Direct identification of hydrophobins and their processing
in Trichoderma using intact-cell MALDI-TOF MS
Torsten Neuhof
1
, Ralf Dieckmann
1,
*, Irina S. Druzhinina
2
, Christian P. Kubicek
2
,
Tiina Nakari-Seta
¨
la
¨
3
, Merja Penttila
¨
3
and Hans von Do
¨
hren
1
1 TU Berlin, Institut fu
¨
r Chemie, FG Biochemie und Molekulare Biologie, Berlin, Germany
2 FB Gentechnik und Angewandte Biochemie, Institut fu
¨
r Verfahrenstechnik, Umwelttechnik und Technische Biowissenschaften, TU Wien,
Vienna, Austria
3 VTT Technical Research Centre of Finland, Espoo, Finland

hren, TU Berlin, Institut fu
¨
r
Chemie, FG Biochemie und Molekulare
Biologie, Franklinstr. 29, 10587 Berlin,
Germany
Fax: +49 30 314 24783
Tel: +49 30 314 22697
E-mail:
*Present address
AnagnosTec, Gesellschaft fu
¨
r Analytische
Biochemie und Diagnostik mbH, Potsdam-
Golm, Germany
(Received 19 September 2006, revised 27
November 2006, accepted 6 December
2006)
doi:10.1111/j.1742-4658.2007.05636.x
Intact-cell MS (ICMS) was applied for the direct detection of hydropho-
bins in various species and strains of Hypocrea ⁄ Trichoderma. In both myce-
lia and spores, dominating peaks were identified as hydrophobins by
detecting mass shifts of 8 Da of reduced and unreduced forms, the analysis
of knockout mutants, and comparison with protein databases. Strain-speci-
fic processing was observed in the case of Hypocrea jecorina (anamorph
Trichoderma reesei). An analysis of 32 strains comprising 29 different spe-
cies of Trichoderma and Hypocrea showed hydrophobin patterns that were
specific at both at the species and isolate (subspecies) levels. The method
therefore permits rapid and direct detection of hydrophobin class II com-
positions and may also provide a means to identify Trichoderma (and other

hydrophobins. Besides the well-known HFB1 and
HFB2, the HFB3 hydrophobin has been identified by
cloning the corresponding gene [13] and further char-
acterization of the protein [14]. We here identified
HFB4, HFB5, and HFB6. In addition, two hydropho-
bin-encloding EST sequences were retrieved from
the TrichEST database ():
one encoding an ortholog of HFB3 from T. longibra-
chiatum (L22T11P141R12690, L14T53P137R01628,
L22T11P138R12431, and L22T11P137R12300), and
the other one encoding an ortholog of HFB1 of
T. atroviride (L12T11P119R10608). Their sequence
relationships and putative processing sites are illustra-
ted in the alignment given in Fig. 1.
ICMS analysis of Trichoderma
Several strains of Trichoderma were studied initially to
examine the effectiveness of ICMS as an analytical
method for distinguishing different species of Trichoder-
ma. A rapid analytical procedure based on ICMS was
established in order to characterize the low-molecular-
weight proteometric (up to 20 000 Da) and peptidomet-
ric (up to 2000 Da) profiles at the same time. Thirty-two
Trichoderma strains belonging to various species were
subcultivated on agar plates at an incubation tem-
perature of 25 °C and analyzed without further pre-
treatments as described in Experimental procedures.
Vegetative mycelia or spores were transferred from the
biomass growing on agar plates directly to the MALDI
sample plate and mixed with an acidic matrix in an
organic solvent mixture. An estimated 10

with a molecular mass of 7540.58 Da, which is further
reduced by disulfide bond formation to 7532.58 Da
[15]. The 86 residue HFB2 precursor with a mass of
8766.28 Da is processed to a 71 amino acid peptide
with a calculated molecular mass of 7196.42 kDa, and
further reduced by disulfide formation to 7188.42 Da
[15]. Both hydrophobins were detected as [MH]
+
sig-
nals of the oxidized forms (Hfb1, m ⁄ z 7533; Hfb2,
m ⁄ z 7189). A minor peak of m ⁄ z 7041 presumably cor-
responds to the processed Hfb2 lacking the terminal
Phe (7041.24 Da) (Fig. 3).
The same peaks were observed in the spectra
obtained from isolated reference substances of HFB1
and HFB2 proteins (Fig. 3C). A second minor peak of
m ⁄ z 7229 correlates with oxidized HFB2 cleaved at
Ala13 lacking the N-terminal Phe. This tentative corre-
Fig. 2. Intact-cell MALDI-TOF mass spectra
of mycelia and spores of Trichoderma
strains. The masses 7347 and 7494 of
T. atroviride spores correlate with two proc-
essed products of the spore hydrophobin
SRH1 [16] cleaved at the N-terminal MQFSI-
VALFATGALA site and the C-terminal Phe,
respectively.
Fig. 3. Intact-cell MALDI-TOF mass spectra
of H. jecorina strain QM 9414 (D), and the
mutant strains QM 9414 Dhfb1 (B) and
QM 9414 Dhfb1Dhfb2 (A). HFB I and HFB II

