 4 Sullivan DJ, Westerneng TJ, Haynes KA, Bennett DE, Coleman DC. Candida dubliniensis sp. nov.: phenotypic and molecular characterisation of a novel species associated with oral candidosis in HIV infected individuals. Microbiology 1995; 141: 1507–21. 5 Al-Mosaid A, Sullivan DJ, Salkin IF, Shanley D, Coleman DC. Differentiation of Candida dubliniensis from Candida albicans on Staib agar and Caffeic acid-ferric citrate agar. J Clin Microbiol 2001; 39: 323–7. 6 Gutierrez J, Morales P, Gonzalez MA, Quindos G. Candida dubliniensis, a new fungal pathogen. J Basic Microbiol 2002; 42: 207–27. 7 Jewtuchowicz VM, Mujica MT, Brusca MI et al. Phenotypic and genotypic identification of Candida dubliniensis from subgingival sites in immunocompetent subjects in Argentina. Oral Microbiol Immunol 2008; 23: 505–9. 8 Sullivan DJ, Coleman DC. Candida dubliniensis: characteristics and identification. J Clin Microbiol 1998; 36: 329–34. 9 Ramage G, Vande Walle K, Wickes BL, Lopez-Ribot JL. Biofilm formation by C. dubliniensis. J Clin Microbiol 2001; 39: 3234–40. 10 Jabra-Rizk MA, Falkler WA Jr, Merz WG, Baqui AMAA, Kelley JI, Meiller TF. Retrospective identification and characterisation of Candida dubliniensis isolates among Candida albicans clinical laboratory isolates from human immunodeficiency virus (HIV) infected and non-HIV infected individuals. J Clin Microbiol 2000; 38: 2423–26. 11 Odds FC, Van Nuffel L, Dams G. Prevalence of Candida dubliniensis isolates in a yeast stock collection. J Clin Microbiol 1998; 6: 2869–73. 12 Coleman DC, Sullivan DJ, Bennett DE, Moran GP, Barry HJ, Shanley DB. Candidiasis: the emergence of a novel species, Candida dubliniensis. AIDS 1997; 11: 557–67. 13 Cooke VM, Miles RJ, Price RG, Midgley G, Khamri W, Richardson AC. New chromogenic agar medium for the identification of Candida spp. Appl Environ Microbiol 2002; 68: 3622–27. 14 Perry JD, Freydiere AM. The application of chromogenic media in clinical microbiology. J Appl Microbiol 2002; 103: 2046–55. 15 Sullivan DJ, Moran GP, Donnelly S et al. Candida dubliniensis: An update. Rev Iberoam Micol 1999; 16: 72–76.

