Updating the Taxonomy of Dermatophytes of the BCCM/ IHEM Collection According to the New Standard: A Phylogenetic Approach

Updating the Taxonomy of Dermatophytes of the BCCM/IHEM Collection According to the New Standard: A Phylogenetic ApproachORIGINAL ARTICLE
Updating the Taxonomy of Dermatophytes of the BCCM/ IHEM Collection According to the New Standard: A Phylogenetic Approach
F. Baert . D. Stubbe . E. D’hooge . A. Packeu . M. Hendrickx
Received: 23 January 2019 / Accepted: 30 April 2019
Springer Nature B.V. 2019
clarifications to dermatophyte (Arthrodermataceae)
dermatophyte strains in the BCCM/IHEM fungal
collection, which contains strains of all dermatophyte
genera except for Ctenomyces. For 688 strains from
this collection, both internal transcribed spacer region
(ITS) and partial b-tubulin (BT) sequences were
aligned and a multilocus phylogenetic tree was
constructed. The ITS ? BT phylogentic tree was able
to distinguish the genera Arthroderma, Lophophyton,
Microsporum, Paraphyton, Nannizzia and Trichophy-
ton with high certainty. Epidermophyton, which is
widely considered as a well-defined genus with E.
floccosum as the only representative, fell within the
Nannizzia clade, whereas the phylogenetic analysis,
based on the ITS region alone, differentiates Epider-
mophyton from Nannizzia as a separate genus. Re-
identification and reclassification of many strains in
the collection have had a profound impact on the
composition of the BCCM/IHEM dermatophyte col-
lection. The biggest change is the decline of preva-
lence of Arthroderma strains; starting with 103 strains,
only 22 strains remain in the genus after reassessment.
Most Arthroderma strains were reclassified into Tri-
chophyton, with A. benhamiae and A. van-
breuseghemii leaving the genus. The amount of
Microsporum strains also dropped significantly with
most of these strains being reclassified into the genera
Paraphyton and Nannizzia.
Introduction
characterized by their capacity to invade and infect
keratinized tissues [1, 2]. These infections of the skin,
Handling Editor: Sybren de Hoog.
Electronic supplementary material The online version of this article (https://doi.org/10.1007/s11046-019-00338-7) con- tains supplementary material, which is available to authorized users.
F. Baert (&) D. Stubbe E. D’hooge A. Packeu M. Hendrickx
Sciensano, Service of Mycology and Aerobiology,
Brussels, Belgium
e-mail: [email protected]
BCCM/IHEM Fungal Collection, Sciensano, Service of
Mycology and Aerobiology, Brussels, Belgium
123
Mycopathologia
https://doi.org/10.1007/s11046-019-00338-7(0123456789().,-volV)( 0123456789().,-volV)
https://doi.org/10.1007/s11046-019-00338-7
commonly known as tinea or ringworm. Diseases
caused by this group of fungi have a worldwide
distribution in both humans and animal hosts [3].
Since the mid-nineteenth century, the taxonomy of
the dermatophytes has extensively been investigated
and was based on both their clinical symptoms and
in vitro morphological characteristics. Since this
classification was phenotype-based, mutants or mor-
photypes of single species were often mistaken for
separate species. Moreover, due to the lack of
reference strains, many species were described sepa-
rately by several independent researchers introducing
many synonymic taxa to this specific fungal family
[4]. Historically, more than 300 names were at some
point defined for the 54 species of dermatophytes that
exist according to the latest taxonomic reports [5].
Since the 1980’s, molecular techniques have been
used for the identification of fungal strains, providing
valuable insights into the taxonomy of the family.
Phylogenetic analysis using different molecular mark-
ers has clarified much of the dermatophyte taxonomy,
although mainly the rDNA internal transcribed spacer
(ITS) region is informative [6–9]. The dermatophytes
are a closely related and highly conserved family of
fungi, so while the molecular approach has been very
successful at determining the main clades of the
dermatophyte family, problematic species complexes
such as Trichophyton rubrum and Trichophyton men-
tagrophytes still exist [7, 10]. Therefore, a combina-
tion of different approaches including ecological,
morphological, clinical and molecular data is recom-
mended for the delineation of species boundaries.
