Cellular Microbiology (2008) 10(11), 2271–2284
x
A novel Phytophthora infestans haustorium-specific
membrane protein is required for infection of potato
Anna O. Avrova,1** Petra C. Boevink,1
Vanessa Young,1 Laura J. Grenville-Briggs,2
Pieter van West,2 Paul R. J. Birch3 and
Stephen C. Whisson1*
1
Plant Pathology Programme, Scottish Crop Research
Institute, Invergowrie, Dundee DD2 5DA, UK.
2
Aberdeen Oomycete Group, College of Life Sciences
and Medicine, Institute of Medical Sciences, University
of Aberdeen, Foresterhill, Aberdeen AB24 2ZD, UK.
3
Division of Plant Sciences, College of Life Sciences,
University of Dundee at SCRI, Invergowrie, Dundee
DD2 5DA, UK.
Summary
Phytophthora infestans causes late-blight, a devastat-
ing and re-emerging disease of potato crops. During
the early stages of infection, P. infestans differentiates
infection-specific structures such as appressoria for
host epidermal cell penetration, followed by infection
vesicles, and haustoria to establish a biotrophic phase
of interaction. Here we report the cloning, from a
suppression subtractive hybridization library, of a
P. infestans gene called Pihmp1 encoding a putative
glycosylated protein with four closely spaced trans-
membrane helices. Pihmp1 expression is upregulated
in germinating cysts and in germinating cysts with
appressoria, and significantly upregulated throughout
infection of potato. Transient gene silencing of Pihmp1
led to loss of pathogenicity and indicated involvement
of this gene in the penetration and early infection
processes of P. infestans. P. infestans transformants
expressing a Pihmp1::monomeric red fluorescent
protein (mRFP) fusion demonstrated that Pihmp1 was
translated in germinating sporangia, germinating
cysts and appressoria, accumulated in the appresso-
rium, and was located at the haustorial membrane
during infection. Furthermore, we discovered that
haustorial structures are formed over a 3 h period,
maturing for up to 12 h, and that their formation is
initiated only at sites on the surface of intercellular
Received 29 April, 2008; revised 30 June, 2008; accepted 7 July,
2008. For correspondence. *E-mail Steve.Whisson@scri.ac.uk; Tel.
(+44) 1382 562731; Fax (+44) 1382 562426; **E-mail Anna.Avrova@
scri.ac.uk; Tel. (+44) 1382 562731; Fax (+44) 1382 562426.
© 2008 Scottish Crop Research Institute
Journal compilation © 2008 Blackwell Publishing Ltd
doi:10.1111/j.1462-5822.2008.01206.
First published online 5 August 2008
hyphae where Pihmp1::mRFP is localized. We propose
that Pihmp1 is an integral membrane protein that pro-
vides physical stability to the plasma membrane of
P. infestans infection structures. We have provided the
first evidence that the surface of oomycete haustoria
possess proteins specific to these biotrophic struc-
tures, and that formation of biotrophic structures
(infection vesicles and haustoria) is essential to suc-
cessful host colonization by P. infestans.
Introduction
Filamentous eukaryotic plant pathogenic microorganisms
such as fungi and oomycetes frequently form highly spe-
cialized structures during invasion and colonization of
their hosts. For many plant pathogens, host invasion com-
mences with the differentiation of a domed or swollen
structure, called an appressorium, at the terminus of a
germination tube or hypha from a spore (Talbot, 2003;
Grenville-Briggs and van West, 2005). The appressorium
secretes adhesive compounds to maintain contact
between host and pathogen during infection (Tucker and
Talbot, 2001; Grenville-Briggs and van West, 2005), and
can generate mechanical force (Howard et al., 1991;
Bechinger et al., 1999) and/or hydrolytic enzymes to pene-
trate host cell walls (Francis et al., 1996; Pryce-Jones
et al., 1999).
After invading a compatible host, fungal and oomycete
hyphae grow either intercellularly or intracellularly,
with or without biotrophic structures called haustoria
(Mendgen and Hahn, 2002). For example, Cladosporium
fulvum produces entirely intercellular hyphae without
haustoria, Magnaporthe grisea grows in planta as intra-
cellular hyphae without haustoria (Kankanala et al.,
2007), while obligate biotrophs, such as the rust fungi,
produce intercellular hyphae with intracellular haustoria.
Haustoria are formed by invagination of the host cell
membrane and are thus in intimate contact with the host
cell. The haustorial plasma membrane of fungal plant
pathogens is differentiated from the intercellular hyphal
plasma membrane and is distinct in its possession of
proteins for the transport of nutrients such as sugars
and amino acids (Hahn et al., 1997; Struck et al., 1998;
Voegele et al., 2001). Haustoria are also the site of
secretion for fungal and oomycete effector proteins,
some of which may act inside the host cell to modulate
2272 A. O. Avrova et al.
the host cell environment (Catanzariti et al., 2006;
Kamoun, 2006; Whisson et al., 2007).
Phytophthora infestans is an oomycete plant pathogen,
best known for its role in precipitating the Irish potato
famines in the mid-19th century. It remains a major con-
straint to potato production worldwide. Although sharing
morphological similarities with fungi such as hyphae and
spores, oomycetes are more closely allied to the stra-
menopiles among the heterokonts, and deeply rooted with
the alveolates (Baldauf et al., 2000; Burki et al., 2007). As
a model oomycete, the life cycle of P. infestans is well
characterized (van West and Vleeshouwers, 2004;
Avrova et al., 2007), as are its interactions with its host
plants. Almost all molecular genetic tools are available to
determine gene function in this pathogen (Birch and
Whisson, 2001; Lamour et al., 2007), including reliable
methods for transient gene silencing and occasional
stable gene silencing, reporter gene assays, and a wealth
of genetic information including greater than 80 000 ESTs
and a genome sequence (van West et al., 1999; Randall
et al., 2005; Whisson et al., 2005; 2007). This establishes
P. infestans as the model oomycete organism for studying
the pathogen molecular and biochemical processes
underlying disease development. Asexual sporangia
are the infectious dissemination propagules and may
germinate directly, or cleave to yield motile, biflagellate
zoospores. Zoospores encyst at an infection site in
response to chemical, electrical and physical signals, dis-
carding flagella and forming a cell wall (van West et al.,
2002, 2003). A germ tube emerges from the cyst or spo-
rangium, the tip of which swells to differentiate into an
appressorium and subsequently a penetration hypha.
After the plant cuticle and cell wall have been breached,
an intracellular, biotrophic infection vesicle is produced in
the epidermal cell. Intercellular hyphae grow into the
mesophyll cell layers, producing intracellular, biotrophic
finger-like haustoria as new host cells are encountered.
After 36–48 h post inoculation (hpi), the mode of interac-
tion between P. infestans and initially infected host cells
becomes necrotrophic and new haustoria are not formed.
From 72 to 96 hpi, initially infected areas of the leaf are
fully colonized and necrotic. Sporulating hyphae emerge
from stomata to release sporangia, allowing aerial dis-
semination of the pathogen. Spatially, this is visualized as
a leaf lesion with a biotrophic margin as new cells are
infected. The centre of the lesion is necrotic where initially
colonized cells have died.
Oomycete haustoria differ from those formed by fungal
plant pathogens in that cell wall and membrane differen-
tiation is less pronounced in oomycetes. For example,
fungal haustoria have a distinct neckband that seals the
extrahaustorial matrix (EHM) from the apoplast, whereas
the oomycete cell wall is continuous from the intercellular
hyphae to the haustoria (Hohl and Suter, 1976; Soylu,
Journal compilation © 2008 Blackwell Publishing Ltd, Cellular Microbiology, 10, 2271–2284
2004). Compared with fungal plant pathogens, molecular
processes underlying oomycete development and patho-
genicity are poorly understood. With the exception of
the AVR3a avirulence protein (Whisson et al., 2007), no
oomycete proteins have been demonstrated to localize
to haustoria, either for secretion or for other haustorial
functions.
Several studies have reported differentially expressed
