Mycologia, 96(4), 2004, pp. 929–934.

q 2004 by The Mycological Society of America, Lawrence, KS 66044-8897

Hirsutella uncinata, a new hyphomycete from Australia

Keith A. Seifert1 species were described by Minter et al (1983), Evans

Hillary Boulay and Samson (1982), Samson et al (1980), Samson
Biodiversity Theme (Mycology and Botany), Eastern and Evans (1985) and Fernández-Garcı́a and Evans
Cereal and Oilseed Research Centre, Agriculture and (1990). The mononematous species of Hirsutella in-
Agri-Food Canada, 960 Carling Ave., Ottawa, clude H. thompsonii F.E. Fisher, an important patho-
Ontario, K1A 0C6 Canada
gen of the citrus-rust mite Phyllocoptruta oleivora
Ashm. and other eriophyoid mites (van der Geest et
al 2000) and the anamorph of Cordyceps sinensis
Abstract: Hirsutella uncinata, collected on follicles
(Berk) Sacc., the famous ‘‘caterpillar fungus’’ of tra-
of Hakea sp. in New South Wales, Australia, is de-
ditional Chinese herbal medicine (Liu et al 2001).
scribed as a new anamorph species. The fungus pro-
The known teleomorphs of Hirsutella species are
duces stiff, erect, verticillate conidiophores with up
classified in Cordyceps (Fr.) Link, Ophiocordyceps
to 10 whorls of hooked, phialidic conidiogenous cells
Petch, Torrubiella Boudier (Morris 1963, Samson et
that have rough-walled necks. The conidia are sub-
al 1982, Hywel-Jones 1994, 1995a, b) and Cordycepioi-
fusiform, flattened on one side and produced in
deus Stifler (Ochiel et al 1997, Suh et al 1998). Com-
groups of 2–3 in a dense slimy ball at the end of each
paratively little DNA sequence data derived for tax-
phialide. Parsimony analysis of aligned internal tran-
onomic purposes has been published specifically on
scribed spacer rDNA sequences suggests that H. un-
cinata is closely related to the Cordyceps sinensis com- Hirsutella, but its phylogenetic affinities are well es-
plex (Clavicipitaceae, Hypocreales), which includes tablished in the Clavicipitaceae, Hypocreales. Several
species with morphologically similar anamorphs. other anamorph genera are associated with Cordy-
Key words: Clavicipitaceae, Cordyceps sinensis, in- ceps, many of which also have species with phialidic
ternal transcribed spacers, phylogenetics, taxonomy conidiogenous cells, such as Akanthomyces Lebert, Gi-
bellula Cavara, Harposporium Lohde, Isaria Pers., Par-
aisaria Samson & B.L. Brady, and Tolypocladium W.
Gams (Hodge 2003).
INTRODUCTION
In 1999, a striking species of Hirsutella emerged
The hyphomycete genus Hirsutella Pat. includes from follicles (fire-resistant, woody, nut-like fruits) of
about 65 species (Hodge 1998), some that are path- Hakea sp. (Proteaceae) collected in New South Wales,
ogenic to insects, mites and other invertebrates. The Australia, and incubated 1–2 wk several months later
species of this genus have unusual phialides that ta- in a damp chamber. The phialides and conidial mas-
per into a long narrow neck, and produce usually ses clearly identified it as a species of Hirsutella, with
only 1–3 conidia in a dense terminal sphere of slime. a conidiophore branching pattern and hooked phial-
In some species, more than one neck emerges from ides that easily distinguished it from all known spe-
the basal part of the conidiogenous cell; in others, cies. This paper presents a description, illustration
the neck branches to form a polyphialide. Many of and preliminary phylogenetic analysis of this new
the known species produce synnematous conidiom- species.
ata, but several species lack conidiomata. Minter and
Brady (1980) revised the taxonomy of 12 monone-
MATERIALS AND METHODS
matous species and distinguished them by the warti-
ness of the conidia and the phialides, characters of Morphological examination.—For isolations, malt yeast agar
the phialide neck (helical versus straight, multiple (MYA: 10.0 g Difco Bacto malt extract, 2.0 g Difco yeast
versus single necks, monophialidic versus polyphial- extract, 20.0 g Difco Bacto agar, 1000 mL distilled water).
idic), the arrangement of conidiogenous cells on co- For morphological studies, cultures were grown on oatmeal
nidiophores, the pigmentation and dimensions of agar (OA), Blakeslee’s malt extract agar with trace metal
microscopic structures. Additional mononematous solution (MEA, made with Difco Bacto malt extract, Sparks,
Maryland) and Synthetischer nährstoffarmer agar with fil-
Accepted for publication January 15, 2004. ter paper (SNA) (Samson et al 2000) in the dark at 25 C,
1 Corresponding author. E-mail: seifertk@agr.gc.ca or in a 12 h:12 h light:dark cycle on a room temperature