⁄ hfb2
–
did
not show HFB1 and HFB2 signals (Fig. 3A), whereas
Fig. 4. Intact-cell MALDI-TOF spectra of mycelia (A, C, E) and sporulating mycelia (B, D, F) of three strains of H. jecorina grown on malt
agar. The masses displayed have an error of about 0.1%, so peaks of 7232 (A), 7237 (E) and 7234 (F) represent similar peptides. Strain 618
mycelia (A) show a variety of peaks, in contrast to strains 665 and 937, shown in (C) and (E). However, there are only few similarities: 7232
and 7237 in (A) and (E), or 7509 and 7514 in (A) and (C). An obvious shift is the appearance of higher mass peaks upon sporulation, presum-
ably related to the only large hydrophobin of H. jecorina.
Trichoderma hydrophobins T. Neuhof et al.
844 FEBS Journal 274 (2007) 841–852 ª 2007 The Authors Journal compilation ª 2007 FEBS
in the knockout mutant hfbI
–
, only the respective mass
peak was missing (Fig. 3B).
Deviating post-translational processing
of hydrophobins in H. jecorina strains
To investigate strain diversity with respect to meta-
bolite production and low-mass proteomics by ICMS,
three phylogenetically described isolates of H. jecorina
were studied. As shown in Fig. 4, spectra of mycelia
and sporulating mycelia directly taken from the
plates after 1 or 3 days, respectively, differ in peak
compositions and intensities. Surprisingly, all spectra
differ with respect to strain QM 9414. Strain
CPK 618 mycelia show a prominent signal of
m ⁄ z 7232 (Fig. 4A), which disappears in the sporula-
tion process, with new signals of m ⁄ z 8859, 8802 and
7521 appearing (Fig. 4B). To obtain a preliminary
correlation of observed masses with hydrophobin

Although they are rather speculative, the interpreted
masses agree with verified cleavage sites observed for
HFB1 and HFB2 and known sites for signal peptidas-
es and Kex2-type peptidases (Table 1). Verification of
these assessments by tryptic digestion and sequencing
is in progress.
Hydrophobin patterns in other T. atroviride and
T. longibrachiatum strains
T. atroviride
A hydrophobin gene (srh1) encoding a class II hydro-
phobin with phylogenetic similarity to H. jecorina
HFB2 (I. S. Druzhinina and C. P. Kubicek, unpub-
lished results) has been found in T. atroviride (therein
named ‘T. harzianum’ [16]). The main components of
the sporulating mycelia of the same strain (T. atrovi-
ride P1) could indeed be assigned to this hydrophobin,
assuming similar post-translational processing as for
the H. jecorina HFB2 (Fig. 2, top spectrum). The
peaks at m ⁄ z 7499 and 7352 correspond to the proc-
essed spore hydrophobin SRH1 with the cleaved signal
sequence MQFSIVALFATGALA and an additional
C-terminal Phe cleavage, respectively, including loss of
8 Da for the disulfide bonds. A minor peak at
m ⁄ z 7741 could be tentatively correlated with the
SRH1 hydrophobin with N-terminal cleavage of
MQFSIVA, C-terminal processing following the two
Glu residues of AAAQGTF, and four disulfide bonds.
Interestingly, these peaks could not be detected in
vegetative mycelia of T. atroviride P1 (Fig. 2), which
displayed a similar peak pattern, but with slightly dif-