candidate pathogenicity factors from P. infestans pre-
infection stages, as they are likely to be enriched for
molecules involved in successful penetration of the host,
and establishment of a compatible interaction (Görnhardt
et al., 2000; Beyer et al., 2002; Avrova et al., 2003; Birch
et al., 2003; Bittner-Eddy et al., 2003; Randall et al., 2005;
Judelson et al., 2008). A high proportion of oomycete
genes identified through transcript profiling techniques
are entirely novel, and identifying those with a role in
pathogenicity remains a significant challenge. Through
gene silencing in stable transformants, a small number of
P. infestans genes have been analysed at the functional
level, revealing an involvement for some in pathogenicity.
However, these have typically encoded proteins involved
in signal transduction or transcriptional regulation (Latijn-
houwers and Govers, 2003; Latijnhouwers et al., 2004;
Blanco and Judelson, 2005). These genes and their
encoded proteins may form part of the core biology of
P. infestans and their silencing may have resulted in
multiple developmental defects, including reduced
pathogenicity. Transient silencing of genes is also pos-
sible in P. infestans, and may facilitate higher-throughput
functional analysis of, and more silencing-recalcitrant,
genes than previously (Whisson et al., 2005; Grenville-
Briggs et al., 2008; Walker et al., 2008). In identifying
genes encoding proteins contributing directly to patho-
genicity, it may be expected that candidate genes encode
proteins that act at the interface between pathogen and
host. Although numerous oomycete genes, such as pro-
tease inhibitors, a glucanase inhibitor, avirulence proteins
and necrosis-inducing proteins, have been identified that
interact with host plant proteins (Rose et al., 2002; Torto
et al., 2003; Tian et al., 2004; 2005; 2007; Birch et al.,
2006), their contribution towards oomycete pathogenicity
remains to be determined.
In a broader transcript profiling study in P. infestans
targeting candidate pathogenicity genes expressed in
pre-infection stages, we selected one transcript for further
study in order to test the hypotheses that it encoded a
protein that may act at the pathogen–host interface, and
play a role in pathogenicity. Here we describe the cloning,
homologues in other oomycetes, role in pathogenicity,
and localization of a P. infestans gene encoding a pre-
dicted integral membrane protein. Translational fusion of
this gene to the monomeric red fluorescent protein
(mRFP) in stable P. infestans transformants demon-
© 2008 Scottish Crop Research Institute
 Phytophthora infestans haustorial membrane protein 2273
strated localization to the haustorial plasma membrane
during infection of potato. Accordingly, it has been named
Pihmp1, for P. infestans haustorial membrane protein 1.
The involvement of Pihmp1 in the penetration process of
P. infestans and establishment of a compatible interaction
was shown using transient gene silencing.
Results
Cloning of Pihmp1
From an existing suppression subtractive hybridization
(SSH) cDNA library (Grenville-Briggs et al., 2005) a
198 bp transcript-derived fragment (clone 2c5) was
selected for further analysis, as BLASTN and TBLASTX
similarity searches of sequence databases (see Experi-
mental procedures) revealed sequence similarity only to
expressed sequence tags (ESTs) from P. infestans (4e-9
to 5e-24 for TBLASTX; 2e-115 to 1e-124 for BLASTN), and the
genome sequences of Phytophthora sojae, Phytophthora
ramorum and Hyaloperonospora parasitica (3e-33 to 1e-40
for TBLASTX; 3e-16 to 6e-20 for BLASTN). As this section of
work was carried out before the availability of the
assembled draft P. infestans genome sequence, genomic
copy number, and the sequence of the complete open
reading frame (ORF) and promoter of Pihmp1 were deter-
mined from three positively hybridizing clones (Fig. S1)
from an existing P. infestans bacterial artificial chromo-
some (BAC) library (Whisson et al., 2001) representing
10¥ genome coverage. The Pihmp1 gene sequences
present on BAC clones 5J1 and 15J23 were identical, and
the sequence on BAC clone 1B18 was a truncated copy of
the gene. Since the release of the P. infestans genome
sequence (see Experimental Procedures for URL), both
the intact and truncated copies of the gene have been
found within the same genomic region (sequence super-
contig 1.1), with no additional copies located elsewhere in
the genome.
A TBLASTN search of the P. infestans genome database
using the predicted amino acid sequence of Pihmp1
as a query revealed 11 supercontigs containing several
homologous regions with E-values of 8 ¥ 10-43-10-172. A
TBLASTN search of the NCBI EST databases, including
human and mouse sequences (http://www.ncbi.nlm.nih.
gov/blast/) with the predicted Pihmp1 protein sequence,
identified two additional homologous ESTs (gi58106741
and gi58105907) from P. infestans germinating sporangia
with E-values of 7 ¥ 10-73 and 2 ¥ 10-37, respectively, one
homologous EST (gi58087999) from P. infestans sporu-
lating mycelium with E-value of 2 ¥ 10-43, and two homolo-
gous ESTs (gi38117833 and gi38062167) from P. sojae
with E-values of 2 ¥ 10-48 and 2 ¥ 10-51 respectively.
The Pihmp1 sequence is deposited in GenBank under
Accession No. EU680858.
© 2008 Scottish Crop Research Institute
Journal compilation © 2008 Blackwell Publishing Ltd, Cellular Microbiology, 10, 2271–2284
Pihmp1 expression is upregulated prior to and
during infection
Real-time reverse transcription polymerase chain reaction
(RT-PCR) analysis was used to characterize expression
of Pihmp1 in P. infestans in vitro stages of cultured myce-
lium, sporangia, zoospores, germinating cysts, germinat-
ing cysts with appressoria and in planta during the
infection of susceptible potato cv. Bintje with a compatible
race (isolate 88069) of P. infestans at 12, 24, 33, 48, 56
and 72 hpi. Microscopic analysis indicated that the timing
of P. infestans development during these infections
closely followed the events described by Vleeshouwers
et al. (2000) and Avrova et al. (2007). Thus, at 12 hpi,
germinating cysts, appressoria and infection hyphae were
clearly visible. At 24–33 hpi, numerous haustoria were
visible in infected tissue, indicative of the biotrophic
phase, whereas few haustoria were visible in infected
tissue by 48 hpi, indicating the transition from biotrophy to
necrotrophy. By 72 hpi, the necrotrophic phase was well
established with sporulation and necrosis clearly visible.
Primer pairs were designed to anneal to Pihmp1, PiactA
and PiipiO1 (Table S1). The PiactA gene from P. infestans
was used as a constitutively expressed endogenous
control (Grenville-Briggs et al., 2008; Judelson et al.,
2008). Expression of PiipiO1 and Pihmp1 in asexual life
cycle- and infection-stage samples was compared with
the level of their expression in a calibrator sample, which
was cDNA synthesized from RNA isolated from non-
sporulating mycelium grown in rye broth. The expression
of PiipiO1 and Pihmp1 in the mycelium cDNA samples
was assigned the value of 1.0.
The P. infestans PiipiO1 gene, known to be upregulated
in invading hyphae during infection (van West et al., 1998),
was used to confirm disease progression in inoculated leaf
samples and yielded the expected expression profile
(Fig. 1). The Pihmp1 expression profile in pre-infection cell
types is very similar to that of PiipiO1: upregulated over
100-fold in zoospores, peaking at over 1600-fold in germi-
nating cysts and then decreasing to 600-fold in germinating
cysts with appressoria (Fig. 1). Unlike PiipiO1, the relative
upregulation of Pihmp1 during infection is lower than that in
germinating cysts with appressoria, suggesting that its
mRNA is transcribed during cyst germination and appres-
sorial development, to be translated for the early biotrophic
stage of interaction. Repeated amplifications, on indepen-
dent occasions with biologically replicated cDNA samples,
resulted in similar expression profiles.
To test whether the truncated paralogue of Pihmp1 was
expressed, primers 1B18F and 1B18R were used in real-
time RT-PCR across the life cycle stages of P. infestans. A
product of the expected size was amplified from genomic
DNA but not from cDNA of any P. infestans life cycle stage
(not shown).
2274 A. O. Avrova et al.