929
930 MYCOLOGIA

FIG. 1. One of 36 equally parsimonious trees based on a heuristic analysis of the internal transcribed spacer of Hirsutella
uncinata and related species (95 steps, CI 0.625, RI 0.741, RC 0.460, HI 0.379). ‘‘H’’ after a Cordyceps name indicates species
with known Hirsutella anamorphs. The single pair of numbers over the long branch indicates the bootstrap and jackknife
support received for the C. sinensis complex. Bold lines indicate the parts of the tree with 100% consensus support.

incubation bench fitted with fluorescent and near UV heuristic searches with uninformative characters removed.
lights. Color codes and capitalized color names refer to Kor- Searches with simple stepwise addition were run, and for
nerup and Wanscher (1978). Means and standard errors the ITS analysis 1000 replicates of random stepwise addition
are presented for some measurements. Cardinal tempera- also were run. Bootstrap and jackknife analyses were run
tures were determined by growing the culture in incubators with 1000 replications using simple stepwise addition.
at 5 C increments, from 5 to 40 C, on cornmeal agar (CMA)
(BBL, Sparks, Maryland) in the dark, with three colonies at
each temperature. RESULTS

DNA extraction, DNA sequencing and phylogenetic analyses.— The original culture of H. uncinata was isolated onto
DNA extraction, purification, amplification and sequencing MYA by removing individual conidial masses from co-
methods were similar to those described by Hambleton et nidiophores using a sterile needle under the dissect-
al (2003), using primers NS1, NS4, ITS1, ITS4 (White et al ing microscope. Conidia germinated only after in-
1990), except that UltraClean Microbial DNA Isolation and cubating the isolation plate for several weeks at 4 C.
UltraClean PCR Purification kits (Mo Bio Laboratories Inc., An oddity of this organism is its ability to grow in the
Solana Beach, California) were used for DNA extraction
dark on SNA (which contains nitrate as the sole ni-
and cleaning of PCR products, and an ABI PRISMt 3700
DNA Analyzer (Applied Biosystems, Foster City, California)
trogen source) and its inability to grow on the same
was used for sequencing. Reference sequences for the small medium when exposed to light.
subunit rDNA (18S) and internal transcribed spacer (ITS) Cardinal temperatures.—Radial growth after 4 wk on
were chosen from GenBank using BLAST searches. After
CMA, 5 C 1 mm, 10 C 2 to 3 mm, 15 C 4 to 5 mm,
initial alignments and parsimony analyses, sequences that
either were relatively distant from the Hirsutella species, or
20 C 8 to 9 mm, 25 C, 9 to 10 mm, no growth at 30,
identical sequences from different accessions of individual 35 or 40 C. Sporulation occurred at all temperatures
species, were discarded. GenBank accession numbers for where growth occurred.
ITS sequences are shown on FIG. 1. These reference se-
Phylogenetic analysis.—Analyses of the small subunit
quences originated in studies by Chen et al (2001), Liu et
al (2001), Nikoh and Fukatsu (2000) and Gernandt and
rDNA (SSU) sequences confirmed that H. uncinata
Stone (1999). Alignments were calculated using Clustal W was related to the teleomorph genus Cordyceps and
as implemented in MegAlign 5.05 (DNAStar, Madison, Wis- supported its classification in Hirsutella. The topolo-
consin), and adjustments were made visually using Se-Al gy of these trees was similar to the more resolved ITS
v.1.01a1 (Rambaut 1996). Parsimony analysis of alignments tree discussed below and ultimately the SSU analysis
were performed with PAUP* 4.0b10 (Swofford 2003) using was used only to select an outgroup for the ITS anal-
 SEIFERT AND BOULAY: HIRSUTELLA UNCINATA 931

