Morphological and molecular characterisation of Penicillium roqueforti and P. paneum isolated from baled grass silage

m y c o l o g i c a l r e s e a r c h 1 1 2 ( 2 0 0 8 ) 9 2 1 – 9 3 2

journa l homepage : www.e l sev i er . com/ loca te /mycres

Morphological and molecular characterisation of Penicilliumroqueforti and P. paneum isolated from baled grass silage

Martin O’BRIENa,b,*, Damian EGANb, Padraig O’KIELYa, Patrick D. FORRISTALc,Fiona M. DOOHANb, Hubert T. FULLERb

aTeagasc, Grange Beef Research Centre, Dunsany, Co. Meath, IrelandbUCD School of Biology and Environmental Science, College of Life Sciences, University College Dublin, Belfield, Dublin 4, IrelandcTeagasc, Crops Research Centre, Oak Park, Co. Carlow, Ireland

a r t i c l e i n f o

Article history:

Received 7 September 2007

Received in revised form

15 January 2008

Accepted 24 January 2008

Corresponding Editor:

Stephen W. Peterson

Keywords:

Cultural features

Forage

Mould

Phylogenetic analyses

Spoilage

* Corresponding author. Tel.: þ353 4690 61E-mail address: [email protected]

0953-7562/$ – see front matter ª 2008 The Bdoi:10.1016/j.mycres.2008.01.023

a b s t r a c t

The morphological and molecular features of Penicillium roqueforti and P. paneum isolated

from baled grass silage were characterised. A total of 315 isolates were investigated, com-

prising 237 P. roqueforti and 78 P. paneum isolates randomly selected from more than 900

Penicillium colonies cultured from bales. The macromorphological features of both species

broadly agreed with the literature, but the micromorphological features differed in some

respects. When observed using SEM, P. roqueforti and P. paneum had finely roughened con-

idia, and conidiophores, phialides and conidia of P. paneum were each larger than those of

P. roqueforti. Based on the phylogenetic analysis of partial sequences of b-tubulin and acetyl

co-enzyme A (CoA) synthetase genes, P. roqueforti and P. paneum isolates were found to be

monophyletic species.

ª 2008 The British Mycological Society. Published by Elsevier Ltd. All rights reserved.

Introduction Until recently, P. roqueforti (s. lat.; subgenus Penicillium) was

Recent research to establish the incidence of fungal growth on

baled grass silage in Ireland have shown that up to 90 % of

bales examined had visible fungal growth present and the

most frequently isolated fungus was Penicillium roqueforti

s. str. (O’Brien 2007). P. roqueforti s. lat. is one of the most com-

mon spoilage moulds of silage (Skaar 1996; Auerbach et al.

1998; Nielsen et al. 2006; Mansfield & Kuldau 2007), but other

fungi of the genera Aspergillus, Monascus, Schizophyllum, and

Pichia are also frequent contaminants of silage (Pelhate 1977;

Skaar 1996; Mansfield & Kuldau 2007; O’Brien et al. 2005,

2007, 2008).

100.

ritish Mycological Society

considered to be a defined species but is now known to consist

of three species, P. roqueforti, P. paneum, and P. carneum, based

on ribosomal and b-tubulin DNA sequence comparisons,

RAPD profiles and secondary metabolite profiles (Boysen

et al. 1996; Samson et al. 2004; Nielsen et al. 2006). Penicillium

species recently isolated from baled grass silage in Ireland

were P. roqueforti and P. paneum, and their incidence of

occurrence was between 42–52 % and 4–5 %, respectively, of

all fungal isolates detected in three surveys undertaken

to date (O’Brien et al. 2005, 2007, 2008). There have been

only a few previous reports of P. paneum occurring on silage

(Boysen et al. 2000; Sumarah et al. 2005; Mansfield & Kuldau

. Published by Elsevier Ltd. All rights reserved.

mailto:[email protected]

http://www.elsevier.com/locate/mycres

922 O. Martin et al.

2007) and on feed grain stored under low oxygen conditions

(Haggblom 1990); originally the isolates were misidentified as

P. roqueforti but later were reclassified as P. paneum by Nielsen

et al. (2006). P. carneum is associated with meat products, such

as sausages, as well as cheese, bread, and barley (Frisvad &

Samson 2004), and perhaps silage (Samson et al. 2002). P. roque-

forti is common on substrates with high levels of organic acids,

high concentrations of carbon dioxide, and low levels of oxygen

(Samson et al. 2002). Accordingly, silage provides a very favour-

able substratum for growth of this mould. Traditionally,

P. roqueforti s. lat. has been used as a secondary starter culture

for the ripening of blue-mould cheeses, such as Gorgonzola

and Roquefort, but only P. roqueforti s. str. is currently used in

the production of blue cheeses (Nielsen et al. 2005).

The most easily observed difference in the macromorpho-

logical features of P. roqueforti compared with P. paneum is the

distinct blackish-green coloured reverse of P. roqueforti on

Czapek yeast autolysate (CYA) agar compared with the beige

to brown reverse colouration of P. paneum cultures and the

exudate droplets formed on CYA by P. paneum cultures

(Frisvad & Samson 2004). However, P. roqueforti and P. paneum

micromorphological features have been reported to be indis-

tinguishable (Frisvad & Samson 2004).