 2009 Blackwell Verlag GmbH • Mycoses 54, 1–16

16 Odds FC, Davidson A. ‘‘Room temperature’’ use of CHROMagarTM Candida. Diagn Microbiol Infect Dis 2000; 38: 147–50. ´ lvarez MI, Sua´rez BL, Caicedo LD. Isolation of Candida 17 A dubliniensis for the first time in Cali, Colombia, and its identification with phenotypic methods. Mycopathologia 2009; 167: 19–24. 18 Kurzai O, Korting H-C, Harmsen D et al. Molecular and phenotypic identification of the yeast pathogen Candida dubliniensis. J Mol Med 2000; 78: 521–9. 19 Ma¨hnß B, Stehr F, Sca¨fer W, Neuber K. Comparison of standard phenotypic assays with a PCR method to discriminate Candida albicans and C. dubliniensis. Mycoses 2005; 48: 55–61. 20 Pincus DH, Coleman DC, Pruitt WR et al. Rapid identification of Candida dubliniensis with commercial yeast identification systems. J Clin Microbiol 1999; 37: 3533–39. 21 Ilkit M, Hilmioglu S, Tasbakan M, Aydemir S. Evaluation of Albicans ID2 and Biggy agar for the isolation and direct identification of vaginal yeast isolates. J Med Microbiol 2007; 56: 762–5. 22 Velegraki A, Logotheti M. Presumptive identification of an emerging yeast pathogen: Candida dubliniensis (sp. nov.) reduces 2,3,5,-triphenyltetrazolium chloride. FEMS Immunol Med Microbiol 1998; 20: 239–41. 23 Giammanco GM, Pizzo G, Pecorella S, Distefano S, Pecoraro V, Milici ME. Identification of Candida dubliniensis among oral yeast isolates from an Italian population of human immunodeficiency virus-infected (HIV+) subjects. Oral Microbiol Immunol 2002; 17: 89–94. 24 Ellepola ANB, Hurst SF, Elie CM, Morrison CJ. Rapid and unequivocal differentiation of Candida dubliniensis from other Candida species using species-specific DNA probes: comparison with phenotypic identification methods. Oral Microbiol Immunol 2003; 18: 379–88. 25 Kirkpatrick WR, Revankar SG, McAtee RK et al. Detection of Candida dubliniensis in oropharyngeal samples from human immunodeficiency virus-infected patients in North America by primary CHROMagar Candida screening and susceptibility testing of isolates. J Clin Microbiol 1998; 36: 3007–12. 26 Al-Mosaid A, Sullivan DJ, Coleman DC. Differentiation of Candida dubliniensis from Candida albicans on PalÕs Agar. J Clin Microbiol 2003; 41: 4787–9. 27 Sahand IH, Moragues MD, Eraso E, Villar-Vidal M, Quindo´s G, Ponto´n J. Supplementation of CHROMagar Candida medium with PalÕs medium for rapid identification of Candida dubliniensis. J Clin Microbiol 2005; 43: 5768–70. 28 Khan ZU, Ahmad S, Mokaddas E, Chandy R. Tobacco agar, a new medium for differentiating Candida dubliniensis from Candida albicans. J Clin Microbiol 2004; 42: 4796–98. 29 Kumar CPG, Menon T. Tobacco agar: a new medium for chlamydosporulation in Candida albicans and Candida dubliniensis. Med Mycol 2005; 43: 473–5.

13

R. Ells et al.

30 Staib P, Morschha¨user J. Chlamydospore formation on Staib agar as a species specific characteristic of Candida dubliniensis. Mycoses 1999; 42: 521–4. 31 Wabale VR, Kagal AS, Mani RS, Bharadwaj RA. A novel method for differentiation of Candida dubliniensis from other Candida species. Indian J Med Microbiol 2007; 25: 304–5. 32 Alves SH, de Loreto ES, Linares CE et al. Comparison among tomato juice agar with other three media for differentiation of Candida dubliniensis from Candida albicans. Rev Inst Med Trop Sao Paulo 2006; 48: 119–21. 33 De Loreto ES, Bolzan AR, Linares CEB, Boff E, Santurio JM, Alves SH. Evaluation of 5 new media containing extracts of seeds applied to Candida dubliniensis screening. Diagn Microbiol Infect Dis 2006; 55: 191–3. 34 De Loreto S, Pozzatti P, Scheid LA, Santurio D, Santurio JM, Alves SH. Differentiation of Candida dubliniensis from Candida albicans on rosemary extract agar and oregano extract agar. J Clin Lab Anal 2008; 22: 172–7. 35 Schoofs A, Odds EC, Colebunders R, Ieven M, Goossens H. Use of specialised isolation media for recognition and identification of Candida dubliniensis isolates from HIV infected patients. Eur J Clin Microbiol Infect Dis 1997; 16: 296–300. 36 Yu¨cesoy M, Esen N, Yulug˘ N. Use of chromogenic tube and methyl blue-sabouraud agar for the identification of Candida albicans strains. Kobe J Med Sci 2001; 47: 161–167. 37 Alves SH, Milan EP, de Laet SantÕAna P, Oliveira LO, Santurio JM, Colombo AL. Hypertonic sabouraud broth as a simple and powerful test for Candida dubliniensis. Diagn Microbiol Infect Dis 2002; 43: 85–86. 38 Ahmad S, Khan Z, Mokaddas E, Khan ZA. Isolation and molecular identification of Candida dubliniensis from non-human immunodeficiency virus-infected patients in Kuwait. J Med Microbiol 2004; 53: 633–7. 39 Ahmad S, Mokaddas E, Al-Sweih N, Khan ZU. Phenotypic and molecular characterization of Candida dubliniensis isolates from clinical specimens in Kuwait. Med Princ Pract 2005; 14(Suppl 1): 77–83. 40 Pinjon E, Sullivan D, Salkin I, Shanley D, Coleman D. Simple, inexpensive, reliable method for differentiation of Candida dubliniensis from Candida albicans. J Clin Microbiol 1998; 36: 2093–5. 41 Gales AC, Pfaller MA, Houston AK et al. Identification of Candida dubliniensis based on temperature and utilization of xylose and a-methyl-D-glucoside as determined with the API 20C AUX and Vitek YBC systems. J Clin Microbiol 1999; 37: 3804–8. 42 Loı¨ez C, Wallet F, Sendid B, Courcol RJ. Evaluation of VITEK 2 colorimetric cards versus fluorimetric cards for identification of yeasts. Diagn Microbiol Infect Dis 2006; 56: 455–7. 43 Ca´rdenes-Perera C-D, Torres-Lana A, Alonso-Vargas R et al. Evaluation of API ID 32C and Vitek-2 to identify