ysis. Therefore, the results are not shown here but branches of the main axis, sometimes with a single
the NS1-4 SSU sequence has been deposited in branch giving rise to a lateral verticillate structure,
GenBank as AY365469. A consideration of the phy- with downwards directed hyphae sometime branch-
logenetic distribution of the various anamorph gen- ing from the basal cell and anchoring the conidio-
era affiliated with Cordyceps is beyond the scope of phore to the substrate. Conidiogenous cells phialidic,
this paper, but preliminary investigations based on 21–36(–50) mm long (mean 5 29 6 1.3, n 5 25,
ITS phylogenies were presented by Liu et al (2002). measured as maximum linear distance of the curva-
The ITS alignment included 570 characters, 45 of ture away from junction with the stipe rather than to
which were informative (TreeBase study S1004, ma- the conidiogenous aperture), in whorls of 3–6(–9),
trix M1696). Although the most closely related spe- sometimes in pairs, hyaline, uncinate, curving up to
cies were chosen on the basis of the 18S analyses, the 1808 and then with the conidiogenous aperture fac-
Clustal W alignment of the data set required consid- ing the stipe, with a smooth-walled, broad base 2.5–
erable adjustment. Parsimony analyses resulted in 36 4.0 mm wide, tapering more or less uniformly to a
equally parsimonious trees of 95 steps, one of which rough-walled to echinulate, curved, cylindrical neck,
is shown as FIG. 1. The different trees were the result conidiogenous aperture about 1.0 mm wide, periclin-
of reorganization of taxa within the C. sinensis com- al thickening inconspicuous, collarettes not seen; po-
plex, which was the only group with significant boot- lyphialides present on some conidiophores, evident
strap support. Cordyceps japonica Lloyd was chosen to as straight or curved lateral necks 4.0–7.5 3 1.0 mm
root the ITS analysis based on its position in the 18S or entirely new secondary phialides emerging from
analysis; this is one of the Cordyceps species parasitic the primary phialide; phialides sometimes with an in-
on other fungi. Hirsutella uncinata is a member of a ternal septum; the terminal phialide on the conidi-
clade of Cordyceps species including three other spe- ophores often uncurved. Conidial mass bright white
cies of mononematous Hirsutella anamorphs. The under the dissecting microscope, hyaline with the
new species was sister of the C. sinensis complex, with compound microscope, 7–11 mm diam, containing
C. robertsii (Hook.) Berk. occupying a basal position. 1–3 conidia. Conidia 5.5–7.0 3 3–4 mm (mean 5 6.4
Comparatively little ITS data is publicly available for 6 0.1 3 3.7 6 0.1, n 5 13), in face view subfusiform
Hirsutella species or allied Cordyceps species, so the or subclavate, in side view flattened on one side, with
significance of these relationships is difficult to assess. a truncate base about 1 mm wide, and a rounded or
However, these data support the classification of H. somewhat pointed apex, hyaline, produced in an el-
uncinata in this anamorph genus, and suggest that if lipsoidal, globose or limoniform slimy mass. In ad-
a teleomorph exists, it would be similar to the mem- dition to the macronematous conidiophores, repent
bers of the Cordyceps sinensis complex. conidiophores spread sparsely over the substrate,
with irregularly scattered single, paired or whorled
phialides and conidia identical to those described
TAXONOMY above.
Colonies on MEA after 6 wk at room temperature:
Hirsutella uncinata Seifert & Boulay, sp. nov. 22–24 mm diam, planar, with no aerial mycelium and
FIGS. 2–8 then Golden Yellow to Orange (5BC6–8) or Brown
Conidiophora 200–400 mm alta, erecta, (1–)3–10 verticil-
(6E8), or with relatively dense, felty aerial mycelium
los densissimos phialidum ferentia; phialides hyalinae, un-
cinatae, 21–36(–50) 3 2.5–4.0 mm, 3–6(–9) verticillatae,
around the inoculum and then Orange Grey (5B2–
ventro plus minusve echinulato, collulo 1.0 mm diam, mod- 4), the surface dense and rubbery, reverse concolo-
ice polyphialidicae. Capitula conidiorum alba, 7–11 mm rous with obverse, often dark brown below inoculum.
diam, 1–3 conidia continentia. Conidia nonseptata, hyalina, Colonies on OA after 6 wk at room temperature: 30–
5.5–7 3 3–4 mm, subfusiforma, ex uno latere applanata, ad 40 mm diam, white, planar, with low, sparse white
basim minute truncata, sursum angustata vel rotundata. aerial mycelium, sometimes with a denser ring of
Culturae in extracto malti ‘‘Blakeslee’’ ad 25 C post 42 dies white aerial mycelium near the margin, reverse gray-
circa 22–24 mm radii crescentes. Holotypus DAOM 229256. ish in center, white in a ring at the margin. Repent
Conidiophores macronematous, scattered and sin- conidiophores spreading sparsely on both media;
gle, or in pairs or caespitose groups, erect or nod- erect conidiophores similar to those on the natural
ding, stiff, hyaline, 200–400 mm tall, the stipe 8–10 substrate mostly arising from the inoculum block, mi-
mm wide at the base, tapering to 4.0–5.0 mm wide croscopically similar to those produced in vivo, but
where the fertile zone begins, walls up to 1.0 mm usually with fewer whorls of phialides, and often few-
thick at the base, sterile for one-half to two-thirds of er phialides per whorl.
the length, and then with (1–)3–10 whorls of coni- Cardinal temperatures. Minimum ,5 C, optimum
diogenous cells 14.5–28 mm apart, usually without 25 C, maximum ,30 C.
932 MYCOLOGIA