The use of biological sequence data for phyloge-
netic inference has shown that the dermatophytes can
be grouped more accurately based on clinical and
ecological traits than based on morphological charac-
teristics [11]. Ecologically speaking, the dermato-
phytes can be divided into three broad groups, namely
the anthropophilic, zoophilic and geophilic species.
This is reflected in the main topology of the Arthro-
dermataceae family [5, 6]. Geophilic species are
separated from the other taxa and located in the
ancestral position of the phylogenetic tree, zoophilic
species are found scattered throughout the middle of
the tree, while the anthropophilic species can be found
in the derived clades. Important to note is that the
morphology of these species does not correspond to
the topology of the family, and some species that can
be found in ancestral clades are morphologically close
to anthropohilic species found in a derived position in
the phylogeny. Because geophilic, anthropophilic and
zoophilic dermatophytes also differ clinically, it is
important to accurately distinguish these species. An
infection by a geophilic species is most often highly
inflammatory but quickly resolved, an infection by a
zoophile is moderately inflammatory and self-limit-
ing, while anthropophilic dermatophytes cause mild,
non-inflammatory and chronic infections [5, 11].
De Hoog et al. have formalized the dermatophyte
topology in their recent paper, where seven main
clades of dermatophytes were identified and judged to
represent genera using a multilocus phylogenetic
approach [5]. The distinguished genera are Arthro-
derma, Microsporum, Lophophyton, Nannizzia, Epi-
dermophyton, Paraphyton and Trichophyton. In this
classification, most geophilic species can be found in
the ancestral Arthroderma genus, while the anthro-
pophiles are almost exclusively located within the
more derived Epidermophyton and Trichophyton
genera.
veterinary fungi holds more than 2000 different strains
of dermatophytes and is thus a valuable resource in
dermatophyte research. The major taxonomical
changes and developments in this family demanded
a thorough revision of the taxonomy and species re-
identification of the BCCM/IHEM strains. Firstly, the
analysis of this large dataset allows to assess whether
taxonomical changes proposed by de Hoog et al. are
confirmed, and secondly, the impact of these taxo-
nomical changes can be evaluated [5]. We aim to re-
evaluate the nomenclature of the collection strains
based on their phylogenetic characteristics. No mor-
phological, physiological, mating or data of the origin
of the strains were taken into account.
M&M
In this study, a total of 688 strains of the BCCM/IHEM
collection were analyzed. All sequences used in this
study were uploaded to genbank (see Online Resource
1). Since for some species type material was not
available in the collection, type sequences from the
reference collection of the Westerdijk Fungal Biodi-
versity Institute were included when available. In
123
Mycopathologia
rated (see Online Resource 2).
Strains were cultured on Sabouraud dextrose broth
for at least 5 days. Genomic DNA of the strains was
extracted using the Invisorb Spin Plant Mini Kit
(Invitek, Berlin, Germany). The extraction kit was
used according to the manufacturer’s instructions,
with some adaptations: (1) Before lysis, a lyophiliza-
tion step and subsequent bead beating was added to
facilitate the disruption of the fungal cell wall and (2)
the lysis time was raised to more than 2 h.
Two gene regions of the genomic DNA were
amplified and sequenced: (1) The primers Bt2b and
Bt2a described by Glass and Donaldson [12] were
used for amplification and sequencing of the partial b- tubulin (BT) gene and (2) the ITS region was amplified
using primers ITS5 and ITS4 [13]. BT was chosen as a
secondary region since it has been shown to provide
the highest resolution when determining clades in the
dermatophyte family when choosing among the most
commonly used markers, with the exception of
barcoding region ITS [5, 14].
PCR amplicons were purified using ExoSAP-IT
PCR Product Cleanup (Affymetrix, Santa Clara, CA,
USA). Sanger sequencing was performed with an ABI
3130xl Genetic Analyser (Applied Biosystems, Wal-
tham, MA, USA).
genetic analysis was performed involving all 688
strains (see Online Resource 3 for full phylogeny). The
multiple sequence alignment was constructed with
MAFFT version 7.394 using the FFT-NS-i iterative
refinement method. The scoring matrix for nucleotide
sequences was set to 1PAM/j = 2. Afterward, the
alignment was manually assessed and checked for
inconsistencies. On the basis of this dataset, a max-
imum likelihood (ML) phylogeny was constructed
using RAxML-HPC Blackbox version 8.2.1 hosted on
the CIPRES Science Gateway [15]. Branch support
was calculated using 1000 bootstrap replicates. The
dataset was subdivided into 5 gene partitions: ITS1 ?