14

44

45

46

47

48

49

50

51

52

53

54

55

56

Candida dubliniensis. Diagn Microbiol Infect Dis 2004; 50: 219–21. Alves SH, Horta JA, Milan EP et al. Carbohydrate assimilation profiles of Brazilian Candida dubliniensis isolates based on ID 32C system. Rev Inst Med Trop Sao Paulo 2005; 47: 109–111. Szabo´ Z, To´th B, Kova´cs M et al. Evaluation of the new Micronaut-Candida system compared to the API ID 32C method for yeast identification. J Clin Microbiol 2008; 46: 1824–25. Tintelnot K, Haase G, Seibold M et al. Evaluation of phenotypic markers for selection and identification of Candida dubliniensis. J Clin Microbiol 2000; 38: 1599–1608. Campanha NH, Neppelenbroek KH, Spolidorio DMP, Spolidorio LC, Pavarina AC. Phenotypic methods and commercial systems for the discrimination between C. albicans and C. dubliniensis. Oral Disease 2005; 11: 392–8. Al-Sweih N, Ahmad S, Khan ZU, Khan S, Chandy R. Prevalence of Candida dubliniensis among germ tubepositive Candida isolates in a maternity hospital in Kuwait. Mycoses 2005; 48: 347–51. Coleman D, Sullivan D, Harrington B et al. Molecular and phenotypic analysis of Candida dubliniensis: a recently identified species linked with oral candidosis in HIVinfected and AIDS patients. Oral Disease 1997; 3(Suppl 1): S96–S101. Bikandi J, Millan RS, Moragues MD et al. Rapid identification of Candida dubliniensis by indirect immunofluorescence based on differential localization of antigens on C. dubliniensis blastospores and C. albicans germ tubes. J Clin Microbiol 1998; 36: 2428–33. Moragues MD, Omaetxebarria MJ, Elguezabal N et al. Serological differentiation of experimentally induced Candida dubliniensis and Candida albicans infections. J Clin Microbiol 2001; 39: 2999–3001. Bliss JM, Sullivan MA, Malone J, Haidaris CG. Differentiation of Candida albicans and Candida dubliniensis by using recombinant human antibody single-chain variable fragments specific for hyphae. J Clin Microbiol 2003; 41: 1152–60. Haidaris CG, Malone J, Sherrill LA et al. Recombinant human antibody single chain variable fragments reactive with Candida albicans surface antigens. J Immunol Methods 2001; 257: 185–202. Marot-Leblond A, Grimaud L, David S et al. Evaluation of a rapid immunochromographic assay for identification of Candida albicans and Candida dubliniensis. J Clin Microbiol 2004; 42: 4956–60. Marot-Leblond A, Beucher B, David S, Nail-Billaud S, Robert R. Development and evaluation of a rapid latex agglutination test using a monoclonal antibody to identify Candida dubliniensis colonies. J Clin Microbiol 2006; 44: 138–42. Sahand IH, Moragues MD, Robert R, Quindos G, Ponton J. Evaluation of Bichro-Dubli Fumouze to distinguish