FIGS. 2–8. Photomicrographs of Hirsutella uncinata holotype, DAOM 229256. 2. Habit of conidiophores. 3. Conidia. 4.
Conidiophore, showing verticillate branching and downward directed hyphae at the base. 5. Detail of phialides and conidial
masses. 6–8. Polyphialides. 2, 4–5 from living material, 3, 6–8 from dried material. Scale bars 5 50 mm for FIG. 2, 10 mm for
FIG. 4, and 10 mm in FIG. 5 for FIGS. 3, 5, 6, 7, 8. FIGS. 2, 3 and 4 are composites.
 SEIFERT AND BOULAY: HIRSUTELLA UNCINATA 933

Etymology. uncinatus, derived from the Latin un- palensis Zang & N. Kinjo) in our ITS analysis, the
cus, meaning ‘‘hooked’’, describing the shape of the relatively distant position of H. uncinata and its other
phialides. closest neighbor, C. robertsii, further supports the rec-
Material examined. Australia. New South Wales: ognition of our fungus as a distinct species.
Mount Tomah Botanical Garden, S33832.49, Of the mononematous species of Hirsutella consid-
E150825.49, ca 1100 m elevation, on woody follicles ered by Minter and Brady (1980), and those de-
of Hakea sp. on the ground, dried after incubation scribed subsequently, only three other species, H. ver-
in a moist chamber for 1–2 wk, K. A. Seifert No. 1316, ticillioides Charles, H. guyana Brady & Minter and H.
17 Aug 1999 (HOLOTY PE DAOM 229256, ex-type necatrix Minter, Brady & R.A. Hall have conidio-
culture deposited in Canadian Collection of Fungal phores with verticillate conidiogenous cells. Of these,
Cultures under the same number, ISOTY PE DAR). only H. necatrix has the multiple whorls along the
central conidiophore axis so characteristic of H. un-
DISCUSSION cinata, but it has shorter, more or less ampulliform
phialides lacking a terminal hook. The conidiogen-
Hirsutella uncinata is recognized easily on the natural ous cells of H. verticillioides usually are conspicuously
substrate by its stiff conidiophores with multiple polyphialidic, whereas those of H. guyana have a
whorls of uncinate phialides. The conidiophores in characteristic constriction in the phialide necks and
culture appear less regularly verticillate and rather only occasionally are polyphialidic. There are no pub-
lax compared to those on the natural substrate. In lished sequence data for these species.
common with some of the mononematous species of
Hirsutella described by Minter and Brady (1980), the
holotype material is sparse but the distinctive mor- ACKNOWLEDGMENTS
phology is diagnostic. All other species of Hirsutella
We are grateful to Dr Brett Summerell, Royal Botanic Gar-
grow on insects, mites or nematodes, but we were
den, Sydney, Australia, for arranging the field trip during
unable to detect an invertebrate host for H. uncinata.
which this fungus was collected. Dr Sarah Hambleton kindly
