Blackwell Science, LtdOxford, UKCMICellular Microbiology 1462-5814Blackwell Publishing Ltd, 200571014181431Original ArticleEntry domain of azurinT. Yamada et al.

Cellular Microbiology (2005) 7(10), 14181431 doi:10.1111/j.1462-5822.2005.00567.x

Internalization of bacterial redox protein azurin in

mammalian cells: entry domain and specificity

Tohru Yamada,1,2 Arsenio M. Fialho,1 Vasu Punj,2 often dictated by a small segment of the protein, which
Laura Bratescu,2 Tapas K. Das Gupta2 and can then be used as a vehicle to transport foreign proteins
Ananda M. Chakrabarty1* into such cells. Amphipathic peptides are used to facilitate
Departments of 1Microbiology and Immunology and uptake of DNA-cleaving metalloporphyrins as potential
2
Surgical Oncology, University of Illinois College of antitumour drugs in human fibroblasts HS68 or murine
Medicine, Chicago, IL 60612, USA. lymphocytic leukaemia L1210 cells (Chaloin et al., 2001).
Peptides, called cell-penetrating peptides, such as pene-
tratin, transportan, Tat (amino acids 4757 or 4860) and
Summary
the model amphipathic peptide MAP, have been widely
Azurin is a member of a group of copper-containing used as delivery vehicles for transporting pharmacologi-
redox proteins called cupredoxins. Different cupre- cally important substances such as antisense oligonu-
doxins are produced by different aerobic bacteria as clotides, proteins and peptides (Lindgren et al., 2000;
agents of electron transfer. Recently, we demon- Hallbrink et al., 2001). Such peptides, particularly the
strated that azurin enters into J774 and several types DNA-binding homeodomain of Antennapedia, a Droso-
of cancer cells leading to the induction of apoptosis. phila transcription factor, or the 21-residue peptide carrier
We now demonstrate that azurin is internalized in Pep-1, are internalized by many types of cells in culture
J774 or cancer cells in a temperature-dependent man- such as human HS68 or murine NIH-3T3 fibroblasts either
ner. Azurin shows preferential entry into cancer com- at 37C or at 4C, suggesting a penetration mechanism
pared with normal cells. An 28-amino-acid fragment different from the classical endocytosis (Morris et al.,
of azurin fused to glutathione S-transferase (GST) or 2001) and which requires no chiral receptor proteins. One
the green fluorescent protein (GFP), which are inca- of the most widely used peptides to transport pharmaco-
pable of entering mammalian cells by themselves, can logically active compounds in mammalian cells is the 11-
be internalized in J774 or human melanoma or breast amino-acid arginine-rich sequence (PTD, protein trans-
cancer cells at 37C, but not at 4C. Competition duction domain) of the human immunodeficiency virus
experiments as well as studies with inhibitors such type 1 (HIV-1) transactivator protein Tat (Schwarze et al.,
as cytochalasin D suggest that azurin may enter cells, 1999; 2000). Intraperitoneal injection of the 120 kDa b-
at least in part, by a receptor-mediated endocytic pro- galactosidase/Tat fusion protein results in the transcellular
cess. The 28-amino-acid peptide therefore acts as a transduction of the fusion protein into virtually all tissues
potential protein transduction domain (PTD), and can in mice, including the passage of the bloodbrain barrier
be used as a vehicle to transport cargo proteins such (Schwarze et al., 1999). This short peptide domain of HIV-
as GST and GSTGFP fusion proteins. Another mem- 1 Tat has been shown to mediate cell internalization of
ber of the cupredoxin family, rusticyanin, that has also large molecules or particles, including magnetic nanopar-
been shown to enter J774 and human cancer cells and ticles, phage vectors, liposomes and plasmid DNA.
exert cytotoxicity, does not demonstrate preferential Interestingly, unlike the other cell-penetrating peptides,
entry for cancer cells and lacks the structural features internalization of cargo proteins by full-length Tat or its 11-
characteristic of the azurin PTD. amino-acid transduction domain is significantly impaired
at 4C (Liu et al., 2000; Suzuki et al., 2002) and requires
interactions with receptors such as the heparan sulphate
Introduction
chains of the cell membrane heparan sulphate proteogly-
Much attention has recently been directed towards an cans (Tyagi et al., 2001). The use of the PTD of the Tat
understanding of how proteins or peptides enter mamma- protein in fusion with a peptide derived from the C-terminal
lian cells. The entry of a protein into a mammalian cell is domain of the tumour suppressor protein p53 (p53C
peptide) in its D-isomeric and inverted sequence (RI-
TATp53C) has been shown to allow activation of p53 in
Received 1 April, 2005; revised 29 April, 2005; accepted 3 May, 2005.
*For correspondence. E-mail pseudomo@uic.edu; Tel. (+1) cancer but not in normal cells. Treatment by RI-TATp53C
312 996 4586; Fax (+1) 312 996 6415. peptide of pre-clinical terminal peritoneal carcinomatosis