 2009 Blackwell Verlag GmbH • Mycoses 54, 1–16

Candida albicans or Candida dubliniensis

57

58

59

60 61

62

63

64

65

66

67

68

69

70

Candida dubliniensis from Candida albicans. Diagn Microbiol Infect Dis 2006; 55: 165–67. Chryssanthou E, Fernandez V, Petrini B. Performance of commercial latex agglutination tests for the differentiation of Candida dubliniensis and Candida albicans in routine diagnostics. AMPIS 2007; 115: 1281–4. Marcos-Arias C, Vicente JL, Sahand IH. Isolation of Candida albicans in denture stomatitis. Arch Oral Biol 2009; 54: 127–131. Neppelenbroek KH, Campanha NH, Spolidorio DMP, Spolidorio LC, Seo´ RS, Pavarina AC. Molecular fingerprinting methods for the discrimination between C. albicans and C. dubliniensis. Oral Disease 2006; 12: 242–53. Cunningham MJ, Noble WC. SDS-PAGE protein patterns of yeasts from human sources. Mycoses 1989; 32: 344–8. Rosa EAR, Rosa RT, Pereira CV, Boriollo MFG, Ho¨fling JF. Analysis of parity between protein-based electrophoretic methods for the characterization of oral Candida species. Mem Inst Oswaldo Cruz 2000; 95: 801–6. Rosa EAR, Rosa RT, Boriollo MFG, Bernardo WLC, Ho¨fling JF. Oral Candida albicans and Candida dubliniensis differentiation by multilocus enzyme electrophoresis and sodium dodecylsulphate-polyacrylamide gel electrophoresis. Rev Argent Microbiol 2003; 35: 24–28. Boerlin P. Applications of multilocus enzyme electrophoresis in medical microbiology. J Microbiol Methods 1997; 28: 221–31. Gil-Lamaignere C, Roilides E, Hacker J, Mu¨ller F-MC. Molecular typing for fungi – a critical review of the possibilities and limitations of currently and future methods. Clin Microbiol Infect 2003; 9: 172–185. Boerlin P, Boerlin-Petzold F, Durussel C et al. Cluster of oral atypical Candida albicans isolates in a group of human immunodeficiency virus positive drug users. J Clin Microbiol 1995; 33: 1129–35. Timmins EM, Howell SA, Alsberg BK, Noble WC, Goodacre R. Rapid differentiation of closely related Candida species and strains by pyrolysis-mass spectrometry and Fourier transform-infrared spectroscopy. J Clin Microbiol 1998; 36: 367–74. Maquelin K, Kirschner C, Choo-Smith L-P et al. Prospective study of the performance of vibrational spectroscopies for rapid identification of bacterial and fungal pathogens recovered from blood cultures. J Clin Microbiol 2003; 41: 324–9. Tietz H-J, Hopp M, Schmalreck A, Sterry W, Czaika V. Candida africana sp. nov., a new human pathogen or a variant of Candida albicans? Mycoses 2001; 44: 437–45. Diaz-Guerra TM, Mellado E, Estrella MC, Laguna F, Rodriguez-Tudela JL. Molecular characterization by PCRfingerprinting of Candida dubliniensis strains isolated from two HIV-positive patients in Spain. Diagn Microbiol Infect Dis 1999; 35: 113–9. Kurzai O, Heinz WJ, Sullivan DJ, Coleman DC, Frosch M, Mu¨hlschlegel FA. Rapid PCR test for discriminating