ITS2, 5.8S ? 28S, BT introns codon 1 and 2, BT
introns codon 3 and BT exons. For each partition, the
GTRCAT approximation of rate heterogeneity was
estimated.
Results
shows strong support for Arthroderma, Paraphyton,
Lophophyton, Trichophyton and Microsporum. The
relationship between Epidermophyton and Nannizzia
remains unresolved. In the obtained topology, Epi-
dermophyton is nested within theNannizzia genus, in a
clade which receives high bootstrap support.
Epidermophyton is considered a well-defined genus
with E. floccosum as its only representative. However,
in the ITS-BT analysis, the species fell within the
Nannizzia clade. Phylogenetic analysis based solely
on the ITS gene region does distinguish Epidermo-
phyton from Nannizzia as a separate clade, but adding
BT information to the analysis casted doubt on this
classification.
are well supported throughout the entire genus. Most
species of this genus were formerly classified in
Microsporum. In the IHEM collection, the strains of 6
species are now transferred from Microsporum to
Nannizzia as a result of the new taxonomy: N. praecox
(= M. praecox), N. gypsea (= M. gypseum), N. nana
(= M. nanum), N. persicolor (= M. persicolor), N.
fulva (= M. fulvum), N. duboisii (= M. duboisii) and N.
incurvata (= M. incurvatum). The genus is also home
to N. aenygmatica [5], but no sequences for this
species were included in the dataset. Arthroderma
corniculatum should also be transferred to Nannizzia,
N. corniculata. Some strains of A. racemosum,
considered a synonym of P. cookei by De Hoog
et al., from the BCCM/IHEM collection formed a fully
supported clade together with the type strain (CBS
450.65) within the Nannizzia clade [5].
Within the genus Arthroderma, all examined
species are well resolved. It is the genus with the
highest interspecific variability in the Arthrodermat-
aceae. The dataset did not contain any strains for the
species A. redelli and did not contain the type
sequence for A. quadrifudum.
of three species. The genus is fully supported in this
study, as are the species clades P. mirabile, P.
cookiellum and P. cookei. The sequences of the type
strain of P. cookei (CBS 228.58) were included in our
analyses. Arthroderma cajetani is confirmed as a
synonym of P. cookei.
canis clade received weak support in contrast to M.
audouinii. Two strains, currently designated as A. otae
within the BCCM/IHEM collection, did not cluster
with either of these two clades nor cluster with each
other. Since no BT sequences were available for M.
ferrugineum, the species is not included in the
multilocus analysis, but the species is clearly distin-
guished and highly supported in the single locus ITS
analysis.
majority of analyzed strains in this study. The genus
consists of eighteen closely related species that are all
represented in this study. Their relationship is not
always resolved in this multilocus analysis. Tri-
chophyton tonsurans and T. equineum are phyloge-
netically indistinguishable; similar situation is
observed between T. quinckeanum and the T. schoen-
leinii. The T. benhamiae species clade is also not
resolved, with a group of 23 strains clustering with T.
concentricum, while a minority of eight strains cluster
with T. bullosum. The species clades T. simii, T.
interdigitale, T. eriotrephon, T. verrucosum and T.
erinacei were well resolved. The T. rubrum complex
contains T. violaceum, T. yaoundei, T. kuryangei, T.
soudanense and T. rubrum. Trichophyton rubrum
strains cluster together with high support. Strains
belonging to the other species in this complex, group
in paraphyletic clades with inconclusive support.