trans-membrane (TM) helices between residues 605 and

A Pihmp1
2000 722, potentially anchoring the protein to the plasma mem-
brane (Fig. 2). Both N- and C-termini of Pihmp1 were
Relative expression

1600 predicted to be located outside the plasma membrane,

with only 10 amino acids in two loops between TM helices
1200
predicted to extend into the cytoplasm. TMHMM posterior
800 probabilities for the sequence were: 0.98 for TM helix 1,
0.90 for TM helix 2, 0.90 for TM helix 3 and 1.00 for TM
400 helix 4.
BLASTP and TBLASTN searches, using default settings, of
0
M S Z C A B12 B24 B33 B48 B56 B72 NCBI GenBank (non-redundant database) yielded only
weak similarity (5e-7 to 1e-13) to a bacterial and fungal
B domain in the N-terminal half of the protein, between
PiipiO residues 293 and 505. In BLASTP-returned sequences from
350
Relative expression

300 Xanthomonas axonopodis pv. citri (NP_644432.1) and

250
200
150
100
50
0 containing three lysine/serine (K/S)-rich repeats of 32–39
M S Z C A B12 B24 B33 B48 B56 B72
Fig. 1. Pihmp1 is expressed in pre-infection stages and throughout
infection of potato. Real-time RT-PCR expression profile of (A)
Pihmp1 and (B) PiipiO1 in pre-infection stages (sporangia – S,
zoospores – Z, germinating cysts – C, germinating cysts with
appressoria – A) and 12 (B12), 24 (B24), 33 (B33), 48 (B48), 56
(B56) and 72 (B72) h post inoculation of potato cv. Bintje with a
compatible race (88069) of P. infestans, relative to their expression
in vegetative non-sporulating mycelium (M). Error bars represent
confidence intervals calculated using three technical replicates for
each sample within the RT-PCR assay.
Pihmp1 is a putative glycosylated membrane protein
The 3072 bp ORF of Pihmp1, which lacks introns,
encodes a predicted 1023-amino-acid, 108.3 kDa protein.
A signal peptide for secretion was predicted at the
N-terminus by SignalP 3.0 (neural network mean 0.929,
hidden Markov model 0.998), with the molecular weight of
the mature protein estimated at 105.5 kDa. The CBS
Prediction Servers (TMHMM v. 2.0) and SOSUI program
(v. 1.11) predicted Pihmp1 to encode a protein with four
Bradyrhizobium sp. (YP_001208179.1), this domain
yielded modest support for a jacalin-like (mannose/
galactose binding) lectin domain. Weak support, below the
thresholds used to construct the Pfam model, for a jacalin-
like lectin domain was identified in Pihmp1 using Pfam 22.0
(E-value 0.0014) between residues 374 and 505. A domain
residues was identified manually between residues 130
and 260 (Fig. 2B). The final two K/S repeats contain four
copies of a DxSxS metal ion-dependent adhesion site
(MIDAS) motif (Tozer et al., 1996). An additional three
serine/glycine (S/G)-rich repeats of 15–17 residues follow
the K/S-rich repeats. CBS Prediction Servers (NetNGlyc
1.0) predicted N-glycosylation at positions 390, 818 and
966, and potential O-glycosylation (mucin type; NetOGlyc
3.1) at positions 93–95, 134, 135, 139, 142, 145 and 226
(Fig. 2B). A mucin domain was also predicted by Pfam in
this region of Pihmp1 but with poor support (E-value 0.55),
below the thresholds used to construct the Pfam model.
Seven cysteine residues are present in the 99-amino-acid
region between residue 505 and the first TM helix of the
predicted mature Pihmp1 protein, and a further three cys-
teines are present in the C-terminal portion of the protein
after the TM helices (Fig. 2B). These protein sequence
features suggest that Pihmp1 is a membrane-anchored,
predominantly extracellular, glycosylated protein that may
act at the interface between P. infestans and host cells
during infection.

Fig. 2. Organization, and multiple alignment of homologues, of the predicted Pihmp1 protein.
A. Sequences in multiple alignment are Pihmp1 from P. infestans, and closest homologues from P. ramorum (Prhmp1), P. sojae (Pshmp1) and
H. parasitica (Hphmp1). Protein sequences were aligned using the CLUSTALW accessory program in the BioEdit Sequence Alignment Editor.
Identical residues across three or more homologues are shaded black, and similar residues across three or more homologues are shaded
grey. The signal peptide is marked by solid overhead bracket, conserved cysteine residues are marked with an asterisk (*) and the
trans-membrane helices are marked by solid black lines.
B. Schematic representation of Pihmp1 showing location of features (not to scale): signal peptide (SP; shaded blue), conserved EER motif
(shaded dark blue), predicted O-glycosylation (vertical red bars), 3¥ K/S-rich repeats (3¥ grey shaded squares), 4¥ DxSxS MIDAS motif
(white squares within K/S repeats), 3¥ S/G-rich repeats (shaded pink), predicted N-glycosylation (vertical black bars), conserved cysteine
residues (vertical blue bars) and trans-membrane helices (TM; shaded green).

© 2008 Scottish Crop Research Institute

Journal compilation © 2008 Blackwell Publishing Ltd, Cellular Microbiology, 10, 2271–2284
 Phytophthora infestans haustorial membrane protein 2275

© 2008 Scottish Crop Research Institute

Journal compilation © 2008 Blackwell Publishing Ltd, Cellular Microbiology, 10, 2271–2284
2276 A. O. Avrova et al.

Pihmp1 has homologues in other oomycetes dsRNA did not exhibit sequence identity over 20 bp to any
other predicted gene or genomic region to initiate off
A TBLASTN search, individually, of the oomycete genome
target gene silencing. The only other sequence exhibiting
databases held within the Virginia Microbial Database
sequence similarity was the truncated and non-expressed
(VMD; Tripathy et al., 2006) with Pihmp1 identified
copy of Pihmp1. In transient gene silencing experiments
homologues in P. sojae (Assembly V3.0, Scaffold
for Pihmp1, lines exhibiting a phenotype after treatment
Super_7, 1918598–1921645 bp), P. ramorum (Assembly
with Pihmp1 dsRNA also showed reduced Pihmp1 mRNA
V1.0, Scaffold 1, 433071–436238 bp) and H. parasitica
abundance, ranging from 8% to 47% compared with
(Assembly V6.0, Contig 6, 566653–569616 bp) with
the non-homologous dsRNA-treated controls (Fig. 3D).
E-values ranging from 9 ¥ 10-88 to 10-140 (Fig. 2A). Homo-
Therefore, successful transient silencing of Pihmp1 was
logues identified by TBLASTN were verified by BLASTN
achieved using dsRNA treatment of protoplasts, allowing
search of the same oomycete genomes with the Pihmp1
the resultant phenotypes to be studied.
DNA sequence to demonstrate that the genomic locations
Both the Pihmp1 dsRNA-treated (silenced) and the
of closest TBLASTN and BLASTN homologues of Pihmp1
non-homologous dsRNA-treated (control) lines grew nor-
were the same. These analyses yielded a single-copy
mally on rye agar medium and produced similar numbers
Pihmp1 homologue from each oomycete genome
of sporangia and zoospores that encysted normally (data
examined. The Pihmp1 homologue in P. ramorum is the
not shown). In control lines, by 24 hpi, cysts and sporan-
most closely related to Pihmp1, as determined by protein
gia had germinated on the leaf surface, formed appres-
distance (Table S2). A homologous sequence to Pihmp1
soria and begun host penetration, forming infection
also exists in the genome of Phytophthora capsici, with
vesicles and haustoria inside host cells (Fig. 3A). Germi-
E-values of 2 ¥ 10-37-10-105, but this genome sequence
nated cysts and sporangia from Pihmp1 dsRNA-treated
has not been assembled to date and so no further
lines developed appressoria-like structures, but infection
conclusions can be made about Pihmp1 homologues in
vesicles and haustoria for in planta infection were not
this species. Although homologues of Pihmp1 in other
observed (Fig. 3B), suggesting a role for Pihmp1 in cell
oomycetes exhibited similarity across the entire length of
membrane, and perhaps cell wall, differentiation and func-
the protein, the greatest divergence in the amino acid
tion during infection. Hyphae from germinated cysts and
sequence was observed in the N-terminal 504 amino
sporangia of Pihmp1-silenced lines occasionally entered
acids of Pihmp1 (Fig. 2A, Table S2). In comparison, these
host leaf tissue via stomata, but infection did not develop
levels of localized protein divergence were not observed
to the levels of control non-homologous dsRNA-treated
for cellular proteins actinA (actA), translation elongation
lines (Fig. 3C); sparse intercellular hyphae and no visible
factor 1a (tef1a) or the trans-membrane protein cellulose
lesion were observed in these instances. Germ tubes of
synthase 3 (cesA3) (Table S2). All 10 cysteine residues in
cysts from Pihmp1-silenced lines also formed atypical
the C-terminal half of Pihmp1 (Fig. 2; seven preceding
distorted hyphae with numerous bulges and swellings that
and three following the TM helices) are conserved in all
grew across the leaf surface by 48 hpi. This latter pheno-
four homologues, suggesting localized conservation of
type was also occasionally observed in wild-type non-
structure in these proteins across different oomycetes.
silenced lines where penetration was unsuccessful or
Additional protein sequences from P. sojae, P. ramorum
delayed and, although this phenotype was significantly
and H. parasitica showed similarity to the P. infestans
more pronounced in Pihmp1-silenced lines, it may be a
Pihmp1 protein. However, these related sequences did
nutrient starvation response by P. infestans, rather than a
not align across the full length of the protein, with little
direct phenotype of Pihmp1 silencing.
or no similarity observed to the (approximately) 300
The observed microscopic phenotype associated with
N-terminal residues, and with some possessing up to
reduced Pihmp1 expression was accompanied by ap-
two predicted mannose/galactose-binding jacalin lectin
parent loss of pathogenicity on potato leaves (Fig. 3C)
domains in the N-terminal half of the protein sequence.
due to inability of Pihmp1-silenced lines to spread within
leaf tissue and form necrotic, sporulating lesions. In our
experiments, Pihmp1 dsRNA-treated lines expressing
Silencing of Pihmp1 leads to loss of pathogenicity
Pihmp1 mRNA up to 47 ⫾ 6% of wild-type Pihmp1 mRNA
Transient gene silencing, triggered by double-stranded levels were non-pathogenic. Pihmp1 dsRNA-treated lines
RNA (dsRNA; Whisson et al., 2005), was used to deter- expressing Pihmp1 mRNA greater than 63 ⫾ 8% pro-
mine the role of Pihmp1 during P. infestans infection of duced appressoria and infection vesicles, colonized leaf
potato. DsRNA for the green fluorescent protein gene tissue by forming typical intercellular hyphae with hausto-
(gfp) or another non-homologous dsRNA not present in ria, leading to lesions indistinguishable in size to the
the wild-type P. infestans isolates 88069 or CY29 were control dsRNA-treated lines. These results clearly show
used as controls. The sequence of the 200 bp Pihmp1 that silencing of Pihmp1 affected P. infestans pathogeni-

© 2008 Scottish Crop Research Institute

Journal compilation © 2008 Blackwell Publishing Ltd, Cellular Microbiology, 10, 2271–2284
 Phytophthora infestans haustorial membrane protein 2277

A B a

iv h

C D 1.8
1.6

Relative expression
1.4

1.2
1

0.8
0.6
0.4
0.2
0
C1 C2 C3 C4 C5 1 2 3 4 5

Independent lines

Fig. 3. Transient silencing of Pihmp1 leads to loss of pathogenicity.