peaks with unique molecular masses. It is therefore
interesting to note that even phylogenetically closely
related species (such as T. hamatum and T. asperel-
lum,orT. harzianum and T. fulvum [17], or T. fas-
ciculatum and T. strictipile, which were recently
revised to be actually the same species [18]), could be
clearly separated. This is in accordance with the data
on H. jecorina shown above, and implies that hydro-
phobin fingerprints can in fact distinguish isolates at
the subspecies level. All spectra are compiled in sup-
plementary Fig. S2.
Discussion
Hydrophobin patterns
Genome sequencing of filamentous fungi has revealed
the presence of multiple hydrophobin genes in filamen-
tous fungi. We here report the the sequences of four
Trichoderma hydrophobins T. Neuhof et al.
846 FEBS Journal 274 (2007) 841–852 ª 2007 The Authors Journal compilation ª 2007 FEBS
new type II hydrophobins for H. jecorina, in addition
to the known HFB1 and HFB2. Likewise, we identi-
fied new hydrophobins in T. atroviride in addition to
the known sporulation-specific one, and in T. longibra-
chiatum. Direct MS analysis of mycelia in differing
physiologic states provides evidence for differential
expression of these genes in relation to the morpho-
logic state. However, there is no clear match of the
observed mass peaks to the predicted propeptides
expected to originate from cleavage of signal peptides.
Instead, further processing has been observed, as has
been demonstrated before from N-terminal sequence

protease
HFB4 8862 TVA ⁄ LFI – Predicted cleavage
within signal peptide
H. jecorina
strain 665
HFB1 7147 EDR ⁄ SNG CQT ⁄ AVG Confirmed and
predicted non-Kex2
sites
HFBII 6999 ALA ⁄ AVC CQK ⁄ AIG Confirmed signal
peptide site, predicted
non-Kex2 site
T. atroviride P1 SRH1 7499 ALA ⁄ SVS – Predicted signal
peptide cleavage
SRH1 7352 ALA ⁄ SVS – F Predicted signal
peptide cleavage,
predicted C-terminal
cleavage
SRH1 7741 IVA ⁄ LFA EE ⁄ AQG Predicted alternative
signal peptide
cleavage,
predicted C-terminal
cleavage
HFB1 7743 AIA ⁄ GPV CQT ⁄ AVG Predicted signal
peptide cleavage,
predicted C-terminal
cleavage
T. longibrachiatum HFB3 7242 RRR ⁄ DQA – Predicted Kex2 site
cleavage
T. Neuhof et al. Trichoderma hydrophobins
FEBS Journal 274 (2007) 841–852 ª 2007 The Authors Journal compilation ª 2007 FEBS 847

ing the sequences by
TRICHOKEY [28] and TRICHOBLAST [29].
Species m ⁄ zm⁄ zm⁄ zm⁄ zm⁄ zm⁄ zm⁄ zm⁄ zm⁄ z
H. tawa CBS 246.63 7014 7214 7718 7895
H. semiorbis CBS 244.63 7024 7570
H. hunua CBS 238.63 7186 7269 7432 7594 7727
H. gelatinosa ATCC 7476 7167 7496
H. citrina CBS 977.69 7171 7440
H. aureoviridis CBS 254.63 7214 7309 7473 7611
T. strigosum CBS 348.93 7129 7490 7634
T. tomentosum CBS 349.93 7171 7399 7501
T. strictipilis CBS 347.93 7155 7346 7490
T. longipile CBS 340.93 7257 7313 7517
T. fasciculatum CBS 118.72
a
7120 7286
H. minutispora CBS 342.93 7282 7459
T. pubescens DAOM 162162 7255 7618
T. viride ATCC 28020 7141 7603
T. stromaticum CBS 101875 7093
T. flavofuscum CBS 248.59
b
7176 7307
T. citrinoviride IMI 232088 7021 7293 7584 7748
T. brevicompactum CBS 109720 7141
T. hamatum CBS 393.92 7055 7342 7616
T. asperellum CBS 358.97 7156 7368 7522
T. croceum DAOM 167068
c
7104 7188 7284

T. croceum is a synonym of T. polysporum
(¼H. pachybasioides).
d
Identified as HFB3 type with processing at the Arg site: MQFLAVAALLFTAAFAAPSSEAHGLRRR.
e
Identified as
processed HFB2 (see Fig. 2).
f
Identified as processed HFB1 (see Fig. 2).
g
Identified as processed HFB1-type hydrophobin: GPVEVRTGGG-
SICPDGLFSNPQCCDTQLLGIIGLGCEVPSQTPRDGADFKNICAKTGDQALCCVLPIAGQDLLCQA.
h
Identified as processed SRH1 (see Fig. 1).
i
Identified as processed SRH1: LASVSVCPNGLYSNPQCCGANVLGVAALDCHTPRVDVLTGPIFQAVCAAEGGKQPLCCVVPVAGQDLLCEE.
Trichoderma hydrophobins T. Neuhof et al.
848 FEBS Journal 274 (2007) 841–852 ª 2007 The Authors Journal compilation ª 2007 FEBS
whereas HFB2 is not cleaved further in the N-teminal
region. Mass spectra provide evidence for a C-terminal
Phe cleavage for HFB2.
The ICMS spectra presented here provide evidence
for similar cleavage patterns of hydrophobins HFBIII
of H. jecorina and T. longibrachiatum, and SRH1 and
HFB1 of T. atroviride, involving signal peptides, Kex2-
type processing, and C-terminal amino acid cleavage
(Table 2). In addition, the recorded masses provide
evidence for alternative processing reactions. One reac-
tion concerns alternative signal peptide cleavage sites
at Ala13 for H. jecorina HFB2 and at Ala7 for T. atr-

