OUP user menu

Purpureocillium, a new genus for the medically important Paecilomyces lilacinus

Jennifer Luangsa-ard, Jos Houbraken, Tineke van Doorn, Seung-Beom Hong, Andrew M. Borman, Nigel L. Hywel-Jones, Robert A. Samson
DOI: http://dx.doi.org/10.1111/j.1574-6968.2011.02322.x 141-149 First published online: 1 August 2011


Paecilomyces lilacinus was described more than a century ago and is a commonly occurring fungus in soil. However, in the last decade this fungus has been increasingly found as the causal agent of infections in man and other vertebrates. Most cases of disease are described from patients with compromised immune systems or intraocular lens implants. In this study, we compared clinical isolates with strains isolated from soil, insects and nematodes using 18S rRNA gene, internal transcribed spacer (ITS) and partial translation elongation factor 1-α (TEF) sequences. Our data show that P. lilacinus is not related to Paecilomyces, represented by the well-known thermophilic and often pathogenic Paecilomyces variotii. The new genus name Purpureocillium is proposed for P. lilacinus and the new combination Purpureocillium lilacinum is made here. Furthermore, the examined Purpureocillium lilacinum isolated grouped in two clades based on ITS and partial TEF sequences. The ITS and TEF sequences of the Purpureocillium lilacinum isolates used for biocontrol of nematode pests are identical to those causing infections in (immunocompromised) humans. The use of high concentrations of Purpureocillium lilacinum spores for biocontrol poses a health risk in immunocompromised humans and more research is needed to determine the pathogenicity factors of Purpureocillium lilacinum.

  • keratitis
  • catheter-related infection
  • taxonomy


Paecilomyces lilacinus is a ubiquitous, saprobic filamentous fungus commonly isolated from soil, decaying vegetation, insects, nematodes and laboratory air (as contaminant), and is a cause of infection in man and other vertebrates. This species can colonize materials such as catheters and plastic implants and can contaminate antiseptic creams and lotions, causing infections in immunocompetent and immunocompromised patients (Castro et al., 1990; Orth et al., 1996; Itin et al., 1998). The prevalence of P. lilacinus in patients has increased recently (Carey et al., 2003; Rosmaninho et al., 2010). A review of 119 infections caused by P. lilacinus after 1964 showed that the most frequent manifestation is keratitis, but other sites of the body were also affected (Pastor & Guarro, 2006). Keratitis caused by P. lilacinus typically occurs by external invasion. Common predisposing factors are chronic keratopathy, environmental trauma, implant surgery following lens and/or cornea replacements and extended use of contact lenses (Domniz et al., 2001; Yuan et al., 2009). Paecilomyces lilacinus infections are reported in patients taking immunosuppressant drugs for transplant surgery for liver, kidneys, bone marrow and heart (e.g. Castro et al., 1990; Orth et al., 1996; Lott et al., 2007; Schooneveld et al., 2007). Although commonly reported as a component of the soil mycobiota, the source of P. lilacinus infections in humans has rarely been traced. Exceptions are a catheter-related P. lilacinus fungemia in an immunocompromised child (Tan et al., 1992), a sodium bicarbonate solution used as a neutralizing agent for a sodium hydroxide sterilizer for artificial lenses (Pettit et al., 1980) and a skin lotion used by patients in a haematology–oncology and bone marrow transplant wards (Orth et al., 1996; Itin et al., 1998).

The first aim of the current study was to clarify the phylogenetic position of P. lilacinus and to find out whether purple-spored species with morphologies similar to P. lilacinus form a monophyletic assemblage within the Hypocreales. The second aim was to determine whether there are clades within P. lilacinus, which only comprise vertebrate or invertebrate pathogens. Towards this aim, translation elongation factor 1-α (TEF) gene and internal transcribed spacer (ITS) sequences from strains obtained from clinical specimens were compared with those from isolates of soil, insects and indoor environments or used as biocontrol agents.

Materials and methods


Strains isolated from various clinical specimens and hospital environments are emphasized in our selection of P. lilacinus isolates. These strains are supplemented with isolates from various other substrates (soil, indoor environment, insects and nematodes), and originate from various collections worldwide. An overview of isolates and sources is shown in Supporting information, Table S1.