A. Control line treated with non-homologous dsRNA. Host penetration has occurred by 24 hpi with infection vesicles (iv) and haustoria (h)
seen within invaded host cells.
B. Pihmp1 transient silenced line showing germinated cyst (c) with two appressoria-like structures (a), but failed host penetration. Scale bar in
(A) and (B) represents 10 mm.
C. Macroscopic disease symptoms at 7 days post inoculation on potato leaves of Pihmp1 transient silenced lines (bottom) compared with
control lines treated with non-homologous dsRNA (top).
D. Levels of transient Pihmp1 silencing observed in experiment shown in (C). Error bars represent confidence intervals calculated using three
technical replicates for each sample within the RT-PCR assay.

city, with silenced lines unable to form intracellular primary
infection vesicles or haustoria and establish infection.
Five independent transient silencing experiments with
P. infestans strains CY29 (three experiments) or 88069
(two experiments), on different dates and with indepen-
dent preparations of in vitro synthesized dsRNA (controls
and Pihmp1), yielded the same loss of pathogenicity phe-
notype for Pihmp1. From these experiments, a total of 100
Pihmp1 dsRNA-treated lines were generated, of which 44
exhibited the loss of pathogenicity phenotype. Seventy-
two control lines, derived from protoplasts treated with
non-homologous dsRNA, were also used, all of which
exhibited wild-type levels of pathogenicity. The frequency
of transient silencing observed among Pihmp1 dsRNA-
treated lines is similar to that observed for other genes
tested by this gene silencing technique (Whisson et al.,
2005; Grenville-Briggs et al., 2008; Walker et al., 2008).
Pihmp1 is localized to the P. infestans haustorial
membrane during infection of potato
The Pihmp1 ORF, including 694 bp of promoter
sequence, was cloned for translational fusion with GFP
in P. infestans stable transformants. Three out of six
transformants transcribed gfp in germinated cysts, in-
dicating the 694 bp region to be sufficient to direct normal
Pihmp1 expression. However, GFP fluorescence was not
observed in germinated cysts either in vitro or in planta
following the inoculation of detached potato leaves. In
addition, no GFP fluorescence was observed in inter-

© 2008 Scottish Crop Research Institute

Journal compilation © 2008 Blackwell Publishing Ltd, Cellular Microbiology, 10, 2271–2284
2278 A. O. Avrova et al.
cellular hyphae or haustoria in planta. This might be
explained by previously reported low fluorescence yield of
GFP in the apoplast (Zheng et al., 2004) and may confirm
bioinformatic predictions suggesting the C-terminus of
Pihmp1 to be located outside the plasma membrane.
Pihmp1 with its native promoter was transferred
from p2c5-GFP to pPLmRFP to yield p2c5-mRFP, a
Pihmp1::mRFP translational fusion. Co-transformation of
P. infestans strain 88069 with p2c5::mRFP and pVW2
(constitutive eGFP expression; van West et al., 1999)
yielded 17 transformants that exhibited normal in vitro and
in planta development. Expression of the Pihmp1::mRFP
fusion in germinated cysts was shown by RT-PCR of
the entire mRFP transcript for nine transformants
(2c5mRFP1, 2, 5, 10, 11, 12, 14, 17 and 19) using mRF-
PclaF and mRFPsacR primers. Of these transformants
expressing the Pihmp1::mRFP fusion, six exhibited
detectable mRFP fluorescence at varying intensities
(2c5mRFP1, 2, 10, 11, 12 and 17), and four transformants
(2c5mRFP10, 11, 12 and 17) also exhibited cytoplasmic
eGFP fluorescence. To confirm that the Pihmp1::mRFP
fusion was expressed from the native promoter in trans-
formants exhibiting mRFP fluorescence, PCR amplifying
the region from the Pihmp1 promoter to mRFP stop codon
was performed with the primers SpeI2C5a and mRFP-
sacR, using transformant genomic DNA and cDNA from
germinated cysts, separately, as templates. The product
of the expected size was amplified only from genomic
Journal compilation © 2008 Blackwell Publishing Ltd, Cellular Microbiology, 10, 2271–2284
Fig. 4. Localization of Pihmp1 to haustoria by
translational fusion to the monomeric red
fluorescent protein (mRFP) and under control
of the Pihmp1 promoter. Enhanced green
fluorescent protein (eGFP) was expressed
constitutively to define P. infestans cytoplasm.
A. Localization of Pihmp1 in germinating
sporangia and cysts with appressoria on the
surface of a potato leaf. Aggregations of
tagged Pihmp1 can be observed in the cyst or
sporangia germ tube, concentrating in the
swollen appressorium (arrowhead) at the
germ tube tip. Arrows and letters indicate
example cyst (c), sporangia (s), germ tube
(g) and appressorium (a) expressing
Pihmp1::mRFP.
B. Red channel of the image shown in (A).
C. Pihmp1 localizes to biotrophic digit-like
intracellular haustoria (arrow) during
P. infestans infection of potato. Additional
small ‘plaques’ of tagged Pihmp1 can also be
observed in the intercellular hyphae
(arrowhead) and may represent sites of either
early or aborted haustorium formation.
D. Red channel of the image shown in (C).
Scale bars represent 50 mm.
DNA of transformants, indicating that the mRNA encoding
the fusion protein was being transcribed under control of
the Pihmp1 promoter.
Confocal microscopy of these transformants showed
that the Pihmp1::mRFP fusion protein was present, in
what appeared to be vesicles, inside sporangia and cysts
germinated both in vitro and on the leaf surface. The
Pihmp1::mRFP-containing ‘vesicles’ moved along the
germ tube, concentrating in the appressorium in contact
with the leaf surface (Fig. 4A and B). After host penetra-
tion and development of infection, mRFP fluorescence
was observed surrounding haustoria, outlining the haus-
torial membrane and concentrated around the haustorial
base (Fig. 4C and D). Plaques of Pihmp1::mRFP were
also observed on the surface of intercellular hyphae, and
these may indicate sites for initiation of haustorial growth.
These observations are consistent with Pihmp1 transla-
tion in the pre-infection stages of germinating cysts and
appressoria for its destined site of localization in the
biotrophic infection structures of P. infestans.
To investigate sites of haustorial growth, development
of intercellular hyphal tips of transformant 2c5mRFP11
were followed by confocal microscopy over 12 h. This
determined that the previously identified plaques of
Pihmp1::mRFP were indeed the sites for initiation of haus-
torial growth (Fig. 5). However, not all Pihmp1 plaques
subsequently developed haustoria. Where haustoria did
develop, initial penetration of host cells was rapid, occur-
© 2008 Scottish Crop Research Institute
 Phytophthora infestans haustorial membrane protein 2279

Fig. 5. Development of P. infestans intracellular haustoria in potato leaf tissue, showing that not all potential sites of haustorial growth develop
into mature haustoria. T = 0. Two adjacent intercellular hyphal tips showing at least six Pihmp1 plaques as sites of potential haustorium
formation. T = 3. Plant cells are invaded by haustoria (arrowheads) within 3 h from Pihmp1 accumulation sites on the intercellular hyphae.
T = 7. After 7 h, haustoria continue to grow into the invaded cells and begin to exhibit a characteristic hooked form. After 9 h (T = 9) and up to
12 h (T = 12) the haustoria continue to grow and hook, accumulating more Pihmp1 at the surface. Note that several potential sites of
haustorium formation failed to develop further (long arrows indicate one of these). Scale bar represents 10 mm.