The substrate, a follicle of a Hakea sp., is dense and advised us on details of DNA sequencing and phylogenetic
woody and not amenable to slicing or subsurface analysis. We appreciate helpful presubmission reviews by Sa-
probing. Seeds of these plants are high in protein rah Hambleton and John Bissett.
and attractive to a variety of insects (Lamont and
Groom 1998), and we assume that the true substrate
of H. uncinata was larvae hidden by the solid tissues LITERATURE CITED
of the follicle.
Chen Y-Q, Wang N, Qu L-Q, Li T-H, Zhang W-M. 2001.
The phylogenetic analyses of the SSU and ITS se-
Determination of the anamorph of Cordyceps sinensis
quences of H. uncinata confirm that it is classified
inferred from analysis of the ribosomal DNA internal
appropriately in Hirsutella and suggest a sister rela- transcribed spacers and 5.8S rDNA. Biochem Syst Ecol
tionship with the Cordyceps sinensis complex. As illus- 29:597–607.
trated by Liu et al (1989), H. sinensis has similar ver- Evans HC, Samson RA. 1982. Entomogenous fungi from the
ticillate conidiophores and rough-walled phialides to Galápagos Islands. Can J Bot 60:2325–2333.
H. uncinata. The microscopic dimensions of the Fernández-Garcı́a E, Evans HC. 1990. Hirsutella cryptoscler-
phialides are similar, but the phialides of H. sinensis otium sp. nov., an entomopathogen of the mealybug
are not uncinate and polyphialides have not been pest Rastrococcus invadens, in West Africa. Mycol Res
reported. The conidia of H. sinensis are described 94:1111–1117.
and drawn as reniform and up to 14 mm long, longer Gernandt DS, Stone JK. 1999. Phylogenetic analysis of nu-
and differently shaped from those of H. uncinata. clear ribosomal DNA places the nematode parasite,
Because H. sinensis is reported only from culture, we Drechmeria coniospora, in Clavicipitaceae. Mycologia 91:
are unable to compare its wild-type morphology with 993–1000.
Hambleton S, Tsuneda A, Currah RS. 2003. Comparative
that of H. uncinata. Liu et al (1989) reported an
morphology and phylogenetic placement of two mi-
optimal temperature of growth for C. sinensis as 15–
crosclerotial black fungi from Sphagnum. Mycologia 95:
20 C, with significant inhibition at 25 C, which cor- 959–975.
relates with its occurrence in the high Himalaya, Hodge KT. 1998. Revisionary studies in Hirsutella (Hypho-
whereas optimal growth of H. uncinata occurs at 25 mycetes) [Doctoral dissertation]. Department of Plant
C, consistent with its temperate home in the Blue Pathology, Cornell University. Ithaca, N.Y.
Mountains of Australia. Given the close relationship . 2003. Clavicipitaceous anamorphs. In White JF, Ba-
among representatives of the C. sinensis complex (in- con CW, Hywel-Jones NL, Spatafora JW, eds. Clavicipi-
cluding C. multiaxialis Zang & N. Kinjo and C. ne- talean fungi: evolutionary biology, chemistry, biocon-
934 MYCOLOGIA