Morphological examination

A selection of isolates (Table S1) were grown for 7–14 days on malt extract agar (MEA) and were incubated in darkness at 25, 30 and 37 °C. Furthermore, three-point inoculations were made on MEA and incubated for 7 days at 25 °C in darkness (medium compositions in Samson et al., 2010). After incubation, colony diameters were measured and cultures were investigated with a light microscope.

Sequencing and phylogeny

Isolates were grown on MEA for 5–10 days, incubated at 25 °C. Total DNA was extracted using the Ultraclean Microbial DNA isolation Kit (MBio, Solana Beach, CA) according the manufacturer's instructions. DNA sequences of the 18S rRNA gene were obtained from the GenBank database, and amplification of the ITS regions and a part of the TEF gene was preformed as described by Houbraken et al. (2011) and Dodd et al. (2002), respectively. The ITS and TEF dataset was combined and maximum likelihood analysis was performed using raxml version 7.2.8. Each dataset was treated as a separate partition. Two Cryptococcus neoformans sequences (GenBank nos AJ560317 and AJ560313) were used to root the 18S rRNA gene phylograms. The phylogram based on combined TEF and ITS sequences were rooted with Paecilomyces marquandii DTO 145E5.

Nucleotide sequence accession numbers

The sequences used for building the 18S rRNA gene phylogram were downloaded from the NCBI GenBank database. Newly generated sequences are deposited in GenBank under accession numbers HQ842812HQ842841.


Generic delimitation

The phylogenetic analysis of the 18S rRNA gene region confirms the data of Luangsa-ard et al. (2004), showing the polyphyletic nature of Paecilomyces. Figure 1 shows that the type species of Paecilomyces, Paecilomyces variotii, is located in the family of the Trichocomaceae (Eurotiales) near Aspergillus, Penicillium and related species, forming a sister clade with the Onygenales. On the other hand, P. lilacinus belongs to the Ophiocordycipitaceae, a family recently introduced by Sung et al. (2007). The purple-spored species P. marquandii is phenotypically similar to P. lilacinus, but failed to group with P. lilacinus in the phylogenetic analysis using 18S rRNA gene sequences, and this species grouped with green-spored species within the family of Clavicipitaceae. Detailed phylogenetic analysis showed that the purple-colored species Paecilomyces nostocoides, P. lilacinus, Isaria takamizusanensis and Nomuraea atypicola are closely related (Sung et al., 2007; this study) and the former three species have identical partial 18S sequence. None of these species are types of a genus, which warrants the introduction of the new genus Purpureocillium for these species. Phenotypically, Paecilomyces sensu stricto (s. str.) (P. variotii) can be differentiated from Purpureocillium by its rapid growth on agar media. Species belonging to Paecilomyces s. str. have a higher optimum and maximum growth temperature (30–45 °C) compared with Purpureocillium (25–33 °C). Furthermore, the conidial color of Paecilomyces s. str. is olive-brown and chlamydospores are frequently formed, while the conidia of Purpureocillium are lilac and chlamydospores absent.

Figure 1

Phylogram based on partial 18S rRNA gene sequences showing that Purpureocillium lilacinum belongs Ophiocordycipitaceae and Paecilomyces variotii to the Trichocomaceae.

Variability within P. lilacinus

Figure 2 shows the results of the maximum likelihood analysis of the combined ITS and TEF sequences and three clades are present in this phylogram. The P. lilacinus isolates split up in two clades. The type culture of P. lilacinus CBS 284.36T is present in one clade, together with the type strain of P. nostocoides and all the examined strains originating from clinical specimens and hospital environments. Furthermore, the majority of P. lilacinus strains from soil, indoor environment, insect larvae, nematodes and decaying vegetation are located in this clade. Minor differences among the ITS and TEF sequences are present within the P. lilacinus clade; however, in various cases, strains originating from insects, nematodes, (indoor) environment and clinical specimens share the same ITS and TEF sequence. No clinical P. lilacinus isolates were present in the other smaller clade. The P. lilacinus isolates from this group are saprobes and seem to have a worldwide distribution (India, Ghana, Israel, Australia). This clade represents a new species and will be described in future (unpublished data). Also I. takamizusanensis and P. nostocoides grouped well with P. lilacinus. The former species is associated with insects, and the latter with corn cyst nematodes. Both species share the ability to form purple-colored conidia. Our results show that P. nostocoides is phylogenetically closely related to P. lilacinus. Comparison of an ITS sequence originating from the ex-type culture of P. nostocoides and deposited in GenBank (AB104884) shows that this sequence is similar to those generated in this study on P. lilacinus. The main difference is the presence of a 29-nucleotide gap in the ITS1 region of P. nostocoides (GenBank AB104884). The ITS regions of the ex-type culture of P. nostocoides (DTO 149E4) were reanalyzed in this study, and in contrast to the sequence deposited on GenBank, these data could not confirm the presence of this 29-nucleotide gap in the ITS1 region. The absence of this gap and the high similarity of the partial TEF sequence of this strain to other P. lilacinus indicates that P. nostocoides is conspecific with P. lilacinus. Furthermore, N. atypicola is phylogenetically related to P. lilacinus (Sung et al., 2007) and possesses lavender-colored conidia similar to those of P. lilacinus (Hywel-Jones & Sivichai, 1995). The taxonomy of the genus Purpureocillium, including the phylogenetic relationship between I. takamizusanensis, N. atypicola, P. nostocoides and P. lilacinus, will be treated elsewhere.