ring within 3 h (Fig. 5), with subsequent elongation and
development of the characteristic hooked shape of the
haustorium occurring for up to 12 h (Fig. 5). The distribu-
tion of Pihmp1 also changed across this time-course,
appearing to become increasingly concentrated in
patches on the haustorial surface during elongation and
maturation.
Discussion
We have provided the first evidence that the surface of
oomycete haustoria possesses proteins specific to these
biotrophic structures, and that formation of biotrophic
structures (infection vesicles and haustoria) is essential to
successful host colonization by P. infestans. We used
SSH to enrich for cDNAs of transcripts upregulated in
P. infestans germinating cysts with appressoria, struc-
tures formed just prior to infection of potato, with the
hypothesis that proteins required for infection would be
translated just prior to host penetration. This strategy has
proven highly successful for identifying genes encoding
effectors probably required for successful establishment
of a compatible interaction (Liu et al., 2005; Randall et al.,
2005; Judelson et al., 2008). While Pihmp1 was most
highly expressed in germinated cysts and appressoria,
the encoded protein was translated in these pre-infection
structures and transported towards the penetration site to
be later incorporated into the plasma membrane of either
the infection vesicle or haustoria.
Pihmp1 is predicted to encode a putative outer mem-
brane protein with four TM helices in the middle of the
peptide sequence, anchoring the protein to the plasma
membrane (Fig. 2). Both N- and C-termini of Pihmp1 are
predicted to be located outside the membrane. Several
domains and features were identified or predicted in
the sequence of Pihmp1, being (from N- to C-terminus):
signal peptide, mucin-like O-glycosylation, K/S-rich
repeats, DxSxS MIDAS-like motifs, S/G-rich repeats,
N-glycosylation, cysteine-rich domain, TM helices and a
C-terminal half without discernible domains of known
function. MIDAS motifs complex metal ions and are found
in animal, and animal pathogen, membrane-localized pro-
teins (such as integrins), where they are involved in medi-
ating cell attachment through interaction with proteins
such as collagen and fibronectin (Tozer et al., 1996;
Akhouri et al., 2004; Kottom et al., 2008).
Pihmp1 is represented by a single full-length functional
copy in all four oomycete genomes examined. Additional
similar proteins are also present in all genomes and, as
such, Pihmp1 appears to form part of a group of similar
proteins encoded by dissimilar genes. All of these similar
proteins exhibit the greatest similarity at the C-terminal
end, typically showing little or no similarity within the
N-terminal 300 amino acids.
The precise role of Pihmp1 in the cell biology of
P. infestans remains to be determined. One possible func-
tion for Pihmp1 may be as a receptor/sensor, but this
appears unlikely as only 10 amino acids are located within
the cell cytoplasm, and no putative phosphorylation sites
were predicted within these short regions to initiate down-
stream signalling. This line of evidence suggests that
Pihmp1 is more likely to be a structural protein. One model
for the structure of Pihmp1 may involve the anchoring of
the protein to the plasma membrane, with the cysteine-rich
domain potentially spanning the cell wall. The K/S-rich,
DxSxS MIDAS-like motifs, S/G-rich, and O-glycosylated
amino acids would thus be exposed to the EHM, and also
extracellular host surveillance and defences. This latter
consideration may in part explain the N-terminal diver-
gence in the homologues of Pihmp1 between P. infestans,
P. sojae, P. ramorum and H. parasitica; host species selec-
tion may have acted to diversify this region. Although a
Pfam search failed to recognize any other well-supported
known functional domains in Pihmp1, this protein might still
contain domains specific to oomycetes, the function(s) of
which remains to be elucidated.
Consistent with a suggested role as a structural protein,
transient gene silencing of Pihmp1 in P. infestans led to an