Although many studies have examined the morphological

and cultural features of P. roqueforti and P. paneum on a wide

variety of substrates (Boysen et al. 1996; Pitt 2000; Samson

et al. 2002; Frisvad & Samson 2004), and some studies have

described their molecular characteristics (Boysen et al. 1996;

Skouboe et al. 1999; Samson et al. 2004), no study, to date,

has described these features in Penicillium isolates from grass

silage to any great extent. The aim of this study was to charac-

terise the morphological, cultural, and molecular characters

of P. roqueforti and P. paneum isolated exclusively from baled

grass silage in Ireland. Their molecular characterisation was

based on the partial sequences of b-tubulin and acetyl co-

enzyme A (CoA) synthetase genes.

Materials and methods

Sample collection and isolate selection

The incidence of fungal growth on baled grass silage (n¼ 464

bales) on Irish farms (n¼ 235 farms) was recorded in three

separate studies undertaken in March 2003 (O’Brien et al.

2005), from November 2003 to March 2004 (O’Brien et al. 2007),

and in February 2004 (O’Brien et al. 2008). A total of 2277 visible

fungal colonies were enumerated on these bales and 1190 fun-

gal colonies were sampled and cultured following an estab-

lished protocol (O’Brien et al. 2005). Isolates were maintained

throughout the study on malt extract agar (MEA) plates (Oxoid,

Basingstoke) at 2–4 �C, in darkness. Of the fungal isolates, 830

were identified as Penicillium roqueforti and 78 as P. paneum by

their macro- and micromorphological features described by

Pitt (2000) (see below); results were confirmed based on liquid

chromatography-ultra violet (LC-UV) and liquid chromatogra-

phy-mass-spectrometry (LC-MS) analysis of secondary metab-

olites produced by both fungi (O’Brien et al. 2006).

A subset of the P. roqueforti isolates (n¼ 237) were ran-

domly selected and their macromorphological features

were examined in more detail (see Supplementary Material

Table S1). Subgroups of these isolates were randomly se-

lected for micromophological (n¼ 38) and molecular analyses

(n¼ 38; Table S1). In the case of P. paneum, the macromorpho-

logical features of all 78 isolates were examined and the

micromorphological features and molecular characteristics

of randomly chosen subsets of isolates (n¼ 20 and 15, respec-

tively) were analysed (see Supplementary Material Table S2).

These fungal isolates selected for morphological and molec-

ular characterisation were thus sourced from 119 bales that

were obtained from 93 farms countrywide.

Representative isolates of P. roqueforti (n¼ 28) and P. pan-

eum (n¼ 4) have been deposited at IBT, Culture Collection at

the Centre for Microbial Biotechnology, Technical University

of Denmark, Lyngby.

Morphological features of penicillia

Using the media and growth conditions specified by Pitt (2000),

a wide range of morphological features and growth rates of

Penicillium roqueforti (n¼ 237 isolates) and P. paneum (n¼ 78 iso-

lates) were examined on CYA at 5, 25 and 37 �C; MEA at 25 �C;

and 25 % glycerol–nitrate agar (G25N) at 25 �C incubated for 7 d

in darkness. The isolates were also inoculated onto yeast ex-

tract–sucrose (YES) agar and incubated at 25 �C for 7 d (Frisvad

& Samson 2004). Colony colours of isolates on all media were

determined with reference to Kornerup & Wansher (1978) un-

der artificial flood lighting [daylight bulbs (4� 100 W), Cooper

Lighting, Doncastor, UK]. Thirty-eight P. roqueforti and 20 P. pan-

eum isolates grown on MEA were examined microscopically

(Olympus BX41) at magnifications of �400 and �1000. Mea-

surements of conidiophores and conidia were made from

mounts in lacto-fuchsin; they are presented as means with

extremes in brackets. In addition, 237 P. roqueforti and 78 P.

paneum isolates were assessed for their ability to grow in Cza-

pek–Dox liquid medium (for formulation see Harrigan 1998)

supplemented with 0.5 % acetic acid (ca pH 3.5), after Engel &

Teuber (1978). Photographs of cultures were taken with

a Nikon 4500 digital camera under artificial flood lighting (as

above) and light micrographs of conidiophores and conidia

were taken with the specimens mounted in clear lactophenol

medium.

Representative isolates were examined using a JEOL 5410

Cryo-Scanning Microscope (Cryo-SEM). For cryo-SEM prepara-

tions, conidia from 7-d-old cultures on MEA were transferred

to an aluminium stub using double-sided adhesive tape. The

specimens were flash frozen (�190 to �212 �C) in slushed

liquid nitrogen under vacuum, transferred to the preparation

chamber, and under vacuum, equilibrated to �170 �C. The

specimens were then sputter coated in the preparation

chamber for 2 min at 2 psi (argon atmosphere) at �170 �C.

Specimens were then transferred to the sample stage and

examined at 10 kV at a working distance of 8–9 mm.

DNA extraction and purification

Fungal cultures used for DNA extraction were derived from

single conidia. Conidia were scraped from the MEA surface

and placed in 2 ml sterile water containing 0.1 % Tween� 80

(Sigma, St Louis, MO). The spore suspension (0.1 ml) was

Characterisation of Penicillium roqueforti and Penicillium paneum isolated from silage 923

spread on the surface of MEA (40 g l�1; Oxoid, Basingstoke) and

incubated for 18–24 h at 25 �C. Single germinating conidia

were isolated and transferred to CYA containing 0.005 %

chloramphenicol (Sigma, St Louis, MO) and 0.005 % chlortetra-

cycline (Sigma, St Louis, MO) and incubated for 3–5 d at 25 �C.