2005 Blackwell Publishing Ltd


and peritoneal lymphoma models results in significant
increase in life span and generation of disease-free ani-
mals (Snyder et al., 2004), thus demonstrating potential
interesting applications of the PTDs in cancer therapy. The
internalization of cargo proteins fused to the Tat PTD is
impaired at 4C, suggesting the role of an active endocytic
process (Tyagi et al., 2001) and indeed such a process
has been shown to originate from the cell membrane lipid
rafts involving caveolar endocytosis (Fittipaldi et al., 2003).
Most of the PTDs to date have been derived from viral
and mammalian sources (Hallbrink et al., 2001; Morris
et al., 2001). Although bacterial proteins such as cholera
toxin are known to enter mammalian cell cytosol (Sofer
and Futerman, 1995), the cytotoxicity of such proteins has
limited the use of bacterial proteins, or PTDs derived from
them, for transporting pharmacologically important cargos
in mammalian cells. We have recently described the
secretion of two redox proteins, azurin and cytochrome
c551, by Pseudomonas aeruginosa that enter mammalian
cells such as J774 cell line-derived macrophages and
induce apoptotic cell death (Zaborina et al., 2000). On
entry into J774 cells, azurin localizes in the cytosol and
nuclear fractions, and forms a complex with the tumour
suppressor protein p53, thereby stabilizing it and enhanc-
ing its intracellular level (Yamada et al., 2002a). A higher
intracellular level of p53 then triggers apoptosis.
Redox-negative apo-azurin devoid of copper or some
site-directed mutants of azurin defective in electron
transfer could form complexes with p53 and were found
to be cytotoxic. A redox-negative site-directed mutant
(M44KM64E), however, appeared to interfere in the oligo-
merization of p53 and demonstrated very little cytotoxicity.
This suggested that the redox activity of azurin was not
important for its cytotoxicity (Goto et al., 2003). The induc-
tion of azurin-mediated apoptosis was not limited to J774
cells. Azurin could also enter cancer cells such as human
melanoma UISO-Mel-2 (Yamada et al., 2002b) or human
breast cancer MCF-7 (Punj et al., 2004) cells. In both
cases, azurin allowed the elevation of the intracellular p53
levels, leading to enhanced Bax formation and induction
of apoptosis in such cells. Most interestingly, intraperito-
neal injection of azurin in nude mice harbouring
xenografted Mel-2 or MCF-7 human cancers led to statis-
tically significant regression of such cancers (Yamada
et al., 2002b; Punj et al., 2004), suggesting the potential
use of azurin in cancer therapy. Cytochrome c551, on the
other hand, could enter J774 cells but induced inhibition
of cell cycle progression at the G1 to S phase (Hiraoka
et al., 2004). A mutant form of cytochrome c551, V23DI59E
cytochrome c551, while having reduced ability to trigger
inhibition of cell cycle progression, allowed induction of
apoptosis in J774 cells (Hiraoka et al., 2004). Interest-
ingly, a similar mutant form of azurin, M44KM64E azurin,
where two hydrophobic amino acids were replaced by two
2005 Blackwell Publishing Ltd, Cellular Microbiology, 7, 14181431
Entry domain of azurin 1419
charged amino acids, had reduced ability to induce apo-
ptosis in J774 cells but had an enhanced ability to allow
inhibition of cell cycle progression at the G1 to S phase
in J774 cells (Yamada et al., 2004a). However, very little
is known about the mode of entry of bacterial redox pro-
teins such as azurin and cytochrome c551 in mammalian
cells. We report the domain of azurin that is involved in its
entry to J774 or UISO-Mel-2 cells and some of the char-
acteristics of the entry domain, including its ability to trans-
port foreign proteins such as glutathione S-transferase
(GST) or enhanced green fluorescent protein (GFP)
inside such cells.
Results
Entry of cupredoxins in mammalian cells
We previously demonstrated that azurin entered both
J774 (Yamada et al., 2002a) and human cancer (Yamada
et al., 2002b; Punj et al., 2004) cells. However, it was not
clear whether azurin was unique in this respect or whether
other purified cupredoxins might also enter mammalian
cells. The secretion of azurin by P. aeruginosa, a known
opportunistic pathogen, also implied that azurin might act
as a virulence factor because of its ability to induce apo-
ptotic death in macrophage-like J774 cells (Zaborina
et al., 2000). J774 cells are, however, not normal primary
macrophages but are ascites forms of murine reticulum
cell sarcoma with macrophage-like properties of adher-
ence, morphology, receptors for immunoglobulin and anti-
body-dependent lysis of target cells (Ralph and Nakoinz,
1975). We previously reported that while azurin induced
apoptosis in vivo in the tumour cells in nude mice, it
showed much less cytotoxicity towards normal tissues in
azurin-treated mice (Yamada et al., 2002b; Punj et al.,
2004) and indeed in in vitro studies, azurin showed much
less cytotoxicity in normal human mammary epithelial
cells such as MCF-10F than in MCF-7 breast cancer cells
(Punj et al., 2004). It was therefore of interest to examine
whether other cupredoxins besides azurin, particularly
those from non-pathogenic bacteria such as the acido-
philic chemolithotroph Acidithiobacillus ferrooxidans
which grows well at pH values 2.03.5 and produces
rusticyanin as an electron transfer agent during oxidation
of ferrous iron (Fe2+) to ferric iron (Fe3+) (Appia-Ayme
et al., 1999; Grossman et al., 2002), will enter J774 and
other mammalian cells to demonstrate cytotoxicity. The
entry of wild-type (WT) azurin and another cupredoxin
rusticyanin, which has also been shown to possess cyto-
toxicity (Yamada et al., 2004b), in J774 cells is shown in
Fig. 1A. In these experiments, the cupredoxins were con-
jugated with Alexa fluor 568 to fluoresce red and incu-
bated with J774 cells for 1 h at 37C at a concentration of
7 mM. The nucleus was stained blue with 4,6-diamidino-
1420 T. Yamada et al.
A B
Control Azurin Rusticyanin
J774
Peritoneal
C 45 min 6h
macrophages
MCF-7
Mast cells
MCF-10F
D E
Mel-2 Fibroblast Control
0.5 h Fibroblast
Mel-2 Fibroblast
MCF-10F
3.5 h
2-phenylindole (DAPI). A control without the proteins was
maintained. In both cases, the cupredoxins were seen to
enter J774 cells in the cytosol (Fig. 1A).
As J774 cells are ascitic forms of murine reticulum cell
sarcoma, it was of interest to determine whether a cupre-
doxin such as azurin can enter normal peritoneal mac-
rophages or mast cells. As previous studies (Punj et al.,
2004) showed a reduced cytotoxic activity of azurin
towards normal breast MCF-10F cells as contrasted with
the MCF-7 cells, it was also of interest to examine whether
azurin could enter such cells equally well. While azurin
could efficiently be internalized in J774 cells during 45 min
incubation, it was internalized very inefficiently in perito-
neal macrophages or mast cells (Fig. 1B). Even after 6 h
incubation, such cells showed only limited entry. Similarly,
while azurin efficiently entered the breast cancer cells
(MCF-7), it showed extremely reduced entry in the normal
mammary MCF-10F cells (Fig. 1C). A reduced level of
azurin in normal cells might explain a low level of apopto-
sis during in vitro treatment with MCF-10F cells (Punj
et al., 2004). Similar results were obtained with human
melanoma Mel-2 and normal fibroblast cells (Fig. 1D). To
demonstrate that indeed cellular entry is the major con-
straint in azurins ability to induce apoptosis in normal
cells, we microinjected azurin, as described previously
(Punj et al., 2004), in fibroblast and MCF-10F cells
(Fig. 1E) and looked for induction of apoptosis, leading to
nuclear DNA condensation and fragmentation. Significant
Fig. 1. Entry of bacterial cupredoxins in mam-
malian cells.
45 min 6h
A. Internalization of azurin and rusticyanin was
visualized by confocal microscopy. Both azurin
and rusticyanin (7 mM) were labelled with red
fluorescent dye Alexa fluor 568 and incubated
with J774 cells for 1 h at 37C. The nucleus is
labelled blue with DAPI. The scale bar corre-
sponds to 20 micrometers.
B and C. Wild-type azurin conjugated with
Alexa fluor 568 was incubated for 45 min or 6 h
with various mammalian cells including J774,
peritoneal macrophages, mast cells, human
breast cancer MCF-7 and human normal epi-
thelial MCF-10F cells as indicated. The inter-
nalization was visualized by confocal
microscopy.
D. Alexa fluor-conjugated azurin at 7 mM was
incubated with equal numbers of cells of vari-
ous cell lines such as Mel-2 and fibroblasts on
a coverslip at 37C for 0.5 and 3.5 h after which
30 min 5h images were taken.
E. Induction of apoptosis in normal fibroblast
and MCF-10F cells after microinjection of
azurin in such cells. Azurin mixed with dextran
and Texas red was microinjected as described
previously (Punj et al., 2004). Fluorescent
images of the microinjected cells were taken
with a Zeiss LSM 510 confocal laser micro-
scope after 30 min and 5 h.
nuclear DNA (labelled blue with DAPI) condensation and
fragmentation were observed in microinjected single cells
in 5 h but not during 30 min (Fig. 1E) incubation with
azurin, demonstrating azurins ability to induce apoptosis
once inside normal cells.
Defining the entry domain of azurin
To define azurin entry domain, we constructed a series of
GST fusions of 128-amino-acid-long azurin truncated at
the central region, the N- and the C-terminals (Fig. 2A),
and purified the fusion protein products (Fig. 2B). Using
Alexa fluor 568-conjugated WT azurin, GST and GST
azurin (GSTazu) fusion derivatives, we examined their
internalization in J774 cells at 37C during 1 h incubation.
The nucleus was stained blue with DAPI. While WT azurin
was internalized, GST remained at the periphery of the
cells and was not internalized (Fig. 3A). GSTazu 36128
and GSTazu 3689 were internalized as was GSTazu
3677 (Fig. 3A). Further truncations, however, demon-
strated that while GSTazu 5077 was internalized, GST
azu 3650 produced a punctate labelling, indicating a slow
and inefficient entry and possible localization on the cell
surface or in the endosomes. Further truncation of GST
azu 5077 to GSTazu 5066 and GSTazu 6777 dem-
onstrated punctate labelling, confirming that for efficient
internalization, amino acids 5077 were important
(Fig. 3A).
2005 Blackwell Publishing Ltd, Cellular Microbiology, 7, 14181431
 Entry domain of azurin 1421