trol, and cultural impacts. New York: Marcel Dekker, Nikoh N, Fukatsu T. 2000. Interkingdom host jumping un-
Inc. p 75–123. derground: phylogenetic analysis of entomoparasitic
Hywel-Jones NL. 1994. Cordyceps khaoyaiensis and C. pseu- fungi of the genus Cordyceps. Mol Biol Evol 17:629–638.
domilitaris, two new pathogens of lepidopteran larvae Ochiel GS, Evans HC, Eilenberg J. 1997. Cordycepioideus, a
from Thailand. Mycol Res 98:939–942. pathogen of termites in Kenya. Mycologist 11:7–9.
. 1995a. Torrubiella iriomoteana from scale insects in Rambaut, A. 1996. Se-Al: Sequence Alignment Editor. Avail-
Thailand and a new related species Torrubiella siamen- able at http://evolve.zoo.ox.ac.uk/.
sis with notes on their respective anamorphs. Mycol Res Samson RA, Evans HC. 1985. New and rare entomogenous
99:330–332. fungi from Amazonia. Proc Ind Acad Sci (Pl Sci) 94:
. 1995b. Cordyceps brunneapunctata sp. nov. infecting 309–317.
beetle larvae in Thailand. Mycol Res 99:1195–1198. , , Hoekstra ES. 1982. Notes on entomogen-
Kornerup A, Wanscher JH. 1978. Methuen handbook of ous fungi from Ghana. VI. The genus Cordyceps. Proc
colour. 3rd ed. London, United Kingdom: Eyre Methu- Kon Ned Akad Wet, Ser. C 85:589–605.
en. 252 p. , Hoekstra ES, Frisvad JC, Filtenborg O. 2000. Intro-
Lamont BB, Groom PK. 1998. Seed and seedling biology of duction to food- and airborne fungi. 6th ed. Utrecht,
the woody-fruited Proteaceae. Austral J Bot 46:387– the Netherlands: Centraalbureau voor Schimmelcultu-
406. res. 389 p.
Liu XJ, Guo YL, Yu YX, Zeng W. 1989. Isolation and iden- , McCoy CW, O’Donnell K. 1980. Taxonomy of the
tification of the anamorphic state of Corydyceps sinensis acarine parasite Hirsutella thompsonii. Mycologia 72:
(Berk.) Sacc. Acta Mycol Sin 8:35–40. 359–377.
Liu Z-Y, Liang Z-Q, Liu A-Y, Yao Y-J, Hyde KD, Yu Z-N. 2002. Suh S-O, Spatafora JW, Ochiel GRS, Evans HC, Blackwell
Molecular evidence for teleomorph-anamorph connec- M. 1998. Molecular phylogenetic study of a termite
tions in Cordyceps based on ITS-5.8S rDNA sequences. pathogen Cordycepioideus bisporus. Mycologia 90:611–
Mycol Res 106:1100–1108. 617.
, Yao Y-J, Liang ZQ, Liu A-Y, Pegler DN, Chase MW. Swofford D. 2003. PAUP*: Phylogenetic Analysis Using Par-
2001. Molecular evidence for the anamorph-teleo- simony (*and other methods). Version 4.0b10. Sunder-
morph connection in Cordyceps sinensis. Mycol Res 105: land, Massachusetts: Sinauer Associates.
827–832. van der Geest LPS, Elliot SL, Breeuwer JAJ, Beerling EAM.
Minter DW, Brady BL. 1980. Mononematous species of Hir- 2000. Diseases of mites. Exp Appl Acarol 24:497–560.
sutella. Trans Br Mycol Soc 74:271–282. White TJ, Bruns T, Lee S, Taylor J. 1990. Amplification and
, , Hall RA. 1983. Five hyphomycetes isolated direct sequencing of fungal ribosomal RNA genes for
from eriophyid mites. Trans Br Mycol Soc 81:455–471. phylogenies. In Innis MA, Gelfand DH, Sninsky JJ,
Morris EF. 1963. The synnematous genera of the Fungi Im- White TJ, eds. PCR protocols: a guide to methods and
perfecti. West Ill Univ Ser Biol Sci 3:1–151. applications. San Diego: Academic Press. p 315–322.