Plugs (6 mm diam) from actively growing Penicillium cultures

were transferred to liquid CYA medium (Boysen et al. 1996)

and incubated at 25 �C in darkness. After 7 d, the mycelium

was harvested by filtration through Whatman� No.1 filter

paper (Whatman, Maidstone, UK) and transferred into sterile

2 ml vials. The mycelium was rinsed with 1 ml TE buffer

(10 mM Tris–HCl, 1 mM EDTA) and excess liquid was removed

following centrifugation (10 000 g, 10 min). The mycelium

was frozen (�85 �C for ca 2 h) and freeze-dried (Labconco

bulk tray drier, Kansas City, MO) at �40 �C for 24 h. A sterile

ball bearing (4 mm diam) was added to each tube and myce-

lium was ground to a fine powder by ribolysis for 40 s at speed

6 using a Hybaid Ribolyser Fast Prep FP120 (Hybaid, Vista, CA).

DNA was extracted from the resulting powder using the proto-

col described in Sambrook & Russell (2001) (and a starting vol-

ume of 700 ml CTAB extraction buffer). Resulting DNA pellets

were rinsed twice in a 70 % (v/v) ethanol, air-dried for

15 min, and resuspended in 100 ml TE buffer. The DNA concen-

tration was measured by UV spectrophotometry (NanoDrop�

ND-1000, Labtech International, Ringmer).

PCR amplification and sequence analysis

Fragments of the genes encoding for b-tubulin and acetyl CoA

synthetase were PCR-amplified using the primer pairs Bt2a

and Bt2b (Glass & Donaldson 1995) and acuA-2F and acuA-1R

(Scott et al. 2004), respectively. Reaction mixtures contained

60–75 ng genomic DNA, 1� PCR buffer and 1 unit Taq DNA

polymerase (Invitrogen, CA), 160 mM of each dNTP, 0.5 or

0.4 mM of each of the forward and reverse primer (for b-tubulin

and acetyl-CoA synthetase, respectively) and 2 mM MgCl2, in

a total volume of 100 ml. Amplifications were performed in

a Peltier Thermal Cycler (PTC) –200 DNA Engine (MJ Research,

Waterstone, MA). The b-tubulin and acetyl CoA synthetase

amplification procedures were as described by Samson et al.

(2004) and Scott et al. (2004), respectively.

PCR products (10 ml) were electrophoresed through 1.5 %

(w/v) agarose gels containing 0.5 mg ml�1 ethidium bromide

and visualised using Imagemaster VDS and Liscap software

(Pharmacia Biotech, San Francisco, CA). The remainder of PCR

products were concentrated by vacuum centrifugation at

45 �C (Eppendorf Concentrator, Hamburg, Germany) to a final

volume of 40 ml. Concentrated products were electrophoresed

through a 1.5 % agarose gel and the bands were visualised by

UV transillumination and excised. PCR product clean up and pu-

rification was performed using the QIAEX II agarose gel extrac-

tion kit (Hilden). Purified PCR products were resuspended in

40 ml ultrapure sterile distilled water and quantified by agarose

gel electrophoresis. PCR products were sequenced by Macrogen

(Seoul) in both directions (50 and 30), using the primers used for

PCR amplification. ClustalW (Thompson et al. 1994) in BioEdit

Sequence Alignment Editor version 7.0.5 was used to generate

consensus sequences for each product (based on 50 and 30

sequence data) and to align the consensus sequences obtained

from different PCR products/isolates to each other and to

penicilla b-tubulin and acetyl CoA synthetase sequences in

the GenBank database (http://www.ncbi.nlm.nih.gov/blast/;

see Supplementary Material Table S3). Homologous sequences

were identified by BLASTn analysis (Altschul et al. 1997) against

the database of GenBank sequences using the NCBI nucleotide

BLAST tool (http://www.ncbi.nlm.nih.gov/blast/). Sequences

obtained in this study were deposited in GenBank [accession

nos EU090071–EU090123 (b-tubulin) and EU121525–EU121577

(acetyl CoA synthetase)].

Phylogenetic analyses

Phylogenetic analysis was performed from aligned sequences

using MP and NJ methods found in PAUP version 4.0b10 for

Macintosh (Swofford 2003). A heuristic search of the individ-

ual and combined datasets (gaps treated as missing) was

performed employing tree bisection and reconnection (TBR)

branch-swapping with MulTrees activated of 300 replicates

of simple sequence addition. The robustness of the most

parsimonous tree was evaluated by 10K BS replications with

MulTrees off and nearest neighbour interchange (NNI)

branch-swapping. Groups with a frequency of greater than

50 % were retained in the BS consensus tree and CI and RI

were calculated.

Results and discussion

Macromorphology features of Penicillium roquefortiand P. paneum

The macromorphological characteristics of both Penicillium

roqueforti and P. paneum in this study (Tables 1 and 2, respec-

tively) were broadly in agreement with the literature (Pitt

2000; Samson et al. 2002; Frisvad & Samson 2004; Boysen

1999), but some minor differences were recorded. For both Pen-

icillium species, colonies growing on CYA (Fig 1) and YES (Fig 2)

(and MEA for P. paneum; Fig 3) were normally olive to olive brown

coloured in their centre, an observation not recorded previ-

ously. However, Pitt (2000) made a similar observation in his

studies of P. roqueforti s. lat. and photographs in Frisvad & Sam-

son (2004): 139 and 147 clearly show olive brown colony centres.

The mean P. paneum colony size on CYA was approxi-

mately 20 % greater than recorded previously in the literature

(Frisvad & Samson 2004) for cultures incubated under similar

conditions. However, the formulation of CYA used in this

study (after Pitt 2000) and in Frisvad & Samson (2004) differed

in the quantity of one individual ingredient (K2HPO4), but this

would not explain the size discrepancy, as the size of P. roque-

forti colonies growing on CYA agreed with the literature.