A
1 128
WT azurin
aecsvdiqgn dqmqfntnai tvdksckqft vnlshpgnlp knvmghnwvl staadmqgvv tdgmasgldk dylkpddsrv iahtkligsg ekdsvtfdvs klkegeqymf fctfpghsal mkgtltlk
14 kDa
36 128
1, GST-azu 36-128 GST--- pgnlp knvmghnwvl staadmqgvv tdgmasgldk dylkpddsrv iahtkligsg ekdsvtfdvs klkegeqymf fctfpghsal mkgtltlk
10 kDa+26 kDa
36 89
2, GST-azu 36-89 GST--- pgnlp knvmghnwvl staadmqgvv tdgmasgldk dylkpddsrv iahtkligs
5.7 kDa+26 kDa
36 77
3, GST-azu 36-77 GST--- pgnlp knvmghnwvl staadmqgvv tdgmasgldk dylkpdd
4.4 kDa+26 kDa
36 50
4, GST-azu 36-50 GST--- pgnlp knvmghnwvl
1.6 kDa+26 kDa B
50 77 (kDa) 1 2 3 4 5 6 7 8
5, GST-azu 50-77 185
GST--- l staadmqgvv tdgmasgldk dylkpdd
2.9 kDa+26 kDa
50 66
6, GST-azu 50-66
GST--- l staadmqgvv tdgmas 52
1.6 kDa+26 kDa
67 77
7, GST-azu 67-77
GST--- gldk dylkpdd 31
1.2 kDa+26 kDa
8, GST
26 kDa 17