Figure 2

Phylogram based on the ITS regions (including 5.8S rRNA gene) and partial TEF sequences showing the diversity among Purpureocillium lilacinum isolates and the relationship with other closely related species. Animal: Embedded Image, human; Embedded Image, nematode; Embedded Image, entomogenous; Embedded Image, other; Embedded Image (indoor) environment/soil; Embedded Image, miscellaneous;?, undetermined.


Purpureocillium Luangsa-ard, Hywel-Jones, Houbraken & Samson gen. nov. Mycobank MB 519529

=Paecillium Luangsa-ard, Hywel-Jones & Samson nomen provisorium– Compendium of soil fungi, 2nd edn, p. 322, 2007.

Type: Penicillium lilacinum Thom.

Latin diagnosis: Conidiophora ex hyphis submersis oriunda, seu mononematosa, phialibus vix in collulum extensi, seu laxis synnematibus connexa, rigida, verticillata; phialidibus collulo distincte angustato praeditis. Conidia in catenis siccis divergentibus adhaerentia, cylindrica (recta vel modice curvata) vel ellipsoidea vel fusiformia, rugulosa, hyalina, aggregata purpurea.

Etymology: The generic name refers to the purple colored conidia produced by its type species, Purpureocillium lilacinum.

Colonies on MEA moderately to fast growing consisting of either a basal or compact crustose felt of numerous conidiophores with a floccose overgrowth of aerial mycelium. Colonies at first white, becoming pink and lilac with the onset of sporulation. Reverse usually in shades of purple or yellow. Conidiophores arising from submerged hyphae, mononematous, stiff, verticillate; phialides ovate to cylindric with distinct neck or erect and densely grouped, forming verticils of branches and cylindrical phialides without or with very short necks. Conidia in dry divergent chains, straight to slightly curved or ellipsoidal to fusiform, slightly roughened, purple in mass.

Purpureocillium lilacinum (Thom) Luangsa-ard, Houbraken, Hywel-Jones & Samson, comb. nov. Mycobank MB 519530

Basionym: Penicillium lilacinum Thom –Bull Bur Anim Ind US Dep Agric, 118: 73 (1910).

=Paecilomyces lilacinus (Thom) Samson –Stud Mycol6: 58 (1974).

=Paecilomyces nostocoides Dunn –Mycologia75: 179 (1983).

Colonies on MEA (Oxoid) fast growing, attaining a diameter of 25–35 mm after 7 days at 25 °C; no or restricted growth at 37 °C, 0–10 (−20) mm. Colonies consisting of a basal felt with or without floccose aerial overgrowth (Fig. 3a and b), some isolates strongly floccose (Fig. 3c), white at first, becoming vinaceous; reverse mostly in shades of purple or sometimes uncolored. Conidiophores arising from submerged hyphae 4–6 μm in length, occasionally forming loose synnemata up to 2 mm high; stalks with roughened thick walls 3–4 μm wide consisting of verticillate branches with whorls of two to four phialides. Phialides 6–9 × 2.5–3 μm, having a swollen basal portion tapering into a short distinct neck about 1 μm wide. Conidia in divergent chains, ellipsoidal to fusiform, smooth-walled to slightly roughened, hyaline, purple en masse, 2–3 × 2–4 μm. Conidial structures formed near the agar atypical: phialides solitary or in verticils, 2–4, variable in length (Fig. 3g and h); shaped like typical Purpureocillium lilacinum phialides, or very long (up to 30 μm) and Acremonium-like. Cylindrical, occasionally slightly curved conidia formed in ‘slimy heads’ on these Acremonium-like structures, conidia on these structures variable in size, measuring 2.0–14 × 1.5–2.5 μm (Fig. 3i). This conidiogenesis was also observed by Okada et al. (1995) for P. nostocoides (=Purpureocillium lilacinum). Chlamydospores absent.