 2009 Blackwell Verlag GmbH • Mycoses 54, 1–16

71

72

73

74

75

76

77

78

79

80

81

82

83

between Candida albicans and Candida dubliniensis isolates using primers derived from the pH-regulated PHR1 and PHR2 genes of C. albicans. J Clin Microbiol 1999; 37: 1587–90. Joly S, Pujol C, Rysz M, Vargas K, Soll DR. Development and characterization of complex DNA fingerprinting probes for the infectious yeast Candida dubliniensis. J Clin Microbiol 1999; 37: 1035–44. Martinez M, Lopez-Ribot JL, Kirkpatrick WR, Coco BJ, Bachmann SP, Patterson TF. Replacement of Candida albicans with C. dubliniensis in Human Immunodeficiency Virus-infected patients with oropharyngeal candidiasis treated with fluconazole. J Clin Microbiol 2002; 40: 3135–9. Blignaut E, Pujol C, Joly S, Soll DR. Racial distribution of Candida dubliniensis colonisation among South Africans. J Clin Microbiol 2003; 41: 1838–42. Fujita S-I, Hashimoto T. DNA fingerprinting patterns of Candida species using HinfI endonuclease. Int J Syst Evol Microbiol 2000; 50: 1381–9. Irobi J, Schoofs A, Goossens H. Genetic identification of Candida species in HIV-positive patients using the polymerase chain reaction and restriction fragment polymorphism analysis of its DNA. Mol Cell Probes 1999; 14: 401–6. McCullough MJ, Clemons KV, Stevens DA. Molecular and phenotypic characterization of genotypic Candida albicans subgroups and comparison with Candida dubliniensis and Candida stellatoidea. J Clin Microbiol 1999; 37: 417–21. Park S, Wong M, Marras SAE et al. Rapid identification of Candida dubliniensis using a species-specific molecular probe. J Clin Microbiol 2000; 38: 2829–36. Graf B, Trost A, Eucker J, Go¨bel UB, Adam T. Rapid and simple differentiation of C. dubliniensis from C. albicans. Diagn Microbiol Infect Dis 2004; 48: 149–51. Ruhnke M, Grosch-Worner I, Lischewski A et al. Genotypic relatedness of yeast isolates from women infected with human immunodeficiency virus and their children. Mycoses 1999; 42: 385–94. Mariano P, Milan EP, da Matta DA, Colombo AL. Candida dubliniensis identification in Brazilian yeast stock collection. Mem Inst Oswaldo Cruz 2003; 98: 533–8. Bautista-Mun˜oz C, Boldo XM, Villa-Tanaca L, Hernandez-Rodriguez C. Identification of Candida spp. by randomly amplified polymorphic DNA analysis and differentiation between Candida albicans and Candida dubliniensis by direct PCR methods. J Clin Microbiol 2003; 41: 414–20. Rocha BA, Barbaro del Negro GM, Yamamoto L, de Souza MVB, Precioso AR, Okay TS. Identification and differentiation of Candida species from paediatric patients by random amplified polymorphic DNA. Rev Soc Bras Med Trop 2008; 41: 1–5. Barros LM, Boriollo MFG, Alves ACBA, Klein MI, Gonc¸alves RB, Ho¨fling JF. Genetic diversity and