Fig. 2. A. Schematic representation of various truncated azurin constructs marked 18 as derived from WT azurin. The detailed procedures for
the constructs are given under Experimental procedures. Various sizes of azu gene were fused at the 3 end of the gst gene in frame.
B. GSTazu fusion proteins were purified after cellular growth and lysis, loaded on SDS-PAGE and visualized by Coomassie blue staining.
GSTazu 3650 (lane 4), 27.6 kDa, shows slightly aberrant mobility compared with GSTazu 5077 (lane 5), 28.9 kDa.

Azurin internalization is an energy-requiring process
To determine whether the internalization of azurin is medi-
ated by an energy-requiring active endocytic process, we
also examined the internalization of WT azurin and the
GSTazu fusion derivatives in J774 cells incubated at
4C. The fact that the 28-amino-acid azu 5077 fused to
GST was internalized in J774 cells at 37C (Fig. 3A) sug-
gested that this moiety could transport GST, which is not
by itself internalized in J774 cells (Fig. 3A). At 4C, how-
ever, internalization of WT azurin inside J774 cells during
1 h incubation was significantly reduced. Similar impair-
ment was also seen with GSTazu 36128 and GSTazu
3689 (Fig. 3B). Interestingly, the shorter GSTazu
3677 and GSTazu 5077 demonstrated more severe
impairment of internalization at 4C, suggesting that
extended sequences beyond amino acids 77 (amino acids
7789 and beyond) may contribute to a slow internaliza-
tion process at 4C, presumably via a cell-penetrating
mechanism that does not require receptor-mediated
endocytosis (Lindgren et al., 2000; Hallbrink et al., 2001).
There is now evidence that the C-terminal part of azurin
has structural similarities with some mammalian surface
protein ligands that bind surface-exposed receptors that
are highly expressed in cancer cells (Gough and Chothia,
2004), and azurin may also be internalized by binding with
such receptors. It is interesting to note that residues 50
77 is an amphipathic peptide with both a hydrophobic
(amino acids 5066) and a hydrophilic (amino acids 67
77) domain (Fig. 3B). While the amphipathic peptide could
transport GST inside J774 cells at 37C, the hydrophobic
(amino acids 5066) or the hydrophilic (amino acids
6777) domains by themselves were not proficient in
this regard even at 37C (Fig. 3A). Neither was active,
however, at 4C (Fig. 3B) so that GSTazu 5066 or
GSTazu 6777 were not internalized at 4C, similar to
the amphipathic GSTazu 5077 (Fig. 3B).
Energy-dependent internalization of the GSTGFPazu
5077 fusion protein in J774 and melanoma UISO-Mel-2
cells
The ability of the azu 5077 moiety to allow internalization
of GST raised an interesting question: can this amphip-
athic azurin peptide allow internalization of even higher-
molecular-weight proteins and will the process be energy-
dependent? To reduce costs, as well as any effect of Alexa
fluor 568 conjugated to the fusion proteins, we fused GST
with GFP to make a GSTGFP fusion derivative. Addition-
ally, we fused azu 5077 moiety to the GSTGFP (molec-
ular mass 53 kDa) fusion protein (Fig. 4A) and examined
the mobility of the purified GST, GSTGFP and GST
GFPazu 5077 fusion derivatives on SDS-PAGE, and
detected by Coomassie blue staining (Fig. 4B) and West-
ern blotting using anti-azurin antibody (Fig. 4C).

2005 Blackwell Publishing Ltd, Cellular Microbiology, 7, 14181431

1422 T. Yamada et al.

A PBS WT azurin GST GST-azu 36-128 GST-azu 36-89

GST-azu 36-77 GST-azu 36-50 GST-azu 50-77 GST-azu 50-66 GST-azu 67-77

PBS WT azurin GST GST-azu 36-128 GST-azu 36-89

B

GST-azu 36-77 GST-azu 50-77 GST-azu 50-66 GST-azu 67-77

50 67 77
lstaadmqgvvtdgmasgldkdylkpdd
3.0

-3.0

Fig. 3. Confocal microscopic images of J774 cells treated with 200 mg ml-1 GSTazu derivatives as depicted in Fig. 2A. Equal numbers of J774
cells were incubated with various GSTazu fusion proteins labelled with Alexa fluor 568. Cells were placed at 37C (A) or 4C (B) during protein
treatment (1 h) before visualization by confocal microscopy. Hydropathy plot of azu 5077 was generated by Kyte and Doolittle method with
GENETYX software.