Figure 3

Purpureocillium lilacinum:(a–c) 14-day-old culture on MEA. (a) DTO 63E5, typical sporulating colonies; (b) DTO 63E1, typical sporulating colonies; (c) floccose colonies, DTO 141C2, (d, e) well-defined conidiophores; (f) typical fusiform conidia; (g, h) Acremonium-like conidiophores; DTO 141C2; (i) cylindrical conidia formed near the agar; DTO 141C2. Scale bar=10 μm.


Identification of Paecilomyces spp.

Species previously assigned to Paecilomyces causing human mycoses include Paecilomyces farinosus, Paecilomyces javanicus, P. lilacinus, P. marquandii, Paecilomyces taitungiacus (=anamorph of Thermoascus taitungiacus), P. variotii and Paecilomyces viridis. Of these, P. variotii is retained in the genus Paecilomyces (as it is the type), P. javanicus and P. farinosus have been returned to the genus Isaria in the Hypocreales (Luangsaard et al., 2004), P. viridis has been transferred to Chamaeleomyces (Sigler et al., 2010) and P. lilacinus is accommodated here in the genus Purpureocillium. P. marquandii is currently maintained in Paecilomyces; however, this species is unrelated to P. variotii and should to be transferred to a new genus. Paranomuraea was suggested for P. marquandii and Paecilomyces carneus (Domsch et al., 2007), but this genus has yet not been published validly. Samson (1974) considered P. lilacinus and P. marquandii to be very close to each other, based on overall morphology and spore color. Paecilomyces marquandii differs from Purpureocillium lilacinum by its hyaline conidiophores and the typical yellow reverse. Although both species have a similar morphology, phylogenies show them to be separated in two families of the Hypocreales (Sung et al., 2007). Some clinical isolates have been identified as P. marquandii (Castro et al., 1990; Naldi et al., 2000). These isolates need to be re-examined using sequence-based methods to determine whether P. marquandii genuinely has the potential for human pathogenicity or whether this is merely a misidentification of Purpureocillium lilacinum. Correct identification is crucial because Purpureocillium lilacinum is significantly more resistant to amphotericin B than P. marquandii (Aguilar et al., 1998).

Purpureocillium lilacinum in clinical settings

Our results and those of other studies show that Purpureocillium lilacinum is a commonly occurring saprobic species and this species is isolated from soil, decaying vegetation, insect and insect larvae, nematodes, humans, animals and (indoor) air (Samson, 1974; Castro et al., 1990; Itin et al., 1998). This species can contaminate antiseptic creams and (skin) lotions, sodium bicarbonate solutions used as a neutralizing agent for a sodium hydroxide sterilizer for artificial lenses, and colonize materials such as catheters and plastic implants (Pettit et al., 1980; Orth et al., 1996; Itin et al., 1998). A 3-year surveillance study showed that Purpureocillium lilacinum was frequently found in water distribution system of a bone marrow transplantation unit. Purpureocillium lilacinum positive sites included water from water tanks and showers, sinks, showers (including drains), toilets and air. This species can thrive on wet and moist surfaces of water distribution systems and form a biofilm, together with other species such as Aspergillus, Fusarium and Acremonium (Anaissie et al., 2003). Although biofilm formation by filamentous fungi has been poorly studied, it is postulated that adhesion, colonization and matrix formation are key criteria in the biofilm formation process (Martinez & Fries, 2010). The capacity of Purpureocillium lilacinum to adhere to the waxy host cuticle of nematodes and its ability to colonize surfaces under harsh conditions with low nutrient concentrations (fungal biofilters, plastics) and low oxygen levels (Mountfort & Rhodes, 1991; Vigueras et al., 2008) suggested that this species is able to form a biofilm. Concordant with our results, Okada et al. (1995) showed that Purpureocillium lilacinum is a dimorphic species and is able to form an Acremonium-state in and/or on agar media. This Acremonium-state phenotypically resembles Fusarium solani, a fungal pathogen causing severe corneal disease and the causal agent of an outbreak of lens-associated keratitis. Remarkably, the most frequent manifestation of Purpureocillium lilacinum is also keratitis (Pastor & Guarro, 2006), suggesting that both species might have similar properties besides their phenotypic similarity. In this respect, it needs to be noted that Imamura et al. (2008) showed that F. solani has the ability to form biofilms on lenses; however, this appears to be strain rather than species dependent.