specific hydrophobins in Cladosporium fulvum is
dependent on the stages of the plant infection pro-
cess, the hydrophobins being either retained on coni-
dia and aerial structures or being excreted [23]. In
Magnaporthe grisea, it has been demonstrated that
the formation of disulfide linkages is required for
secretion and cell wall localization [24].
Indeed, we have shown here that sporulating and
nonsporulating mycelia of several species differ in
hydrophobin composition. Unexpectedly, the patterns
observed indicate diverse cleavage reactions of the
respective prepropeptides. These patterns are unlikely
to be proteolytic artefacts of extraction, as proteases
are unlikely to be active in methanol ⁄ acetonitrile mix-
tures. As MALDI-TOF MS involves an especially gen-
tle ionization process, cleavages of peptide bonds are
generally not observed. Differences in hydrophobin
processing are thus interpreted as being dependent on
the presence and concentrations of specific proteinases
acting on the respective propeptides.
Hydrophobins as biomarkers for ICMS
of filamentous fungi
Hydrophobins are proposed to be suitable ICMS bio-
markers for the following reasons: (a) fungi contain a
set of hydrophobin genes, generally with developmen-
tally regulated expression; (b) the hydrophobic pep-
tides can be selectively dissolved and rapidly analyzed
by MALDI-TOF MS; (c) hydrophobin patterns are
diverse, due to post-translational processing; (d) the
presence of the characteristic four disulfide bonds can

Hypocrea jecorina QM 9414 and its mutants were cultiva-
ted in liquid cultures on microtiter plates (200 lL volume)
for 4 days in buffered minimal medium [15] complemented
with 3% glucose and 0.2% peptone. The other strains of
H. jecorina, as well as strains of other Trichoderma spp.,
were cultivated on malt extract agar (3%) at 25 °C.
Extraction and preparation of mycelia
for MALDI-TOF analysis
A few micrograms of fungal mycelia were suspended in
acetonitrile ⁄ methanol ⁄ water (1 : 1 : 1), and 1 lL of the sus-
pension was directly spotted onto target wells of a 100-posi-
tion sample plate and immediately mixed with 1 lLof
matrix solution [10 mgÆmL
)1
2,5-dihydroxybenzoic acid in
acetonitrile ⁄ methanol ⁄ water (1 : 1 : 1) and 0.3% trifluoro-
acetic acid]. The sample matrix mixture was allowed to air
dry prior to analysis. Alternatively, freeze-dried mycelium
obtained from shaken cultures or fungi grown on plates
was homogenized in 60% ethanol and centrifuged at
13 000 g using a Beckman Microfuge 11 (Beckman Coulter,
Unterscheissheim, Germany). One microliter of the protein
solution was spotted on a MALDI target plate and mixed
with matrix.
Reduction of disulfide bonds
For reduction of proteins containing disulfide bonds, cells
were suspended in 60% methanol, vortexed, and centri-
fuged at 13 000 g using a Beckman Microfuge 11, and the
supernatant was concentrated to dryness. The residual was
redissolved in 50 mm Tris ⁄ HCl (pH 8) and 1 mm dithio-

leotide) program. We used the hydrophobin class II protein
sequences of other fungal species as queries to search the
H. jecorina genome. Then, all putative hydrophobins, inclu-
ding the newly identified hydrophobin from H. jecorina,
were used to identify further proteins with similar domains,
and finally all hypothetical proteins encoding hydrophobins
from the annotated genomes of the Broad Institute (http://
www.broad.mit.edu/), Neurospora crassa, Gibberella zeae
(Fusarium graminearum) and Magnaporthe griseae, were
also used.
Calculations of monoisotopic molecular masses of hydro-
phobins was performed with the expasy proteomics server
( or the peptide mass calculator
( To
correlate observed mass peak data, masses of various
N- and C-terminally processed and oxidized peptides were
calculated and compared.
Acknowledgements
This work was supported by a fellowship from the Deut-
sche Forschungsgemeinschaft (Do270 ⁄ 10) and by the
Fifth Framework program (Quality of Life and Man-
agement of Living Resources; Project EUROFUNG 2;
QLK3-1999-00729) of the European Community. The
T. reesei genome sequencing project was funded by the
Department of Energy. The authors thank M. Salohei-
mo for helpful comments and discussions.
References
1 Linder MB, Szilvay GR, Nakari-Setala T & Penttila
¨
ME (2005) Hydrophobins: the protein-amphiphiles