2005 Blackwell Publishing Ltd, Cellular Microbiology, 7, 14181431

 Entry domain of azurin 1423

A D E
PBS GSTGFP GSTGFPazu 5077 PBS GSTGFP GSTGFPazu 5077
NH2-- GST GFP azu 5077 --COOH

7
-77
77

0-7
50-

u5
5 0
zu
FP
B C
GS urin

-az
rin

zu
P-a

FP
TG

azu
37C 37C

Ta
az

-GF

TG
T
WT

WT
GS

GS
GS

GS
(kDa) (kDa)
185 185
98 98

52 52

31 31
19 19
17
4C 4C
6 9
3
6

Fig. 4. Construction of GSTGFPazu 5077 fusion protein.

AC. (A) The gfp gene was introduced at the 3 end of gst gene (for GSTGFP) and azu 5077 oligonucleotide fragment was then ligated at the
3 end of the gfp gene in frame to produce GSTGFPazu 5077 fusion protein. GST, GSTGFP and GSTGFPazu 5077 were purified as
single fusion proteins from the cell lysates. Purified proteins, including WT azurin, were run on SDS-PAGE and detected by Coomassie blue
staining (B) and also by Western blotting using anti-azurin antibody (C). In (C), both monomeric azurin (14 kDa) and small amounts of a dimeric
form are seen. It should be noted that anti-azurin antibody does not cross-react with GST.
D and E. Temperature-dependence of GSTGFP or GSTGFPazu 5077 internalization in either J774 (D) or UISO-Mel-2 cells (E) cells. The
cells were treated with 4 mM GSTGFP or GSTGFPazu 5077 and incubated at 37C for 1 h. Green fluorescence derived from GSTGFP or
GSTGFPazu 5077 was visualized by confocal microscopy.

In order to see whether there is any difference in the
binding profile of GSTGFP and GSTGFPazu 5077,
we incubated both J774 and UISO-Mel-2 cells with GST
GFP and GSTGFPazu 5077 at 37C and at 4C and
the green fluorescence was localized using confocal
microscopy. In J774 cells, GSTGFP fusion protein bound
to the surface and was not internalized both at 37C and
at 4C (Fig. 4D). In contrast, GSTGFPazu 5077 was
found to be internalized at 37C but not at 4C, suggesting
that the azu 5077 peptide could allow internalization of
the 53 kDa GSTGFP in J774 cells in a temperature-
dependent manner (Fig. 4D). In UISO-Mel-2 cells, the
GSTGFP fusion protein was retained on the surface both
at 37C and at 4C (Fig. 4E). In contrast, similar to J774
cells, GSTGFPazu 5077 fusion protein was seen to
be internalized at 37C but not at 4C (Fig. 4E). Thus,
the azu 5077 peptide promotes internalization of a
53 kDa cargo protein in J774 or UISO-Mel-2 cells by an
energy-dependent process, presumably mediated by an
endocytic pathway.
Kinetics of the azurin PTD (GSTGFPazu 5077) entry
in J774 cells and the effect of inhibitors
The ability of the azurin PTD (azu 5077) to carry a
53 kDa cargo such as GSTGFP inside J774 or Mel-2
cells (Fig. 4D and E) allowed us to study the kinetics of
entry of this fusion protein in J774 cells and some param-
eters that affect this process. To obtain quantitative
assessment of the cargo-carrying PTD entry, we used
flow cytometry to measure GFP fluorescence, varying
either the concentrations of the fusion protein or the time
of incubation. Significant azurin PTD entry was observed
at concentrations above 1.0 mM and increasing entry was
observed with increasing concentrations of the fusion pro-
tein in a linear manner up to 8 mM (Fig. 5A). The flow
cytometric data were plotted diagrammatically as shown
in the right hand panel of Fig. 5A. When 0.45 mM or
1.8 mM fusion protein was used for entry determination at
different periods of incubation, maximal entry was
observed in about 1.0 h (Fig. 5B). As the entry of the
fusion protein was shown to be low at 4C, implying the
entry process (but not necessarily surface binding) as
energy-requiring (Fig. 4D and E), we tested the effect of
the protonophore carbonyl cyanide m-chlorophenylhydra-
zone (CCCP), a mitochondrial uncoupler of energy gen-
eration, both by flow cytometry (Fig. 5C) and by confocal
microscopy (using Alexa fluor-conjugated azurin; Fig. 5D,
panel CCCP, 20 mM) on azurin PTD (or intact azurin)
entry. There was substantial (about 40%) inhibition of
azurin PTD or azurin entry by CCCP under such condi-
tions. We also tested the effect of two other inhibitors,
cytochalasin D, a known inhibitor of receptor-mediated
endocytosis that disrupts the cellular microfilament net-
work, and Brefeldin A, which is known to disrupt the Golgi
apparatus and inhibit classical vesicle-mediated secretion
(Lippincott Schwartz et al., 1990; Elliott and OHare, 1997)
on the entry of Alexa fluor-conjugated azurin by confocal

2005 Blackwell Publishing Ltd, Cellular Microbiology, 7, 14181431

1424 T. Yamada et al.

A B
100
GST-GFP-azu 50-77 (0.45 mM) 70

Events
Control GST-GFP-azu 50-77
80 Control 0.5 h 60

Positive cells (%)

0.9 mM

Positive cells (%)

1h
1.8 mM 3h
Events

3.6 mM 50
60
7.2 mM 40
40 GST-GFP-azu 50-77 (1.8 mM) 30

Events
Control 20
20 0.5 h
1h 10
3h
100 101 102 103 104 0
0 2 4 6 8 10 0
Green fluorescence GST-GFP-azu 50-77 (mM) 100 101 102 103 104 0 1 2 3
Green fluorescence Time (h)