© 2008 Scottish Crop Research Institute

Journal compilation © 2008 Blackwell Publishing Ltd, Cellular Microbiology, 10, 2271–2284
2280 A. O. Avrova et al.
inability to form biotrophic structures after appressorium
formation. Pihmp1 may stabilize the P. infestans cell
membrane, and perhaps cell wall, during penetration. The
observed microscopic phenotype was accompanied by an
apparent loss of pathogenicity, even though Pihmp1-
silenced lines were able to enter the leaf via stomata but
were unable to spread further, indicating that Pihmp1 is
required for successful establishment of infection in addi-
tion to penetration.
Haustoria of fungal and oomycete plant pathogens
differ in at least one major respect: differentiation of a
neckband close to the intercellular hyphae. The neckband
functions to seal the EHM from the apoplast such that the
matrix becomes a unique and isolated compartment
(Szabo and Bushnell, 2001). Electron microscopy studies
of the haustorium-forming oomycetes Albugo candida
and H. parasitica (syn. Peronospora parasitica) revealed
some differences in electron-dense material between
intercellular hyphae and the intracellular haustorial cell
wall and membrane (Soylu, 2004; Soylu et al., 2004),
although these differences were not as pronounced as for
fungal haustoria. The localization of Pihmp1 has revealed
the first evidence that although the P. infestans hausto-
rium membrane and wall are apparently undifferentiated
(Hohl and Suter, 1976), it does have a different protein
complement to intercellular hyphae. This finding is con-
sistent with haustoria of rust fungi where it has been
shown that the uptake of sugars and amino acids occurs
only in the haustoria through membrane-localized
transporters (Hahn et al., 1997; Voegele et al., 2001). As
stated, fungal haustoria are considered to be sealed from
the host apoplast by a neckband. However, recent results
with mRFP-tagged, translocation-deficient, Avr3a aviru-
lence protein showed that some leakage of proteins from
the P. infestans EHM to the apoplast could occur
(Whisson et al., 2007), indicating that it does not have a
sealed EHM. Furthermore, those results highlighted that
the haustorium is likely to be a major site of effector
protein secretion during infection.
Also highlighted in this study was the rapidity with which
haustoria are formed in plant cells and the time taken for
a haustorium to fully mature. From the initial formation of
Pihmp1::mRFP plaques on intercellular hyphae, haustoria
are nearly fully formed within 3 h, but continue maturing
for up to 12 h. The time-courses also suggested that not
all available sites for haustoria are utilized, or that the
earliest events in initiation of haustorium formation may
be halted if a site is sensed as unsuitable. This observa-
tion is in contrast with those of Hohl and Suter (1976)
where haustorial development was thought to only be
halted by host wall appositions after host cell penetration.
Our results imply that sensing of suitable sites for haus-
torium development may occur earlier. The variation in
Pihmp1 localization during haustorial growth and matura-
Journal compilation © 2008 Blackwell Publishing Ltd, Cellular Microbiology, 10, 2271–2284
tion, becoming increasingly aggregated by 12 h, showed
that the structures of the haustorial membrane and wall
are likely to be in constant flux over the course of its
development and function, possibly in response to host
cues.
The hypersensitive response, triggered by recogni-
tion of specific pathogen avirulence proteins by host
resistance proteins, targets the biotrophic stages of
P. infestans infection, with host programmed cell death
leading to control of infection. The inability of Pihmp1-
silenced lines to form haustoria, and thus infect potato,
also highlights the importance of haustorium formation to
the success of P. infestans infection. These observations
indicate that haustoria, and the essential functions they
perform, provide ideal targets for late-blight control.
Experimental procedures
Growth of P. infestans, potato plants and
plant inoculation
Growth of potato plants, P. infestans isolates CY29 and 88069,
and plant inoculation were carried out as described in Grenville-
Briggs et al. (2005). Life cycle stages of P. infestans (axenic
cultured mycelium, sporangia, zoospores, germinated cysts with
or without appressoria) were also prepared as described by
Grenville-Briggs et al. (2005).
RNA isolation, cDNA synthesis, SSH, real-time RT-PCR
expression analysis and mRFP RT-PCR
Total RNA was extracted using the QIAGEN RNeasy Plant Mini
Kit using the protocol supplied by the manufacturer, with the
addition of on-column DNase digestion. Synthesis of first-strand
cDNA was carried out using the First Strand cDNA Synthesis kit
(GE Healthcare, UK) using oligo dT priming, following the proto-
col supplied by the manufacturer.
Suppression subtractive hybridization using the PCR-Select
cDNA subtraction kit (Clontech) was performed to generate a
cDNA library enriched for sequences specifically expressed
during appressorium formation. SSH conditions were as
described by Grenville-Briggs et al. (2005).
For gene expression analysis, SYBR green real-time RT-PCR
assays were carried out as described in Avrova et al. (2003). The
P. infestans actinA (PiactA) gene was used as a constitutively
expressed endogenous control and relative expression of
Pihmp1 was normalized against expression levels in non-
sporulating mycelium (assigned a relative expression value of
1.0) as described in Grenville-Briggs et al. (2008) and Judelson
et al. (2008).
Pihmp1::mRFP fusion gene expression was detected by
RT-PCR using the following conditions: each 20 ml of PCR con-
tained 0.5 U Taq DNA polymerase (Promega, UK), 2 ml of 10¥
reaction buffer, 1.5 mM MgCl2, 15 mM deoxynucleotide triphos-
phates (Promega), 30 mM mRFPClaF and mRFPSacR primers
(Table S1), and 10 ng of cDNA. Thermocycling conditions were
as follows: 94°C for 5 min, followed by 40 cycles of 94°C for 30 s,
60°C for 30 s and 72°C for 1 min. A final extension step of 72°C
for 10 min was included.
© 2008 Scottish Crop Research Institute
 Phytophthora infestans haustorial membrane protein 2281
P. infestans DNA manipulations
Transfer of the P. infestans BAC library to nylon membranes,
colony and Southern hybridization was as described in Whisson
et al. (2001). BAC clone plasmid preparations were as described
in Whisson et al. (2001; 2005). For fingerprinting, BAC clones
were restriction digested with HindIII restriction endonuclease,
and fragments were separated by 1% agarose gel electrophore-
sis and stained with SYBR Gold nucleic acid stain (Invitrogen,
UK).
Phytophthora infestans genomic DNA was extracted from
freeze dried mycelium using the QIAGEN DNeasy Plant Mini Kit,
using the protocol supplied by the manufacturer. Southern
transfer of either restriction endonuclease (HindIII)-digested
P. infestans genomic DNA or BAC DNA from agarose gels to
Hybond N+ filters (GE Healthcare) was by standard methods
(Sambrook et al., 1989).
Preparation of single-sequence hybridization probes used
25 ng of probe DNA, 50 mCi of [a-32P]-dCTP and a single reaction
of the High Prime labelling kit (Roche).
DNA from BAC clones 1B18, 5J1 and 15J23 was partially
digested with Sau3AI (Promega) and subcloned into vector
pGEM-3Z (digested with BamHI restriction endonuclease and
alkaline phosphatase; Promega, UK) as described in Bell et al.
(2002). For each BAC clone, 768 recombinant Escherichia coli
strain DH10B (Invitrogen) subclones were selected, grown at
37°C in 384-well microtitre plates in freezing medium (Whisson
et al., 2001) with ampicillin for 24 h, and then stored at -70°C.
BAC subclone plasmids were prepared using the QIAGEN
Plasmid Miniprep kit and sequenced in both directions with SP6
and T7 primers (Table S1) using the Perkin Elmer ABI PRISM
BigDye Terminator 3.1 cycle sequencing kit, manufacturer-
recommended thermal cycling conditions, and ABI Model 377
DNA sequencer. The 5′ end of the Pihmp1 sequence was
obtained by direct sequencing from BAC clones 1B18, 5J1 and
15J23 using primers 1B18R, 5J1R, 5J1R1, 5J1R2, 15J23R3,
2C5prR2 and 2C5prF2 (Table S1).
DNA and protein sequence analyses
DNA and protein sequence similarity searches were made with
BLASTN, TBLASTN, TBLASTX and BLASTP algorithms, using default
search parameters, at NCBI GenBank (http://www.ncbi.nlm.nih.
gov/blast/Blast.cgi; non-redundant nucleotide and protein
databases, ESTs including human and mouse sequences),
Phytophthora Functional Genomics Database (PFGD; ESTs from
P. infestans and P. sojae; http://www.pfgd.org/), Virginia Microbial
Database (individual genome databases for P. sojae, P. ramorum
and P. capsici and H. parasitica; http://phytophthora.vbi.vt.edu/)
and P. infestans genome database (http://www.broad.mit.edu/
annotation/genome/phytophthora_infestans). BLASTN algorithm
parameters were: expect threshold = 10, word size = 28, linear
gap cost, match and mismatch scores of 1 and -2 respectively.
BLASTP and TBLASTN algorithm parameters were: expect
threshold = 10, word size = 3, BLOSUM62 matrix, gap costs of
existence = 11 and extension = 1, and conditional compositional
score matrix adjustment. Algorithm parameters for TBLASTX were
as for BLASTP, but without conditional compositional score matrix
adjustment. ORF finder (http://www.ncbi.nlm.nih.gov/gorf/gorf.
html) and DNA Tools (http://biology.semo.edu/cgi-bin/dnatools.pl)
were used for predicting the ORFs in the sequences obtained.
© 2008 Scottish Crop Research Institute
Journal compilation © 2008 Blackwell Publishing Ltd, Cellular Microbiology, 10, 2271–2284
Additional internet sites used for the structural analysis of the
predicted proteins are (all using default parameters): CBS Pre-
diction Servers (TMHMM v. 2.0; NetOGlyc v. 3.1; NetNGlyc v. 1.0;
SignalP v. 3.0; http://www.cbs.dtu.dk/services/), SOSUI program v.
1.11 (http://bpnuap.nagoya-u.ac.jp/sosui/) and Pfam v. 22.0
(http://pfam.sanger.ac.uk/). Protein alignments were performed
using the CLUSTALW v. 1.4 program, using default parameters,
within the BioEdit Sequence Alignment Editor v. 7.0.5.3. Protein
distance calculations between homologues of Pihmp1 were per-
formed on aligned protein sequences using default parameters
(Jones Taylor Thornton matrix; Jones et al., 1992) for the