15

R. Ells et al.

84 85

86

87

88

89

90

91

92

16

exoenzyme activities of Candida albicans and Candida dubliniensis isolated from the oral cavity of Brazilian periodontal patients. Arch Oral Biol 2008; 53: 1172–78. Soll DR. The ins and outs of DNA fingerprinting the infectious fungi. Clin Microbiol Rev 2000; 13: 332–70. Borst A, Theelen B, Reinders E, Boekhout T, Fluit AC, Savelkou PHM. Use of amplified fragment length polymorphism to identify medically important Candida spp., including C. dubliniensis. J Clin Microbiol 2003; 41: 1357–62. Ball LM, Bes MA, Theelen B, Boekhout T, Egeler RM, Kuijper EJ. Significance of amplified fragment length polymorphism in identification and epidemiological examination of Candida species colonization in children undergoing allogenic stem cell transplantation. J Clin Microbiol 2004; 42: 1673–79. Jabra-Rizk MA, BAqui AAMA, Kelley JI, Falkler WA Jr, Merz WG, Meiller TF. Identification of Candida dubliniensis in a prospective study in the United States. J Clin Microbiol 1999; 37: 321–6. Magee BB, Sanchez MD, Saunders D, Harris D, Berriman M, Magee PT. Extensive chromosome rearrangements distinguish the karyotype of the hypovirulent species Candida dubliniensis from the virulent Candida albicans. Fungal Genet Biol 2008; 45: 338–50. Zwart G, van Hannen EJ, Kamst-van Agterveld MP et al. Rapid screening for freshwater bacterial groups by using reverse line hybridization. Appl Environ Microbiol 2003; 69: 5875–83. Xiang H, Xiong L, Liu X, Tu Z. Rapid simultaneous detection and identification of six species Candida using polymerase chain reaction and reverse line hybridization assay. J Microbiol Methods 2007; 69: 282–87. Schabereiter-Gurtner C, Selitsch B, Rotter ML, Hirschl AM, Willinger B. Development of novel real-time PCR assays for detection and differentiation of eleven medically important Aspergillus and Candida species in clinical specimens. J Clin Microbiol 2007; 45: 906–14. Somogyvari F, Doczi I, Serly J, Ahmad S, Nagy E. Rapid discrimination between Candida albicans and Candida

93

94

95

96

97

98

99

100

101

dubliniensis by using real-time polymerase chain reaction. Diagn Microbiol Infect Dis 2007; 58: 367–9. Lau A, Sorrell TC, Chen S, Stanley K, Iredell J, Halliday C. Multiplex tandem PCR: a novel platform for rapid detection and identification of fungal pathogens from blood culture specimens. J Clin Microbiol 2008; 46: 3021–7. Gilfillan GD, Sullivan DJ, Haynes K, Parkinson T, Coleman DC, Gow NAR. Candida dubliniensis: phylogeny and putative virulence factors. Microbiology 1998; 144: 829–38. Ramos JP, Rosa CA, Carvalho EMM, Leoncini O, Valente P. Heteroduplex mobility assay of the 26S rDNA D1 ⁄ D2 region for differentiation of clinically relevant Candida species. Antonie Van Leeuwenhoek 2006; 89: 39–44. Oliveira K, Haase G, Kurtzman C, Hydlig-Nielsen JJ, Stender H. Differentiation of Candida albicans and Candida dubliniensis by fluorescent in situ hybridization with peptide nucleic acid probes. J Clin Microbiol 2001; 39: 4138–41. Donnely SM, Sullivan DJ, Shanley DB, Coleman DC. Phylogenetic analysis and rapid identification of Candida dubliniensis based on analysis of ACT1 intron and exon sequences. Microbiology 1999; 145: 1871–82. Kanbe T, Horii T, Arishima T, Ozeki M, Kikuchi A. PCRbased identification of pathogenic Candida species using primer mixes specific to Candida DNA topoisomerase II genes. Yeast 2002; 19: 973–89. Romeo O, Criseo G. First molecular method for discriminating between Candida africana, Candida albicans and Candida dubliniensis by using hpw1 gene. Diagn Microbiol Infect Dis 2008; 62: 230–3. Innings A, Ullberg M, Johansson A et al. Multiplex realtime PCR targeting the RNase P RNA gene for detection and identification of Candida species in blood. J Clin Microbiol 2007; 45: 874–880. Manarelli BM, Kurtzman CP. Rapid identification of Candida albicans and other human pathogenic yeasts by using short oligonucleotides in a PCR. J Clin Microbiol 1998; 36: 1634–41.

 2009 Blackwell Verlag GmbH • Mycoses 54, 1–16