Although P. roqueforti and P. paneum have always been

regarded as having very similar macromorphology features

(Frisvad & Samson 2004), several differences between the

two species emerged in this study, perhaps reflecting the large

numbers of isolates examined for each species. P. paneum iso-

lates grew faster on all three media (i.e. CYA, YES, and MEA)

and colony colours on CYA (Fig 1B) and YES (Fig 2B) were

a darker shade of green than were those of P. roqueforti (Figs

1A and 2A, respectively). P. roqueforti colony on CYA had

http://www.ncbi.nlm.nih.gov/blast/

http://www.ncbi.nlm.nih.gov/blast/

Table 1 – Macromorphological characteristics ofPenicillium roqueforti s. str. isolated from baled grasssilage in Ireland and comparisons with characteristicsreported in the literature for P. roqueforti s. str. and s. lat.sourced from a wide range of substrates and geographicallocations

Character Present study(n¼ 237 isolates)a,b

Literatureb,c

CYA colony

Diam (mm) (11–)40(–70) (17–)40(–77)

Colour Olive brown (centre)

to dull greend

Green

Reverse Dark green (to blacke) Blackish green

Colony texture Velutinous Velutinous

Medium buckling Absent Radially sulcatef

Colony margins Arachnoid Arachnoidg

Exudate droplets

on colony

Absent Absent

Diffusable colours Absent Absent

YES colony

Diam (mm) (30–)54(–72) 38–61

Colour Olive (centre)

to dull green

Greeng

Reverse Dull green to dark

green (to blacke)

Blackish green

Colony texture Velutinous ND

Medium buckling Wrinkled Wrinkledh

Colony margins Entire ND

Exudate droplets

on colony

Absent ND

Diffusable colours Absent ND

MEA colony

Diam (mm) (27–)50(–71) 26–43

Colour Dull greend Dull greenf

Reverse Beige to greyish green Pale to brown

to blackf

Colony margins Arachnoid Arachnoidg

CYA @ 5 �C (diam, mm) (0–)4(–11) 2–4

CYA @ 37 �C (diam, mm) No growth No growth

G25N (diam, mm) (7–)20(–25) 20–22(–28)f

Growth on 0.5 %

acetic acid

Yes Yes

CYA, Czapek yeast autolysate; YES, yeast extract–sucrose agar;

MEA, malt extract agar.

a Observations were recorded after incubation for 7 d at 25 �C,

unless otherwise stated.

b Measurements are presented as means with extremes in

brackets.

c Primarily adapted from Frisvad & Samson (2004) unless stated

otherwise; ND, not described.

d ca 10 % of isolates were distinctively olive brown to greyish

green.

e Observation recorded after incubation for 14 d at 25 �C.

f From Pitt (2000).

g From Boysen et al. (1996).

h From photograph (in Frisvad & Samson 2004).

Table 2 – Macromorphological characteristics ofPenicillium paneum isolated from baled grass silage inIreland and comparisons with characteristics reported inthe literature for P. paneum sourced from a wide range ofsubstrates and geographical locations

Character Present study(n¼ 78 isolates)a,b

Literaturec

CYA colony

Diam (mm) (30–)48(–60) 38–41


to dull green to

dark green

Blue green

to green

Reverse Greyish orange (to

brownish oranged)

Beige to brown

Colony texture Velutinous Velutinous

Medium buckling Irregular wrinkling ND

Colony margins Arachnoid to entire Entiree

Exudate droplets

on colony

Clear to olive brown

(ca 50 % of isolates)

Copious, clear

Diffusable colours Absent Absent

YES colony

Diam (mm) (44–)60(–73) 52–71


to dull green to

dark green

Bluish-grey–

greene

Reverse Beige to blond (to

yellowish greyd)

Cream

yellow/beigef

Colony texture Velutinous ND

Medium buckling Wrinkled Wrinkled

Colony margins Entire Entiree

Exudate droplets

on colony

Absent ND

Diffusable colours Absent ND

MEA colony

Diam (mm) (39–)56(–69) 43–67

Colour Olive to jade green Greeng

Reverse Beige to greyish green ND

Colony margins Arachnoid Entiree

CYA @ 5 �C (diam, mm) (0–)2(–7) 2–4

CYA @ 37 �C (diam, mm) No growth No growth

G25N (diam, mm) (0–)15(–27) ND

Growth on 0.5 %

acetic acid

Yes Yes

CYA, Czapek yeast autolysate; YES, yeast extract–sucrose agar;

MEA, malt extract agar.

a Observations were recorded after incubation for 7 d at 25 �C,

unless otherwise stated.

b Measurements are presented as means with extremes in

brackets.

c Primarily adapted from Frisvad & Samson (2004) unless stated

otherwise; ND, not described.

d Observation recorded after incubation for 14 d at 25 �C.

e From Boysen et al. (1996).

f Often turns to strawberry red with age and the colour diffuses

into the medium (after Frisvad & Samson 2004).

g From Boysen (1999).


a distinctive dark green to black coloured reverse after 7 d

incubation (Fig 1A) compared with the intense orange brown

observed for P. paneum (Fig 1B), and colours approximating

these have been previously observed by Boysen et al. (1996)

and Frisvad & Samson (2004). Another difference between

species on CYA medium was the ability of ca 50 % of the

P. paneum isolates to produce exudate droplets (Fig 1B); none

of the P. roqueforti isolates exhibited this characteristic on

CYA.