By applying the recently proposed changes in
dermatophyte taxonomy by de Hoog et al. [5], the
composition of the dermatophyte strains in the
BCCM/IHEM collection has changed significantly
(see Fig. 2). The majority of name changes is seen in
the genera Arthroderma andMicrosporum. In total, 81
out of 103 Arthroderma strains are reclassified to five
different genera, leaving only 22 strains in Arthro-
derma (see Fig. 3). The number of strains in Tri-
chophyton is largely unchanged, and the amount of
Epidermophyton strains remains the same.
Figure 4 shows the distribution of the origin of the
analyzed strains, classified into the categories human,
animal and environmental. Arthroderma has the
highest proportion of strains from environmental
origin. Paraphyton also shows a high proportion of
environmental strains as two out of three representa-
tives are geophylic, namely P. cookei and P. cookeil-
lum. The proportion of strains from human infection is
visibly higher in the more derived clades. Strains from
veterinary origin are found in each genus except
Epidermophyton. Lophophyton contains only one
species, L. gallinae, which is zoophilic on poultry,
and thus, almost all strains were of animal origin.
Discussion
strong support for the major clades of the Arthroder-
matacae family. Combining these genes into one
multilocus analysis provided a phylogenetic tree with
more significant support than using any one gene
separately. The backbone of the dermatophyte phy-
logeny is firmly supported with bootstrap values above
85 for all internal nodes. The genera Arthroderma,
Fig. 1 Maximum likelihood analysis of the concatenated dataset of ITS and BT sequences of 688 BCCM/IHEM strains and reference
strains. Bootstrap support is mentioned below each branch
123
Mycopathologia
Trichophyton are all well supported, which is in line
with the findings of recent publications [5, 14]. The
tree topologies of the single and multiple locus
analyses did not show any conflicts on the genus
level, except for the position of the genus Epidermo-
phyton relative to Nannizzia. ITS analysis alone
indicates that Epidermophyton is a sister genus of
Nannizzia, while based on addition of the BT gene, E.
floccosum (the only representative of the genus)
1
10
100
1000
revision of the taxonomy
molecular phylogeny.
Logarithmic scale
Arthroderma Epidermophyton Microsporum Trichophyton Trichophyton 59.55% 0.00% 0.00% 97.98% Paraphyton 4.49% 0.00% 8.82% 0.00% Nannizzia 17.98% 0.00% 29.41% 0.00% Microsporum 2.25% 0.00% 56.86% 0.00% Lophophyton 1.12% 0.00% 4.90% 0.00% Epidermophyton 0.00% 100.00% 0.00% 0.00% Arthroderma 14.61% 0.00% 0.00% 2.02%
0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 60.00% 70.00% 80.00% 90.00%
100.00%
Name in old classificaon
Fig. 3 Depiction of the proportional flow of strains in the BCCM/IHEM collection for each genus from the old classification to the new
classification of genera as proposed by de Hoog et al. [5].
123
Mycopathologia
Rezaei-Matehkolaei et al., who analyzed 54 BT
sequences and found E. floccosum to be closely
related to species now comprised in Nannizzia [8]. A
similar result was also obtained using the genes for
calmodulin [16] and translation elongation factor
1 - a [17]. Seemingly, the protein-encoding genes
paint a different picture regarding the classification of
E. floccosum than the non-encoding barcoding region,
ITS. Since protein-encoding genes are under more
selective pressure and thus less variable, ITS is most
suited to delineate these closely related genera. Use of
other non-encoding and variable regions could provide
more definitive insights into the exact phylogeny.
An interesting case is that ofM. racemosum. Based
on syntype strains CBS 423.74 and CBS 424.74, de
Hoog et al. concluded this name was a synonym for P.
cookei [5]. There exists, however, a type strain IHEM
21235 (= CBS 450.65 = ATCC 16135 = IMI
128984) isolated by Borelli and which has different
ITS and BT sequences. Rush-Munro et al. [18]
originally introduced IMI 135823 (= CBS
424.74 = ATCC 18911) and IMI 135822 (= CBS
423.74 = ATCC 18910) as single spore isolates of IMI
128984 representing both mating types in 1970, but
did not mention them being syntypes. In our study, it is
clear that the type strain (IHEM 21235) belongs to
Nannizzia, and that N. racemosa cannot be considered
a synonym of P. cookei, but should be accepted as a
separate species.