PROTDIST v. 3.5c accessory program in the BioEdit Sequence
Alignment Editor.
Transient RNAi and trypan blue staining of
Pihmp1-silenced P. infestans in planta
Transient RNAi was performed as described in Whisson
et al. (2005). Primer T7SSHac2C5Fs in combination with
SSHac2C5Rs, and SSHac2C5Fs in combination with
T7SSHac2C5Rs (Table S1) were used to amplify an approxi-
mately 200 bp section of Pihmp1 for in vitro synthesis of dsRNA
using the Megascript RNAi kit (Ambion, UK) following the proto-
col supplied by the manufacturer.
After treatment with dsRNA (Pihmp1 or controls), single colo-
nies from regenerated (typically 5–6 days after treatment) proto-
plasts were transferred to rye agar medium. Thirteen days after
the exposure of protoplasts to dsRNA, these colonies were
flooded with cold sterile distilled water to induce zoospore
formation. After 3 h at 4°C, the suspension was collected and
used for leaf inoculations. Leaf samples were taken at 48 hpi
(15 days after exposure to dsRNA) for RNA extraction. Subse-
quently real-time RT-PCR was used to determine whether the
Pihmp1 dsRNA-treated lines were silenced. Given the small
amounts of biological material available from each dsRNA-
treated line, the individual lines formed the biological replicates
for the experiment as a whole, while three technical replicates
were performed for each real-time RT-PCR assay.
Leaf samples were taken at 16, 24 and 48 hpi for trypan blue
staining (Koch and Slusarenko, 1990) and light microscopy. Both
the infection site and developing lesion margin (where infection
progressed) were examined. Digital photomicrographs were cap-
tured on an Olympus BH2 light microscope using Olympus DP70
or C5050 digital cameras with automatic exposure settings.
GFP and mRFP tagging of Pihmp1, and P. infestans
transformation
The Pihmp1 ORF preceded by 694 bp was amplified from
P. infestans BAC clone 5J1 using primers SpeI2C5 and
NotI2C5R. Each 50 ml of PCR contained 1 U Taq DNA poly-
merase (Promega, UK), 5 ml of 10¥ reaction buffer, 1.5 mM
MgCl2, 15 mM deoxynucleotide triphosphates (Promega), 30 mM
forward and reverse primers (Table S1), and 10 ng of BAC DNA.
Thermocycling conditions were as follows: 94°C for 5 min, fol-
lowed by 40 cycles of 94°C for 30 s, 62°C for 30 s and 72°C for
4 min. A final extension step of 72°C for 10 min was included. The
Pihmp1 PCR product was digested with SpeI and NotI restriction
endonucleases, and introduced into the pSam SpeI and NotI
cloning sites in the sense orientation to yield construct p2C5-
2282 A. O. Avrova et al.
GFP. pSam is a promoterless GFP construct also containing the
neomycin phosphotransferase gene (nptII) under control of the
Bremia lactucae hsp70 promoter for expression in oomycetes.
Pihmp1 with its native promoter was transferred from p2C5-GFP
into pPL-mRFP to yield p2c5-mRFP. pPL-mRFP was constructed
by excision of the B. lactucae Ham34 promoter from pTor (Blanco
and Judelson, 2005) and ligation of the mRFP gene into the KpnI
and SacII sites, with a linker for KpnI, NotI and ClaI cloning sites
at the 5′ end of the gene to yield a promoterless mRFP vector for
oomycete expression. Pihmp1 was excised from p2c5-GFP with
SpeI and SacII restriction endonucleases, blunt-ended with
Klenow polymerase (New England Biolabs), cut with NotI restric-
tion endonuclease, purified and ligated into the pPL-mRFP
vector. All fusions were sequenced in both directions to verify that
the fusions were in frame and without introduced cloning errors.
Oomycete expression vector pVW2 for constitutive eGFP
expression has been described elsewhere (van West et al.,
1999).
Stable transformation of P. infestans was achieved using a
modified PEG-CaCl2-Lipofectin protocol described by Judelson
et al. (1991). An updated version of this protocol can be found at
http://138.23.152.128/protocols/protocols.html. For use here,
the protocol was modified by substituting Novozym 234 with
a mixture of 5 mg ml-1 lysing enzymes (from Trichoderma
harzianum; Sigma L1412) and 2 mg ml-1 cellulase (from Tricho-
derma reesei; Sigma C8546) to release protoplasts from germi-
nating sporangia. For each transformation, 50 mg of plasmid DNA
was mixed with Lipofectin reagent (Invitrogen, UK) in the propor-
tions described in the protocol. Protoplasts were regenerated for
48–72 h in pea broth (Whisson et al., 2005) containing 20 g l-1
sucrose and 1 M mannitol, prior to selection of transformants on
rye agar amended with 5 mg ml-1 geneticin. Transformants were
then transferred to, and maintained on, rye agar amended with
20 mg ml-1 geneticin.
Samples of mycelium, and cysts germinated in vitro for 16 h at
10°C, were taken for DNA or RNA extraction and PCR or RT-PCR
analysis respectively.
Confocal imaging of fluorescent proteins
Zoospores (5 ¥ 104 ml-1) of stable transformants were inoculated
in 10 ml droplets onto detached potato leaflets. Inoculated leaflets
were incubated at 20°C for 24–72 h to allow cysts to germinate
on the leaf surface, form appressoria, penetrate host cells and
form intercellular hyphae with haustoria. For each transformant,
six leaflets were inoculated on two independent occasions.
All GFP (488 nm excitation; emission 500–530 nm) and mRFP
(561 nm excitation; emission 600–630 nm) imaging was con-
ducted on a Leica TCS-SP2 AOBS confocal microscope using
HC PL FLUOTAR 63X0.9, HCX APO L U-V-I 40X0.8 or HCX APO
L U-V-I 20X0.5 water dipping lenses. For all images shown in this
article, the 561 nm line from the ‘lime’ diode laser was set at
100% on the AOBS. The mRFP emissions were detected with the
photomultiplier tube set at either 670 V or 603 V.
Acknowledgements
This work was supported by the Scottish Government Rural and
Environment Research and Analysis Directorate (RERAD)
(A.O.A., P.C.B., V.Y., P.R.J.B. and S.C.W.), the Biotechnology and
Journal compilation © 2008 Blackwell Publishing Ltd, Cellular Microbiology, 10, 2271–2284
Biological Sciences Research Council (BBSRC) (L.J.G.-B.) and
The Royal Society (P.v.W.). The authors thank L. Pritchard at SCRI
for bioinformatics advice. Genetic modification of P. infestans was
carried out under plant health licence PH/11/2008.
References
Akhouri, R.R., Bhattacharyya, A., Pattnaik, P., Malhotra, P.,
and Sharma, A. (2004) Structural and functional dissection
of the adhesive domains of Plasmodium falciparum
thrombospondin-related anonymous protein (TRAP).
Biochem J 379: 815–822.
Avrova, A.O., Venter, E., Birch, P.R., and Whisson, S.C.
(2003) Profiling and quantifying differential gene transcrip-
tion in Phytophthora infestans prior to and during the early
stages of potato infection. Fungal Genet Biol 40: 4–14.
Avrova, A.O., Whisson, S.C., Venter, E., De Luca, S., Hein, I.,
and Birch, P.R.J. (2007) A novel, non-protein coding
infection-specific gene family is clustered throughout the
genome of Phytophthora infestans. Microbiology 153:
747–759.
Baldauf, S.L., Roger, A.J., Wenk-Siefert, I., and Doolittle,
W.F. (2000) A kingdom-level phylogeny of eukaryotes
based on combined protein data. Science 290: 972–977.
Bechinger, C., Giebel, K.-F., Schnell, M., Leidere, P.,
Deising, H.B., and Bastmeyer, M. (1999) Optical measure-
ments of invasive forces exerted by appressoria of a plant
pathogenic fungus. Science 285: 1.
Bell, K.S., Avrova, A.O., Holeva, M.C., Cardle, L., Morris, W.,
De Jong, W., et al. (2002) Sample sequencing of a
selected region of the genome of Erwinia carotovora
subsp. atroseptica reveals candidate phytopathogenicity
genes and allows comparison with Escherichia coli.
Microbiology 148: 1367–1378.
Beyer, K., Jimenez, S.J., Randall, T.A., Lam, S., Binder, A.,
Boller, T., et al. (2002) Characterization of Phytophthora
infestans genes regulated during the infection with potato.
Mol Plant Pathol 3: 473–486.
Birch, P.R.J., and Whisson, S.C. (2001) Phytophthora
infestans enters the genomics era. Mol Plant Pathol 2:
257–263.
Birch, P.R.J., Avrova, A.O., Armstrong, M., Venter, E., Taleb,
N., Gilroy, E., et al. (2003) The potato–Phytophthora
infestans interaction transcriptome. Can J Plant Pathol 25:
226–231.
Birch, P.R., Rehmany, A.P., Pritchard, L., Kamoun, S., and
Beynon, J.L. (2006) Trafficking arms: oomycete effectors
enter host plant cells. Trends Microbiol 14: 8–11.
Bittner-Eddy, P.D., Allen, R.L., Rehmany, A.P., Birch, P., and
Beynon, J.L. (2003) Use of suppression subtractive hybrid-
ization to identify downy mildew genes expressed during
infection of Arabidopsis thaliana. Mol Plant Pathol 4: 501–
507.
Blanco, F.A., and Judelson, H.S. (2005) A bZIP transcription
factor from Phytophthora interacts with a protein kinase
and is required for zoospore motility and plant infection.
Mol Microbiol 56: 638–648.
Burki, F., Shalchian-Tabrizi, K., Minge, M., Skjaeveland, A.,
Nikolaev, S.I., Jakobsen, K.S., and Pawlowski, J. (2007)
Phylogenomics reshuffles the eukaryotic supergroups.
PLoS ONE 2: e790. doi:10.1371/journal.pone.0000790.
© 2008 Scottish Crop Research Institute
 Phytophthora infestans haustorial membrane protein 2283
Catanzariti, A.M., Dodds, P.N., Lawrence, G.J., Ayliffe, M.A.,
and Ellis, J.G. (2006) Haustorially expressed secreted pro-
teins from flax rust are highly enriched for avirulence
elicitors. Plant Cell 18: 243–256.
Francis, S.A., Dewey, F.M., and Gurr, S.J. (1996) The role of
cutinase in germling development and infection by Ery-
siphe graminis f. sp. hordei. Physiol Mol Plant Pathol 49:
201–211.
Görnhardt, B., Rouhara, I., and Schmelzer, E. (2000) Cyst
germination proteins of the potato pathogen Phytophthora
infestans share homology with human mucins. Mol Plant
Microbe Interact 13: 32–42.
Grenville-Briggs, L.J., and van West, P. (2005) The biotrophic
stages of oomycete–plant interactions. Adv Appl Microbiol
57: 217–243.
Grenville-Briggs, L.J., Avrova, A.O., Bruce, C.R., Williams,
A., Whisson, S.C., Birch, P.R.J., et al. (2005) Elevated
amino acid biosynthesis in Phytophthora infestans during
appressorium formation and potato infection. Fungal Genet
Biol 42: 244–256.
Grenville-Briggs, L.J., Anderson, V.L., Fugelstad, J., Avrova,
A.O., Bouzenzana, J., Williams, A., et al. (2008) Cellulose
synthesis in Phytophthora infestans is required for normal
appressorium formation and successful infection of potato.
Plant Cell 20: 720–738.
Hahn, M., Neef, U., Struck, C., Gottfert, M., and Mendgen, K.
(1997) A putative amino acid transporter is specifically