On MEA, colony colour was dull green or greyish green for

P. roqueforti (Fig 3A) and jade green for P. paneum (Fig 3B). The

ability of isolates to grow at low aw was assessed with G25N,

Fig 1 – Colony morphology of two typical (A) Penicillium roqueforti isolates [(KE28G (top) and LS04A (bottom))] and (B) P. paneum

isolates [(OY127E (top) and LD104I2 (bottom)] cultured on CYA agar at 25 �C. Plates 1 and 2: front and reverse colony view,

respectively, after 7 d growth; plate 3, reverse colony view after 14 d growth. Petri dish size [ 9 cm diam.


which has a reduced water activity of ca 0.93. P. roqueforti was

found to grow ca 25 % quicker than P. paneum after 7 d on this

medium and this has not been reported previously in the liter-

ature. Frisvad & Samson (2004) described the reverse colour

of older P. paneum cultures on YES as turning strawberry red

with the pigment diffusing into the medium; in this study

the reverse colouration of P. paneum cultures on YES was beige

to blond with no pigments diffusing into the medium (Fig 2B).

No comprehensive comparative studies of cheese and

spoilage/silage isolates of P. roqueforti could be found in the

published literature. However, a comparison of the cultural

features of silage isolates of P. roqueforti with those of two

blue cheese isolates, highlight little, if any, differences in

either growth or macromorphological characteristics (M.O’B.,

unpubl. data).

Micromorphological features of Penicillium roquefortiand P. paneum

The micromorphological features of Penicillium roqueforti and

P. paneum (Tables 3 and 4, respectively; Fig 4) were broadly in

agreement with the literature (Pitt 2000; Samson et al. 2002;


isolates [(OY127E (top) and LD104I2 (bottom))] cultured on YES agar at 25 �C. Plates 1 and 2: front and reverse colony view,

respectively, after 7 d growth; plate 3, reverse colony view after 14 d growth. Petri dish size [ 9 cm diam.


Frisvad & Samson 2004; Boysen 1999). The micromorphologi-

cal differences between the species included P. paneum having

larger conidiophore structures (i.e. stipes, rami, metulae),

phialides, and conidia than P. roqueforti. Although the conidi-

ophores, phialides, and conidia were within the ranges listed

in the literature for both species (Pitt 2000; Frisvad & Samson

2004), P. roqueforti was at the lower end, whereas P. paneum

was at the mid to high end of these ranges.

Prior to this study, both species were regarded as having

smooth-walled conidia as described by Pitt (2000) and Frisvad

& Samson (2004). SEM of conidia from both P. roqueforti and

P. paneum isolates collected in this study show that the conidia

have a finely rough surface texture (Fig 5), agreeing with

previous observations (M.O’B., unpubl. data). P. paneum coni-

dia had a rougher surface texture than P. roqueforti (Fig 5A–B,

respectively). Under light microscopy, all P. paneum conidia

examined were observed to be finely rough surfaced, whereas

only 28 % of P. roqueforti conidia were observed to be finely

rough. There are two possible explanations as to why both

species are reported as having smooth conidia in the


isolates [(OY127E (top) and LD104I2 (bottom))] cultured on MEA at 25 �C for 7 d. Plates 1 and 2 depict the front and reverse

colony view, respectively. Petri dish size [ 9 cm diam.


literature. First, the degree of conidial roughness was very

difficult to observe by LM in this and possibly other studies,

and therefore, other researchers may have inadvertently

described the conidia as being smooth. The difficulty of ob-

serving conidial surface texture using LM has previously

been documented; for example, Peterson (2004) reported that

conidia of P. biourgeianum appeared smooth using LM but

were finely roughened using SEM. An alternative explanation

for the discrepancy in observations on conidial surface texture

is that ecotypes of P. roqueforti and P. paneum occurring on

baled silage in Ireland have finely roughened conidia, whereas

isolates previously described from other substrata do not. As

Table 3 – Micromorphological characteristics ofPenicillium roqueforti s. str. isolated from baled grasssilage in Ireland and comparisons with characteristicsreported in the literature for P. roqueforti s. str. and s. lat.sourced from a wide range of substrates and geographicallocations

Character Present study(n¼ 38 isolates)a

Literatureb

Conidiophore

Branching pattern Ter-, occassionly

bi- or quater-verticillate

Ter-, occassionly

quater-verticillate

Stipe texture Rough Rough

Length (mm) (5–)94(–252) 100–250

Rami

Length (mm) (6–)16(–40) 17–33

Texture Rough Roughc

Metulae

Length (mm) (9–)11(–15) 10–17

Phialides

Length (mm) (6–)9(–11) 8–10

Type Ampulliform Ampulliform

Conidia

Shape Globose Globose

Colour Greyish green Greend

Texture Smooth (72 %),

finely rough (28 %)

Smooth

Size (mm) (2–)3.5(–6) (2.5–)3.5–5(–6)e

a Observations were recorded after incubation on malt extract

agar for 7 d at 25 �C. Structures were examined using LM (�400–

1000 magnification). Measurements are presented as means with

extremes in brackets.

b Primarily adapted from Frisvad & Samson (2004) unless stated

otherwise.

c From Pitt (2000).

d Conidia taken from Czapek yeast autolysate.

e Extreme values reported by Shimada & Ichinoe (1998).