Antifungals and Purpureocillium lilacinum

Paecilomyces can cause hyalohyphomycosis, and two species, Purpureocillium lilacinum (=P. lilacinus) and P. variotii, are the most frequently encountered (Walsh et al., 2004; Houbraken et al., 2010). The phylogenies described here and elsewhere explain why some treatments will work for one species and fail for the others. Major differences in antifungal susceptibility profiles were found between P. variotii and Purpureocillium lilacinum in vitro. Amphotericin B showed good activity against P. variotii and related species in vitro, as was the case for flucytosine (Aguilar et al., 1998; Castelli et al., 2008; Houbraken et al., 2010). However, these antifungals are not active against Purpureocillium lilacinum, and treatments of infections may therefore be complicated. Also itraconazole has limited efficacy against Purpureocillium lilacinum in vitro. Voriconazole, terbinafine, ravuconazole and posaconazole were active against Purpureocillium lilacinum, with posaconazole being the drug with the best in vitro activity (e.g. Martin et al., 2002; Pastor & Guarro, 2006; Sponsel et al., 2006; Houbraken et al., 2010). Posaconazole may be the only appropriate alternative agent, although the lack of an intravenous formulation and limited penetration into the cerebrospinal fluid might limit its use (Rodríguez et al., 2009; Houbraken et al., 2010). On the other hand, Ortoneda et al. (2004) showed that a combination of terbinafine combined with ravuconazole and voriconazole gave the best results in vitro.

The in vitro susceptibility of Purpureocillium lilacinum for itraconazole seems to be strain dependent and both susceptible and resistant strains are reported (Pastor & Guarro, 2006; Castelli et al., 2008; Houbraken et al., 2010). Kitami et al. (2005) and Zendri et al. (2006) found that orally administered itraconazole successfully treated cutaneous infections. Recently, a large body of literature has accumulated on the successful treatment of keratitis and other Purpureocillium lilacinum infections with voriconazole alone or in combination with terbinafine (Martin et al., 2002; Chang et al., 2008; Yuan et al., 2009). The efficacy of voriconazole was also successfully demonstrated in a murine model, when compared with amphotericin B (Rodríguez et al., 2010).

Purpureocillium lilacinum in biocontrol and other benefits

There is a significant body of literature that has demonstrated the negative impact of Purpureocillium lilacinum to mankind in the form of medically important infections. However, there is also a wealth of literature reporting the use of Purpureocillium lilacinum for the control of nematode pests (e.g. Brand et al., 2003; Kalele et al., 2007). It is therefore possible that isolates of Purpureocillium lilacinum used as biological control agents of nematodes could form opportunistic mycoses in humans as well as other vertebrates. Literature suggests that Purpureocillium lilacinum is most often a problem in immunocompromised patients with very few instances of it occurring in apparently immunocompetent subjects. Our ITS and TEF data suggest that it is not possible to separate harmful from beneficial isolates of Purpureocillium lilacinum. Other genotyping techniques such as multilocus sequence typing, microsatellite analysis or amplified fragment length polymorphism have a higher resolution and might show a genetic structure within Purpureocillium lilacinum. Furthermore, these typing techniques might enable tracking of the biocontrol Purpureocillium lilacinum strain(s) released into the environment.

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Table S1. Purpureocillium lilacinum isolates examined in this study.


We thank Martin Meijer (CBS-Fungal Biodiversity Centre) and Adrien Szekely (UK Mycology Reference Laboratory) for their excellent technical assistance. Various Purpureocillium lilacinum isolates were kindly provided by Stephen W. Peterson and Jens C. Frisvad. Dr Uwe Braun kindly advised us on the new genus name.


  • Editor: Holger Deising


View Abstract