C D Control CCCP Brefeldin A Cytochalasin D

70
20 mM 18 mM 0.25 mM 0.5 mM
60
Positive cells (%)

50
Events

40
GST-GFP-azu 50-77 + CCCP
30
GST-GFP-azu 50-77
20
10
100 101 102 103 104
0
Green fluorescence
GST-GFP-azu 50-77 + + +
CCCP 10 20 mM

Fig. 5. A. Kinetic study for internalization of GSTGFPazu 5077 fusion protein. Green fluorescence was assayed in J774 cells treated with
various concentrations of GSTGFPazu 5077 at 37C for 1 h. Ten thousand cells were analysed by flow cytometry. The untreated cells (control)
in flow cytogram constituted an internal negative control. The percentages of GSTGFPazu 5077-positive cells are represented in the right
panel.
B. Time-dependence of internalization of GSTGFPazu 5077. J774 cells were incubated with 0.45 mM (upper left panel and open circle in right
panel) or 1.8 mM (lower left panel and closed circle in right panel) GSTGFPazu 5077 for indicated times at 37C and collected by flow cytometry.
C. J774 cells were pre-treated with 0, 10 or 20 mM CCCP for 1 h and incubated with 1.8 mM GSTGFPazu 5077 for an additional hour at 37C.
Entry of GSTGFPazu 5077 in the presence (10 mM, red; 20 mM, blue) or absence (green) of CCCP was analysed by FACS.
D. Inhibition of azurin internalization by CCCP and cytochalasin D but not by Brefeldin A. Alexa fluor 568-conjugated azurin (14 mM) was incubated
with UISO-Mel-2 at 37C for 1 h in the presence of PBS (control) or specified concentrations of CCCP, Brefeldin A and cytochalasin D. The extent
of azurin internalization in each case was followed by confocal microscopy.

microscopy. While Brefeldin A had no effect, increasing conjugated GSTazu 5077 (7 mM) in the presence of 7,
concentrations of cytochalasin D caused significant inhi- 14 and 56 mM of unlabelled azurin for 1 h at 37C before
bition of azurin internalization (Fig. 5D), suggesting that determining GSTazu 5077 entry. The results clearly
azurin internalization is mediated, at least in part, by a demonstrated that in comparison with labelled 7 mM GST
receptor-mediated endocytic process. The modest inhibi- azu 5077 alone (Fig. 6A, top), the presence of increasing
tion of azurin internalization by CCCP or cytochalasin D amounts of unlabelled azurin (7, 14 and 56 mM; Fig. 6A,
implies that part of azurin is bound on the surface, pre- middle) increasingly competed with the entry of labelled
sumably due to azurins structural similarity at the C- 7 mM GSTazu 5077, thus demonstrating a common
terminal end with a group of mammalian proteins that are receptor binding by unlabelled azurin when used in
ligands for binding with a group of surface-exposed recep- excess. In contrast, when similar concentrations (6, 12
tor tyrosine kinases that are known to be hyperexpressed and 48 mM) of unlabelled rusticyanin (Fig. 6A, bottom)
in cancer cells (Gough and Chothia, 2004). were used in the presence of labelled GSTazu 5077,
very little effect on the entry of GSTazu 5077 was seen.
These results appear to imply that azurin entry involves
Azurin and rusticyanin use different modes of entry:
binding with receptors that do not recognize rusticyanin.
competition experiments and structural differences
To determine whether a lack of competition of the entry
To demonstrate that azurin internalization involves recep- of the azurin PTD by rusticyanin may imply a different
tor binding, we did a competition experiment with unla- mode of entry of rusticyanin, we checked the specificity
belled azurin at 37C in the presence of 7 mM Alexa fluor- of azurin and rusticyanin entries in cancer and normal
conjugated GSTazu 5077. We incubated J774 cells cells. As shown earlier (Fig. 1BD), Alexa fluor-
without Alexa fluor-conjugated azurin, with Alexa fluor- conjugated azurin entered efficiently in UISO-Mel-2 and
conjugated GSTazu 5077 (7 mM) and with Alexa fluor- MCF-7 cancer cells but not in the normal mammary MCF-

2005 Blackwell Publishing Ltd, Cellular Microbiology, 7, 14181431

 Entry domain of azurin 1425

A B PBS Azurin Rusticyanin

0 mM 7 mM

Mel-2

7 mM 14 mM 56 mM
MCF-7

6 mM 12 mM 48 mM
MCF10A1

C
1 20 40 54 60 78 80 100 120 128
Azurin (1JZG_A)
16 140
Auracyanin B (1QHQ_A)
36 155
Rusticyanin (1A3Z)
1 98
Plastocyanin (1IUZ)
1 93
Pseudoazurin(1BQK)

D
Microorganism Accession number

Psae AAG08307 (51) -STAADMQGVVTDGMASGLDKDYLKPDD (77)

Pssy AAO58351 (51) -SKKADASAITTDGMSVGIDKDYVKPDD (77)
Neme AAF41888 (89) IAKAEDMDGVFKDGVGA-ADTDYVKPDD (115)
Vipa NP_800938 (52) -ADTANIQAVGTDGMSAGADNSYVKPDD (78)
Bobr AZBR (51) -TKTADMQAVEKDGIAAGLDNQYLKAGD (77)
Consensus DG D Y K D