Table 4 – Micromorphological characteristics ofPenicillium paneum isolated from baled grass silage inIreland and comparisons with characteristics reported inthe literature for P. paneum sourced from a wide range ofsubstrates and geographical locations

Character Present study(n¼ 20 isolates)a

Literatureb

Conidiophore

Branching

pattern

Ter-, occassionly

bi- or quater-verticillate

Ter-, occassionly

quater-verticillate

Stipe texture Rough Rough

Length (mm) (17–)134(–336) 100–250

Rami

Length (mm) (6–)18(–42) 17–33

Texture Rough ND

Metulae

Length (mm) (9–)14(–20) 10–17

Phialides

Length (mm) (8–)10(–13) 8–10

Type Ampulliform Ampulliform

Conidia

Shape Globose Globose

Colour Greyish green Blue green to greenc

Texture Smooth (8 %),

Finely rough (92 %)

Smooth

Size (mm) (2.3–)4.1(–4.9) 3.5–5

a Observations were recorded after incubation on malt extract

agar for 7 d at 25 �C. Structures were examined using LM (�400–

1000 magnification). Measurements are presented as means with

extremes in brackets.

b From Frisvad & Samson (2004); ND, not described.

c Conidia taken from Czapek yeast autolysate.


no SEMs of either P. roqueforti or P. paneum conidia have been

published previously to our knowledge, comparisons with

the present study are not possible.

The stipe texture of both species was quite variable. In

P. roqueforti, 8 % of isolates examined had predominantly

smooth stipes, 8 % were finely rough, 68 % were rough, 11 %

were very rough, and 3 % were tuberculate. In the case of

P. paneum, 30 % of isolates had predominantly rough stipes,

whereas the remaining 70 % had very rough stipes. Pitt

(2000) does not mention that this variation of stipe texture

occurs in P. roqueforti. However, Shimada & Ichinoe (1998)

found the stipes of 29 P. roqueforti contaminants isolated

from blue-veined cheeses to be rough (83 % of isolates), finely

rough (3 %), and some stipes were observed with warts (14 %).

Shimada & Ichinoe (1998) also recorded P. roqueforti conidial

size (range 2.5–6 mm) to be comparable with the size range

observed in this study (i.e. 2–6 mm), whereas the size range

reported by Frisvad & Samson (2004) was narrower (i.e.

3.5–5 mm).

The identity of many of the P. roqueforti and P. paneum iso-

lates used in this study were previously confirmed by their

secondary metabolite profiles (O’Brien et al. 2006). In that

study, the range of secondary metabolites produced by both

species broadly agreed with the literature, but not all metabo-

lites were consistently produced. The micromorphological

characters of cheese isolates of P. roqueforti were very similar

to those made for silage isolates in this study (M. O’B., unpubl.

data).

Phylogenetic analysis

b-tubulin plays a role in the biosynthesis of the globular

protein tubulin, which is the basic structural constituent of

microtubules in eukaryotic cells (Deacon 1997). Glass &

Donaldson (1995) developed a series of primers for amplifying

b-tubulin from filamentous fungi (Bt2a and Bt2b), they ampli-

fied a 426–427-bp fragment from Penicillium roqueforti and

P. paneum that spanned three introns separated by protein-

coding sequences (exons). Acetyl CoA synthetase is possibly

involved in an accessory step of penicillin biosynthesis, in ad-

dition to its role in primary metabolism; Martınez-Blanco et al.

(1993) characterised the gene encoding acetyl CoA synthetase

in P. chrysogenum and showed that the coding region was

interrupted by five introns. The primers acuA-2F and

acuA-1R used in this study amplified 282-bp (P. roqueforti)

and 290-bp (P. paneum) fragments spanning introns 3 and 4

(Scott et al. 2004).

BLASTn analysis of the b-tubulin sequences identified

P. roqueforti and P. paneum as the closest homologues of our

Fig 4 – Micromorphology of Penicillium roqueforti strains KE28G (A) and LS04A (B) and P. paneum strains KY222B (C) and

OY127E (D) following growth on MEA at 25 �C for 7 d. Bars [ 9 mm.


P. roqueforti and P. paneum isolates, respectively (99–100 %

homology; results not shown). There were no acetyl CoA syn-

thetase sequences available in GenBank for either P. roqueforti

or P. paneum, so no comparisons could be made. Alignment

(using ClustalW) showed that the sequences of the b-tubulin

and acetyl CoA synthetase gene fragments were highly

conserved within the two Penicillium species used in this

study. The partial b-tubulin sequence distinguished the 38

Fig 5 – Scanning electron micrographs of Penicillium roqueforti isolates (A) KE28G and (B) KY219A and P. paneum isolates

(C) KY222B and (D) OY123A following growth on MEA at 25 �C for 7 d. Bars [ 3.5 mm.


P. roqueforti isolates examined into three groups. Fig S1 depicts

an alignment of representatives from each group and of the P.

roqueforti b-tubulin sequences in the GenBank database. A

group comprising 27 isolates differed in one bp from a group

comprising two isolates and in two bp and a single nucleotide

deletion from a group comprising nine isolates. The sequence

of the latter group was identical to one of the four published

P. roqueforti b-tubulin sequences available in GenBank (i.e.,

AY674382; isolated from mouldy baker’s yeast); the other

three published sequences (from cheese-derived P. roqueforti

isolates) had a small number of insertions/deletions and bp

substitutions to the silage isolates. Samson et al. (2004) found

that four P. roqueforti strains from blue cheeses and mouldy

baker’s yeast had up to 5 bp substitutions in b-tubulin

sequences.