expressed in haustoria of the rust fungus Uromyces fabae.
Mol Plant Microbe Interact 10: 438–445.
Hohl, H.R., and Suter, E. (1976) Host–parasite interfaces in a
resistant and a susceptible cultivar of Solanum tuberosum
inoculated with Phytophthora infestans: leaf tissue. Can J
Bot 54: 1956–1970.
Howard, R.J., Ferrari, M.A., Roach, D.H., and Money, N.P.
(1991) Penetration of hard substrates by a fungus employ-
ing enormous turgor pressures. Proc Natl Acad Sci USA
88: 11281–11284.
Jones, D.T., Taylor, W.R., and Thornton, J.M. (1992) The
rapid generation of mutation data matrices from protein
sequences. Comput App Biosci 8: 275–282.
Judelson, H.S., Tyler, B.M., and Michelmore, R.W. (1991)
Transformation of the oomycete pathogen, Phytophthora
infestans. Mol Plant Microbe Interact 4: 602–607.
Judelson, H.S., Ah-Fong, A.M.V., Aux, G., Avrova, A.O.,
Bruce, C., Cakir, C., et al. (2008) Profiling the asexual
development transcriptome of the late blight pathogen Phy-
tophthora infestans. Mol Plant Microbe Interact 21: 433–
447.
Kamoun, S. (2006) A catalogue of the effector secretome of
plant pathogenic oomycetes. Ann Rev Phytopathol 44:
41–60.
Kankanala, P., Czymmek, K., and Valent, B. (2007) Roles for
rice membrane dynamics and plasmodesmata during bio-
trophic invasion by the blast fungus. Plant Cell 19: 706–724.
Koch, E., and Slusarenko, A. (1990) Arabidopsis is suscep-
tible to infection by a downy mildew fungus. Plant Cell 2:
437–445.
Kottom, T.J., Kennedy, C.C., and Limper, A.H. (2008) Pneu-
mocystis PCINT1, a molecule with integrin-like features
that mediates organism adhesion to fibronectin. Mol Micro-
biol 67: 747–761.
© 2008 Scottish Crop Research Institute
Journal compilation © 2008 Blackwell Publishing Ltd, Cellular Microbiology, 10, 2271–2284
Lamour, K.H., Win, J., and Kamoun, S. (2007) Oomycete
genomics: new insights and future directions. FEMS Micro-
biol Lett 274: 1–8.
Latijnhouwers, M., and Govers, F. (2003) A Phytophthora
infestans G-protein beta subunit is involved in sporangium
formation. Eukaryot Cell 2: 971–977.
Latijnhouwers, M., Ligterink, W., Vleeshouwers, V.G., van
West, P., and Govers, F. (2004) A G-alpha subunit controls
zoospore motility and virulence in the potato late blight
pathogen Phytophthora infestans. Mol Microbiol 51: 925–
936.
Liu, Z., Bos, J., Armstrong, M., Whisson, S.C., da Cunha, L.,
Torto, T., et al. (2005) Patterns of diversifying selection in
the phytotoxin-like scr74 gene family of Phytophthora
infestans. Mol Biol Evol 22: 659–672.
Mendgen, K., and Hahn, M. (2002) Plant infection and the
establishment of fungal biotrophy. Trends Plant Sci 7: 352–
356.
Pryce-Jones, E., Carver, T., and Gurr, S.J. (1999) The roles
of cellulase enzymes and mechanical force in host pen-
etration by Erysiphe graminis f. sp. hordei. Physiol Mol
Plant Pathol 55: 175–182.
Randall, T.A., Dwyer, R.A., Huitema, E., Beyer, K., Cvitanich,
C., Kelkar, H., et al. (2005) Large-scale gene discovery in
the oomycete Phytophthora infestans reveals likely com-
ponents of phytopathogenicity shared with true fungi. Mol
Plant Microbe Interact 18: 229–243.
Rose, J.K., Ham, K.S., Darvill, A.G., and Albersheim, P.
(2002) Molecular cloning and characterization of gluca-
nase inhibitor proteins: coevolution of a counterdefense
mechanism by plant pathogens. Plant Cell 14: 1329–1345.
Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989) Molecu-
lar Cloning: A Laboratory Manual, 2nd edn. Cold Spring
Harbor, NY: Cold Spring Harbor Laboratory Press.
Soylu, S. (2004) Ultrastructural characterisation of the host–
pathogen interface in white blister-infected Arabidopsis
leaves. Mycopathologia 158: 457–464.
Soylu, E.M., Soylu, S., and Mansfield, J.W. (2004) Ultrastruc-
tural characterisation of pathogen development and host
responses during compatible and incompatible interactions
between Arabidopsis thaliana and Peronospora parasitica.
Physiol Mol Plant Pathol 65: 67–78.
Struck, C., Siebels, C., Rommel, O., Wernitz, M., and Hahn,
M. (1998) The plasma membrane H(+)-ATPase from the
biotrophic rust fungus Uromyces fabae: molecular charac-
terization of the gene (PMA1) and functional expression of
the enzyme in yeast. Mol Plant Microbe Interact 11: 458–
465.
Szabo, L.J., and Bushnell, W.R. (2001) Hidden robbers: the
role of fungal haustoria in parasitism of plants. Proc Natl
Acad Sci USA 98: 7654–7655.
Talbot, N.J. (2003) On the trail of a cereal killer: exploring the
biology of Magnaporthe grisea. Ann Rev Microbiol 57:
177–202.
Tian, M., Huitema, E., Da Cunha, L., Torto-Alalibo, T., and
Kamoun, S. (2004) A Kazal-like extracellular serine pro-
tease inhibitor from Phytophthora infestans targets the
tomato pathogenesis-related protease P69B. J Biol Chem
279: 26370–26377.
Tian, M., Benedetti, B., and Kamoun, S. (2005) A second
Kazal-like protease inhibitor from Phytophthora infestans
2284 A. O. Avrova et al.
inhibits and interacts with the apoplastic pathogenesis-
related protease P69B of tomato. Plant Physiol 138: 1785–
1793.
Tian, M., Win, J., Song, J., van der Hoorn, R., van der Knaap,
E., and Kamoun, S. (2007) A Phytophthora infestans
cystatin-like protein targets a novel tomato papain-like apo-
plastic protease. Plant Physiol 143: 364–377.
Torto, T.A., Li, S., Styer, A., Huitema, E., Testa, A., Gow,
N.A., et al. (2003) EST mining and functional expression
assays identify extracellular effector proteins from the plant
pathogen Phytophthora. Genome Res 13: 1675–1685.
Tozer, E.C., Liddington, R.C., Sutcliffe, M.J., Smeeton, A.H.,
and Loftus, J.C. (1996) Ligand binding to integrin aIIbb3 is
dependent on a MIDAS-like domain in the b3 subunit. J Biol
Chem 271: 21978–21984.
Tripathy, S., Pandey, V.N., Fang, B., Salas, F., and Tyler,
B.M. (2006) VMD: a community annotation database for
oomycetes and microbial genomes. Nucleic Acid Res.
34 (Database issue): D379–D381.
Tucker, S.L., and Talbot, N.J. (2001) Surface attachment and
pre-penetration stage development by plant pathogenic
fungi. Ann Rev Phytopathol 39: 385–417.
Vleeshouwers, V.G., van Dooijeweert, W., Govers, F.,
Kamoun, S., and Colon, L.T. (2000) The hypersensitive
response is associated with host and nonhost resistance to
Phytophthora infestans. Planta 210: 853–864.
Voegele, R.T., Struck, C., Hahn, M., and Mendgen, K. (2001)
The role of haustoria in sugar supply during infection of
broad bean by the rust fungus Uromyces fabae. Proc Natl
Acad Sci USA 98: 8133–8138.
Walker, C.A., Köppe, M., Grenville-Briggs, L.J., Avrova, A.O.,
Horner, N.R., McKinnon, A.D., et al. (2008) A DEAD-box
RNA helicase is required for normal zoospore development
in the potato late blight pathogen Phytophthora infestans.
Fungal Genet Biol 45: 954–962.
van West, P., and Vleeshouwers, V.G.A.A. (2004) The Phy-
tophthora infestans–host interaction. In Plant Pathogen
Interactions. Annual Plant Reviews, Vol. 11. Talbot, N.J.
(ed.). Oxford: Blackwell Publishing UK, pp. 219–242.
van West, P., de Jong, A.J., Judelson, H.S., Emons, A.M.C.,
and Govers, F. (1998) The ipiO gene of Phytophthora
infestans is highly expressed in invading hyphae during
infection. Fungal Genet Biol 23: 126–138.
van West, P., Kamoun, S., Van’t Klooster, J.W., and Govers,
F. (1999) Internuclear gene silencing in Phytophthora
infestans. Mol Cell 3: 339–348.
van West, P., Morris, B.M., Reid, B., Appiah, A.A., Osborne,
M.C., Campbell, T.A., et al. (2002) Oomycete plant patho-
gens use electric fields to target roots. Mol Plant Microbe
Interact 15: 790–798.
van West, P., Appiah, A.A., and Gow, N.A.R. (2003)
Advances in research on oomycete root pathogens. Phys
Mol Plant Pathol 62: 99–113.
Journal compilation © 2008 Blackwell Publishing Ltd, Cellular Microbiology, 10, 2271–2284
Whisson, S.C., van der Lee, T., Bryan, G., Waugh, R.,
Govers, F., and Birch, P.R.J. (2001) Physical mapping
across an avirulence locus of Phytophthora infestans using
a highly representative, large-insert bacterial artificial chro-
mosome library. Mol Genet Genomics 266: 289–295.
Whisson, S.C., Avrova, A., van West, P., and Jones, J.T.
(2005) A method for double-stranded RNA mediated tran-
sient gene silencing in Phytophthora infestans. Mol Plant
Pathol 6: 153–163.
Whisson, S.C., Boevink, P.C., Moleleki, L., Avrova, A.O.,
Morales, J.G., Gilroy, E.M., et al. (2007) A translocation
signal for delivery of oomycete effector proteins into host
plant cells. Nature 450: 115–118.
Zheng, H., Kunst, L., Hawes, C., and Moore, I. (2004) A
GFP-based assay reveals a role for RHD3 in transport
between the endoplasmic reticulum and Golgi apparatus.
Plant J 37: 398–414.
Supporting Information
Additional Supporting Information may be found in the online
version of this article:
Fig. S1. Copy number and alleles of Pihmp1 in P. infestans
isolate T30-4.
A. Pihmp1-hybridizing BAC clones digested with HindIII restric-
tion endonuclease, separated on 1% agarose gel and stained
with SYBR Gold.
B. Southern blot of gel in (A) hybridized to radiolabelled probe to
Pihmp1, revealing three hybridizing restriction fragments of dif-
fering sizes. Fragments shown in BACs5J1 and 15J23 represent
different RFLP alleles and contain the full Pihmp1 ORF, while the
hybridizing fragment shown in BAC 1B18 represents a truncated
non-expressed copy of Pihmp1. Molecular sizes are given at left
as kilo base pairs.
Table S1. Oligonucleotide primers used in sequencing, real-time
RT-PCR and transient RNAi.
Table S2. Protein distance among homologues of hmp1, cellu-
lose synthase 3 (cesA3), actin A (actA) and translation elongation
factor 1a (tef1a) from Phytophthora infestans (Pi), P. ramorum
(Pr), P. sojae (Ps) and Hyaloperonospora parasitica (Hp). Above
diagonal for each gene: protein distance measured across full
length of protein. Below diagonal for each gene (top to bottom):
protein distance measured across N-terminal half of protein,
protein distance measured across C-terminal half of the protein,
difference between protein distance from N- and C-terminal
halves. Protein distance was calculated from CLUSTALW align-
ment using the PROTDIST accessory program in the BioEdit
Sequence Alignment Editor.
Please note: Blackwell Publishing are not responsible for the
content or functionality of any supporting materials supplied by
the authors. Any queries (other than missing material) should be
directed to the corresponding author for the article.
© 2008 Scottish Crop Research Institute