9 Chen H-Y & Chen Y-C (2005) Characterization of
intact Penicillium spores by matrix-assisted laser
desorption ⁄ ionization mass spectrometry. Rapid Com-
mun Mass Spectrom 19, 3564–3568.
10 Li TY, Liu BH & Chen YC (2000) Characterization of
Aspergillus spores by matrix-assisted laser
desorption ⁄ ionization time-of-flight mass spectrometry.
Rapid Commun Mass Spectrom 14, 2393–2400.
11 Welham KJ, Domin MA, Johnson K, Jones L & Ash-
ton DS (2000) Characterization of fungal spores by laser
desorption ⁄ ionization time-of-flight mass spectrometry.
Rapid Commun Mass Spectrom 14, 307–310.
12 Welham KJ, Domin MA, Johnson K, Jones L & Ash-
ton DS (2000) Matrix-assisted laser-desorption ⁄ ioniza-
tion time-of-flight mass spectrometry and its application
to the analysis of fungal spores. Pharmacy Pharmacol
Commun 6, 107–111.
13 Penttila
¨
M, Nakari-Seta
¨
la
¨
T, Fagerstro
¨
m R, Selber K,
Kula M-R, Linder M & Tjerneld F (2000) Process for
partitioning of proteins. International Patent Applica-
tion PCT ⁄ Fi00 ⁄ 00249, 24.3.2000.
14 Rintala E, Linder M & Nakari-Seta

aggregate species to species clusters? J Zhejiang Univ Sci
B 6, 100–112.
18 Chaverri P, Castlebury LA, Overton BE & Samuels GJ
(2003) Hypocrea ⁄ Trichoderma: species with conidio-
phore elongations and green conidia. Mycologia 95,
1100–1140.
19 Bendtsen JD, Nielsen H, von Heijne G & Brunak S
(2004) Improved prediction of signal peptides: SignalP
3.0. J Mol Biol 314, 783–795.
20 Zhang Z & Henzel WJ (2004) Signal peptide prediction
based on experimentally verified cleavage sites. J Prot
Sci 13, 2819–2824.
21 Goller SP, Schoisswohl D, Baron M, Parriche M & Kubi-
cek CP (1998) Role of endoproteolytic dibasic proprotein
processing in maturation of secretory proteins in Tricho-
derma reesei. Appl Environ Microbiol 64, 3202–3208.
22 Dieckmann R, Kaesler I, Graeber I, Szewzyk U & von
Do
¨
hren H (2005) Rapid screening and dereplication of
microbial isolates from marine sponges by intact-cell-
MALDI-TOF mass spectrometry. J Appl Microbiol
Biotechnol 67, 539–548.
23 Whiteford JA, Lacroix H, Talbot NJ & Spanu PD
(2004) Stage-specific cellular localisation of two hydro-
phobins during plant infection by the pathogenic
fungus Cladosporium fulvum. Fungal Genet Biol 41,
624–631.
24 Kershaw MJ, Thornton CA, Wakley GE & Talbot NJ
(2005) Four conserved intramolecular disulphide lin-

choderma and Hypocrea identifications. Mycol Res 109,
658–660.
Supplementary material
The following supplementary material is available
online:
Doc. S1. Sequence data of hydrophobins.
Fig. S1. Dithiothreitol reduction of H. jecorina QM
9414 hydrophobins.
Fig. S2. Intact-cell MALDI-TOF spectra (6000–
10 000 m ⁄ z)ofTrichoderma ⁄ Hypocrea strains. All indi-
cated masses have been identified as hydrophobins by
an m ⁄ z shift of 8 upon dithiothreitol reduction.
This material is available as part of the online article
from
Please note: Blackwell Publishing is not responsible
for the content or functionality of any supplementary
materials supplied by the authors. Any queries (other
than missing material) should be directed to the corres-
ponding author for the article.
Trichoderma hydrophobins T. Neuhof et al.
852 FEBS Journal 274 (2007) 841–852 ª 2007 The Authors Journal compilation ª 2007 FEBS