The partial acetyl CoA synthetase sequence also distin-

guished three groups of P. roqueforti isolates (Fig S2); 35 isolates

comprised group 1, which differed in two bp from two isolates

that comprised group 2 and from one isolate that comprised

group 3. There was no variation in the sequence of the b-tubu-

lin and acetyl CoA synthetase gene fragments between the 15

P. paneum isolates. The P. paneum b-tubulin sequences were

identical sequences to two of the three published P. paneum

sequences available in GenBank (AY674387 and AY674389;

isolated from rye bread) and differed in a single nucleotide de-

letion from an isolate originating from mouldy baker’s yeast

(AY674388; Fig S3). Samson et al. (2004) found only 1 bp

substitution in b-tubulin sequences among three strains of

P. paneum from rye breads and mouldy baker’s yeast.

Most parsimonious trees (MPT) generated from individual

gene datasets showed compatible topologies, supporting the

analysis of these datasets in combination (data not shown).

Combined analysis of data from the partial b-tubulin and ace-

tyl CoA synthetase sequences included 53 isolates comprising

38 P. roqueforti and 15 P. paneum. An exhaustive search of the

combined dataset (728 bp, 117 parsimony-informative charac-

ters) produced 23 MPT (CI and RI of 0.883 and 0.979, respec-

tively). NJ analysis supported the overall topology from the

MP analysis and the resulting phylogenetic tree is presented

in Fig 6. The isolates assigned to each species in this study

were well supported by BS (100 %). P. paneum strains were

monophyletic and variation within P. roqueforti isolates did

not receive strong BS support (52 %). The fact that all isolates

were sourced from a common substratum in an island geo-

graphical region may explain the lack of variability within

each species.

This is the first significant record of the morphological,

cultural, and molecular characteristics of P. roqueforti and

P. paneum isolates from grass silage. Considering the impor-

tance of grass silage as a feed source for livestock in Ireland

and western Europe with some 47 M tonnes harvested annu-

ally (Wilkinson & Toivonen 2003), this description of two com-

mon spoilage and toxigenic moulds will greatly help other

investigators to correctly identify contaminants.

CN100KOY127ETN103AWH121IOY123ALD104I2OY112HDL205ADL227AKY222BWD208AG218AMO204AG216BOY112C

TN115KCE214ALK213AC204ARN203BG206BWD210BCK202BWX206AMO212AMN210ACN210ASO208ADL213BKE17EKE25DKE28GKE29IKE18FKE28IKK09BLS01GLS06HMH02FMHO8FTN12BMH07BKY219ALM210AOY129NKE19EKE18G1LS08AMH04FMH10A2

WH121FMH14C

KE11BNRRL13487

DAOM21670NRRL824

C200NRRL13485

NRRL911CBS484.84

0.005 substitutions/site

P

en

ic

illiu

m p

an

eu

mP

en

ic

illiu

m ro

qu

efo

rti

100

100

100

Penicillium dipodomyicola

Penicillium chrysogenum

Penicillium dipodomyis

Penicillium nalgiovense

Penicillium aethiopicum

0.005 substitutions/site

Fig 6 – Phylogenetic analysis of the Irish Penicillium roqueforti and P. paneum isolates, based on the sequence of b-tubulin and

acetyl CoA synthetase gene fragments. The dendrogram was derived by NJ analysis, complemented by BS analysis (10K

replicates). BS supports greater than 80 % are given above the branch points. CI [ 0.883 and RI [ 0.979. P. aethiopicum,

P. chrysogenum, P. dipodomyis, P. dipodomyicola, and P. nalgiovense were the outgroup species (for GenBank accession no, see

Table S3).


Acknowledgements

We thank Dr Tommy Gallagher, Dr Emma Teeling and

Ms Gwyneth MacMaster for their help with the phylogenetic

analysis, Brendan Bury for the preparation of SEM images

and Dr Josephine Brennan for her expertise in the laboratory.

We are grateful to farmers for permitting sampling on their

farms. A Teagasc Walsh Fellowship Research Scholarship

awarded to M.O’B. supported this study.

Supplementary material

Supplementary data associated with this article can be found

in the online version, at doi:10.1016/j.mycres.2008.01.023.

r e f e r e n c e s

Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W,Lipman DJ, 1997. Gapped BLAST and PSI-BLAST: a newgeneration of protein databases search programs. Nucleic AcidsResearch 25: 3389–3402.

Auerbach H, Oldenburg E, Weissbach F, 1998. Incidence of Peni-cillium roqueforti and roquefortine C in silages. Journal of theScience of Food and Agriculture 76: 565–572.

Boysen M, Skouboe P, Frisvad J, Rossen L, 1996. Reclassification ofthe Penicillium roqueforti group into three species on the basisof molecular genetic and biochemical profiles. Microbiology142: 541–549.

Boysen ME, 1999. Molecular identification and quantification of thePenicillium roqueforti group. PhD thesis, Swedish University ofAgricultural Sciences, Uppsala.

Boysen ME, Jacobsson KG, Schnurer J, 2000. Molecular identifica-tion of species from the Penicillium roqueforti group associated

http://dx.doi.org/doi:10.1016/j.mycres.2008.01.023


with spoiled animal feed. Applied and Environmental Microbiol-ogy 66: 1523–1526.

Deacon JW, 1997. Modern Mycology, 3rd edn. Blackwell Science,Oxford.

Engel G, Teuber M, 1978. Simple aid for the identification of Pen-icillium roqueforti Thom. European Journal of Applied Microbiologyand Biotechnology 6: 107–111.

Frisvad JC, Samson RA, 2004. Polyphasic taxonomy of Penicilliumsubgenus Penicillium. A guide to identification of the food andair-borne terverticillate penicillia and their mycotoxins.Studies in Mycology 49: 1–173.