Fig. 6. A. Internalization of GSTazu 5077 and its competitive inhibition in the presence of unlabelled cupredoxins. UISO-Mel-2 cells were
incubated with PBS or Alexa fluor 568-conjugated 7 mM GSTazu 5077 at 37C for 1 h (first row). Detailed procedures for competition with
unlabelled proteins are given under Experimental procedures. The same amounts (7 mM) of Alexa fluor-conjugated GSTazu 5077 were mixed
with 7, 14 and 56 mM unlabelled azurin (second row) or with 6, 12 and 48 mM unlabelled rusticyanin (third row). The fusion protein alone or the
mixtures of the fusion proteins with unlabelled cupredoxins were incubated with equal numbers of UISO-Mel-2 cells for 1 h at 37C before the
images were taken.
B. Entry specificity of cupredoxins, azurin and rusticyanin in cancer and normal cells. Alexa fluor 568-conjugated azurin (7 mM) or rusticyanin
(6 mM) was incubated with human melanoma (Mel-2) and breast (MCF-7) cancer and normal mammary epithelial cells (MCF-10A1). Internalization
of cupredoxins was determined by confocal microscopy.
C. Diagram showing the structural alignment of azurin with other cupredoxins, as computed by the VAST algorithm. The N-terminal extended
parts of auracyanin B and rusticyanin, being absent in azurin, have been omitted in the figure. The grey bars indicate the regions of azurin that
can be superimposed on residues from each neighbour. The blank spaces are unaligned regions. The azurin PTD where there is no alignment
with rusticyanin is highlighted by vertical dashed lines. The numbers in parentheses after the names of the cupredoxins are the protein database
accession numbers.
D. Multiple amino acid sequence alignment of the residues comprising the middle part in some known bacterial azurins. Bacterial genus and
species are abbreviated as follows: Psae, Pseudomonas aeruginosa; Pssy, Pseudomonas syringae; Neme, Neisseria meningitidis; Vipa, Vibrio
parahaemolyticus; Bobr, Bordetella bronchiseptica. Numbers of intervening amino acids are given in parenthesis. The CLUSTAL X software (Higgins
and Sharp, 1988) was used to generate this multiple sequence alignment.

10A1 cells (Fig. 6B). Alexa fluor-conjugated rusticyanin,
however, not only entered the cytosol of UISO-Mel-2 and
MCF-7 cancer cells, but also in the normal MCF-10A1
cells (Fig. 6B). Interestingly, however, unlike in the cancer
cells where rusticyanin was evenly distributed in the cyto-
sol, in MCF-10A1 cells, much of the rusticyanin was
sequestered in the perinuclear space surrounding the
nucleus. Thus, azurin and rusticyanin show different
modes of entry and cellular localization in cancer and
normal cells.
To explore the reasons for the differences in the mode
of entry of these two members of the cupredoxin family
towards mammalian cells, we checked both the amino
acid sequence identity and the structural similarities of the
azurin PTD with other members of the cupredoxin family.
The sequence identity between azurin and rusticyanin in