olution on the species level was low. Although the
examined M. audouinii and M. canis strains formed
separate clades, the 2 strains previously designated to
M. otae (IHEM 26531, IHEM 04242) did not fall
within one of these clades. Intraspecies variation
between the strains previously named M. otae,
amounting to a difference of 6 basepairs, is close to
the number of base differences per sequence from
averaging over all sequence pairs between strains
previously named M. otae and M. canis/M. audouinii
(amounting to an average of 6 and 9 base differences,
respectively). This problem hindered species delin-
eation within Microsporum.
differ on an ecological basis, with only one or two
different nucleotides, T. tonsurans and T. equinum
were phylogenetically indistinguishable in this study.
There exists, however, a C/T polymorphism at posi-
tion 18 in the ITS gene that can be used to distinguish
these species [19]. This difference was present in the
strains analyzed here, but proved insufficient to
distinguish them in the phylogenetic tree. The BT
sequences showed no interspecific variation, in con-
trast to the observations of Rezaei-Matehkolaei [8],
detecting a difference of one nucleotide. The same
situation was observed for T. quinckeanum and T.
schoenleinii; as previously observed by Beguin et al.
[20], their ITS sequences only differ by 3 basepairs,
namely a GC insertion at position 533 and a C/T
0% 10% 20% 30% 40% 50% 60% 70% 80% 90%
100%
Environmental
Animal
Human
environmental, animal-
completely identical. These differences are conse-
quent though, so they are sufficient to differentiate
between the 2 species, in addition to their morpho-
logical and clinical differences [20]. The exact topol-
ogy of the clade remained unclear as the more basal
nodes were only weakly supported.
Multigene phylogenetic analysis by Suh et al.
subdivided T. benhamiae into 3 phylogroups, one of
which turned out to be very distinct from the other two
and consisted of 2 African strains of T. benhamiae
[21]. This phylogroup was judged to represent a novel
species separate from the typical T. benhamiae strains.
Our analysis corroborates this claim: while most T.
benhamiae strains clustered together with T. concen-
tricum as a subclade in the ITS ? BT phylogeny, all
African strains plus some strains of European origin
clustered with T. bullosum forming a highly supported
clade.
highly supported clades even though the average
interspecies variation amounted to only 7 SNP’s.
While the species T. soudanense, T. kuryangei and T.
yaoundei are closely related to both T. rubrum and T.
violaceum, there are few though distinct SNP’s
differentiating these taxa. However, these differences
were insufficient to produce monophyletic clades in
the phylogenetic analysis, rather T. soudanense, T.
kuryangei and T. yaoundei appear to reside in
paraphyly with T. violaceum.
dermatophyte collection has changed significantly
(see Figs. 2, 3). The decline of the amount of
Arthroderma strains is remarkable, but mainly caused
by reclassification of some important species such as
A. vanbreuseghemii and A. benhamiae into Trichophy-
ton, A. fulvum and A. corniculatum intoNannizzia, etc.
In total 79% of Arthroderma strains were reclassified
to a different genus, leaving only 22 of 103 analyzed
strains in the Arthroderma genus. The number of
strains in Microsporum also declined by 45%, mainly
due to reclassification to the genera Nannizzia and
Paraphyton (Fig. 3). The number of strains in Tri-
chophyton remained largely unchanged, while the
amount in Epidermophyton remained exactly the
same. The collection does not contain any sequences
for strains of the genera Guarromyces and
Ctenomyces.
dermatophytes remains limited for frequently encoun-
tered clinical species. The nomenclature for T.
rubrum, T. tonsurans,M. canis andM. audouinii stays
unchanged, while the position of some closely related
species is more clear. As mentioned by de Hoog and
confirmed by the present study, anthropophilic strains
causing low-inflammatory infections correspond with
T. interdigitale while more inflammatory human
infections are caused by isolates of the zoophilic T.
mentagrophytes s. st. [5]. Another important change
for the routine identification of dermatophytes is the
reclassification ofA. benhamiae into T. benhamiae and
of A. vanbreuseghemii into T. mentagrophytes.
In conclusion, analyzing the ITS and BT genes of
the BCCM/IHEM strains provides great support for
the taxonomic changes proposed by De Hoog et…