Glass NL, Donaldson GC, 1995. Development of primer setsdesigned for use with the PCR to amplify conserved genesfrom filamentous ascomycetes. Applied and EnvironmentalMicrobiology 61: 1323–1330.

Haggblom P, 1990. Isolation of roquefortine C from feed grain.Applied and Environmental Microbiology 56: 2924–2926.

Harrigan WF, 1998. Laboratory Methods in Food Microbiology, 3rdedn. Academic Press, London.

Kornerup A, Wansher JH, 1978. Methuen Handbook of Colour, 3rdedn. Eyre Methuen, London.

Mansfield MA, Kuldau GA, 2007. Microbiological and moleculardetermination of mycobiota in fresh and ensiled maize silage.Mycologia 99: 269–278.

Martınez-Blanco H, Orejas M, Reglero A, Luengo JM, Penalva MA,1993. Characterisation of the gene encoding acetyl-CoAsynthetase in Penicillium chrysogenum: conservation of intronposition in plectomycetes. Gene 130: 265–270.

Nielsen KF, Dalsgaard PW, Smedsgaard J, Larsen TO, 2005.Andrastins A-D, Penicillium roqueforti metabolites consistentlyproduced in blue-mold-ripened cheese. Journal of Agriculturaland Food Chemistry 53: 2908–2913.

Nielsen KF, Sumarah MW, Frisvad JC, Miller JD, 2006. Productionof metabolites from the Penicillium roqueforti complex. Journalof Agricultural and Food Chemistry 54: 3756–3763.

O’Brien M, 2007. The mycobiota of baled grass silage in Ireland. PhDthesis, University College Dublin.

O’Brien M, O’Kiely P, Forristal PD, Fuller HT, 2005. Fungi isolatedfrom contaminated baled grass silage on farms in the IrishMidlands. FEMS Microbiology Letters 247: 131–135.

O’Brien M, Nielsen KF, O’Kiely P, Forristal PD, Fuller HT,Frisvad JC, 2006. Mycotoxins and other secondary metabolitesproduced in vitro by Penicillium paneum Frisvad and Penicilliumroqueforti Thom isolated from baled grass silage in Ireland.Journal of Agricultural and Food Chemistry 54: 9268–9276.

O’Brien M, O’Kiely P, Forristal PD, Fuller HT, 2007. Visible fungalgrowth on baled silage during the winter feeding season inIreland and silage characteristics associated with the

occurrence of fungi. Animal Feed Science and Technology 139:234–256.

O’Brien M, O’Kiely P, Forristal PD, Fuller HT, 2008. Fungalcontamination of big-bale grass silage on Irish farms:predominant mould and yeast species and features of balesand silage. Grass and Forage Science 63: 121–137.

Pelhate J, 1977. Maize silage: incidence of moulds during conser-vation. Folia Veterinaria Latina 7: 1–16.

Peterson SW, 2004. Multilocus DNA sequence analysis shows thatPenicillium biourgeianum is a distinct species closely related to P.brevicompactum and P. olsonii. Mycological Research 108: 434–440.

Pitt JI, 2000. A Laboratory Guide to Common Penicillium Species, 3rdedn. Food Science Australia, North Ryde, NSW.

Sambrook J, Russell DW, 2001. Molecular Cloning: a laboratorymanual, vol. 1. Cold Spring Harbor Laboratory Press, New York.

Samson RA, Hoekstra ES, Frisvad JC, Filtenborg O, 2002. Introduc-tion to Food- and Airborne Fungi, 6th edn. Centraalbureau voorSchimmelcultures, Utrecht.

Samson RA, Seifert KA, Kuijpers AFA, Houbraken JAMP,Frisvad JC, 2004. Phylogenetic analysis of Penicillium subgenusPenicillium using partial b-tubulin sequences. Studies inMycology 49: 175–200.

Scott J, Untereiner WA, Wong B, Straus NA, Malloch D, 2004.Genotype variation in Penicillium chysogenum from indoorenvironments. Mycologia 96: 1095–1105.

Shimada T, Ichinoe M, 1998. Fungal species from imported anddomestic mold-ripened cheese. Journal of the Food HygienicSociety of Japan 39: 350–356.

Skaar I, 1996. Mycological survey and characterisation of the mycobiotaof big bale grass silage in Norway. PhD thesis, Norwegian Collegeof Veterinary Medicine, Oslo.

Skouboe P, Frisvad JC, Taylor JW, Lauritsen D, Boysen M, Rossen L,1999. Phylogenetic analysis of nucleotide sequences from theITS region of terverticillate Penicillium species. MycologicalResearch 103: 873–881.

Sumarah MW, Miller JD, Blackwell BA, 2005. Isolation andmetabolite production by Penicillium roqueforti, P. paneum, andP. crustosum isolated in Canada. Mycopathologia 159: 571–577.

Swofford DL, 2003. PAUP: phylogenetic analysis using parsimony.Version 4. Sinnauer Associates, Sunderland, MA.

Thompson JD, Higgins DG, Gibson TJ, 1994. CLUSTAL W: im-proving the sensitivity of progressive multiple sequencealignment through sequence weighting, position-specific gappenalties and weight matrix choice. Nucleic Acids Research 22:4673–4680.

Wilkinson JM, Toivonen MI, 2003. World Silage: a survey offorage conservation around the world. Chalcombe Publications,Lincoln.

Documents

Morphological and molecular characterisation of Penicillium roqueforti and P. paneum isolated from baled grass silage