2005 Blackwell Publishing Ltd, Cellular Microbiology, 7, 14181431

1426 T. Yamada et al.
the azu 5077 region is less than 20%, as is true of
several other cupredoxins (Rienzo et al., 2000; Murphy
et al., 2002). A structural analysis using VAST algorithm
(Gibrat et al., 1996) between azurin and several members
of the cupredoxin family demonstrated a significant
similarity between azu 5077 region and that of the
cupredoxin auracyanin B, a cupredoxin from the green
thermophilic photosynthetic bacterium Chloroflexus
aurantiacus (Bond et al., 2001). Interestingly, however,
other members of the cupredoxin family, while demon-
strating structural similarity with other regions of azurin,
lacked this similarity with the azurin PTD (Fig. 6C).
Indeed, when compared with other proteins whose struc-
tures have been deposited in the protein database, there
was very little structural similarity between the azurin PTD
and other proteins, demonstrating the unique nature of the
azurin PTD. It remains to be seen whether the auracyanin
B domain, having structural identity with azurin PTD but
very little amino acid sequence identity (< 20%), will dem-
onstrate any entry specificity towards cancer cells.
A lack of structural identity between the azurin PTD and
other cupredoxins, except auracyanin B, raised an inter-
esting question: is this PTD conserved in other azurins
found in other bacteria, particularly pathogenic bacteria?
As the crystal structure of azurins from other pathogenic
bacteria has not been solved or presented in databases,
we performed a multiple amino acid sequence alignment
of the residues in the P. aeruginosa PTD region with
known sequences of other bacterial azurins from patho-
gens using the CLUSTAL X alignment program (Higgins and
Sharp, 1988). While the phytopathogenic Pseudomonas
syringae azurin showed high identity with P. aeruginosa
azurin PTD region (Fig. 6D), an azurin-like protein from
Neisseria meningitidis (Gotschlich and Seiff, 1987;
Kawula et al., 1987; Cannon, 1989) also showed signifi-
cant identity with the PTD domain of P. aeruginosa azurin
(Fig. 6D), as did azurins from Vibrio parahaemolyticus
and Bordetella bronchiseptica (Fig. 6D). A motif sequence
DGX5DX2YXKX2D was found conserved in all surface of normal cells. It would be of great interest to
these azurins.
Discussion
Cupredoxins such as azurin and rusticyanin have been
studied extensively for their electron transfer properties
(Rienzo et al., 2000; Murphy et al., 2002) but there is very
little information on their ability to enter mammalian cells
and produce cytotoxic effects. Our recent demonstration
that the copper-containing redox protein azurin and the
iron (haem)-containing redox protein cytochrome c551
enter J774 (Yamada et al., 2002a; Hiraoka et al., 2004)
and human cancer cells (Yamada et al., 2002b; Punj et al.,
2004) and induce either apoptosis or inhibition of cell
cycle progression thus raised the question: are cupredox-
ins in general capable of entering mammalian cells? If
they do, is there a common mode of entry? Are there
specificities with regard to which cells they enter prefer-
entially and why? It is important to note that not only
bacterial cupredoxins and cytochromes enter J774-like
mammalian cells, but we have reported that even mam-
malian cytochrome c can enter J774 cells when added
exogenously (Hiraoka et al., 2004). Is the mode of entry
of mammalian cytochrome c similar to that of bacterial
cytochrome c? The study reported here is the first step
towards answering many of these questions. The fact that
azurin and rusticyanin can enter mammalian cells and the
internalization of azurin is significantly impaired at 4C
suggests the possible involvement of a receptor-mediated
endocytic process in azurin internalization. This is con-
firmed by the competition experiments with unlabelled
azurin and by studies with inhibition by cytochalasin D. Do
all cupredoxins use similar receptors for entry? The com-
petition studies indicate that rusticyanin cannot compete
for azurin entry and presumably enters by a different
mechanism. Nothing is known about the entry mecha-
nisms of other cupredoxins that can exert cytotoxicity
towards mammalian cells (Yamada et al., 2004c).
An important question that we have to address in future
relates to the specificity of entry. As previously mentioned,
J774 cells are not normal macrophages but are ascites
forms of murine reticulum cell sarcoma (Ralph and
Nakoinz, 1975). It is interesting to note that azurin exhibits
preferential internalization in J774 and human breast can-
cer MCF-7 cells than the corresponding normal cells such
as primary peritoneal macrophages and mast cells or
mammary epithelial MCF-10F or MCF-10A1 cells. Similar
selective entry of azurin and induction of apoptosis has
also recently been reported for the human osteosarcoma
cell line U20S but not the normal L-02 cells (Ye et al.,
2005). If azurin entry is indeed due to a receptor-mediated
endocytic process, then it is likely that such receptors are
hyperexpressed on the surface of cancer cells than on the
isolate and characterize any putative receptors for azurin
binding and compare their abundance in cancer and nor-
mal cells. The entry domain of azurin can then be used
as a vehicle to transport pharmacologically important
compounds preferentially to cancer cells. It remains to be
seen whether other cupredoxins or cytochromes will
exhibit similar specificity for entry into cancer cells for
ultimate use in the treatment of human cancer.
An interesting aspect of the present study is our dem-
onstration of the ability of the azu 5077 peptide to enter
J774 and cancer cells and carry cargo proteins with them,
thus acting as a PTD. It is not clear from our data whether
the azurin PTD may still harbour the cytotoxicity domain.
We have recently obtained synthetic peptides correspond-
ing to azu 5077 and demonstrated their cytotoxicity in
2005 Blackwell Publishing Ltd, Cellular Microbiology, 7, 14181431

J774 and Mel-2 cells (unpubl. obs.). As azurin shows
preferential entry towards cancer cells (Figs 1BD and
6B), it could potentially be used as a vehicle to carry cargo
proteins such as the P. aeruginosa exotoxin A domain III
(PED III) that kills eukaryotic cells by ADP-ribosylating
elongation factor 2 and inhibiting protein synthesis (Kreit-
man and Pastan, 1995; Kreitman and Pastan, 2003). PED
III itself cannot enter eukaryotic cells and needs a vehicle
such as the domain II of P. aeruginosa exotoxin A for its
entry to the cytosol. Pastan and colleagues have used
various recombinant immunotoxins for targeting CD22,
CD25 or other cancer-specific targets in order to kill
cancer cells (Kreitman and Pastan, 1995; Kreitman and
Pastan, 2003). As azurin exerts its cytotoxicity through an
association with tumour suppressor p53 (Yamada et al.,
2002b; Punj et al., 2004) while PED III exerts its cytotox-
icity through a KDEL type of sequence in its C-terminus
for generalized inhibition of protein synthesis, it would be
interesting to use a bacterial protein vehicle such as the
azurin PTD by fusing it in the N-terminal of PED III to allow
the fusion protein to preferentially enter cancer cells and
kill them by a combined cytotoxic mechanism. Such inves-
tigations are presently under progress.
The preferential entry of azurin in cancer cells to exert
cytotoxicity raises another interesting question: what is the
physiological function of azurin in P. aeruginosa? Azurin-
like proteins are produced by many bacteria such as
Alcaligenes xylosoxidans, Methylobacillus flagellatum and
others (Fig. 6D). The major function of azurin in such
bacteria is believed to be its role in respiration as a com-
ponent of the electron transfer chain (Rienzo et al., 2000;
Murphy et al., 2002). Although widely believed to be
involved in dissimilatory nitrate reduction in P. aeruginosa,
in vivo studies have disproved an obligatory role of azurin
in denitrification and no significant physiological function
of this cupredoxin has been discerned in P. aeruginosa
(Vijgenboom et al., 1997). Vijgenboom et al. (1997) dem-
onstrated that the expression of the azurin gene (azu) was
mediated from two promoters P1 and P2. The proximal
promoter P1 was shown to be regulated by ANR, an
anaerobiosis regulator protein, while the distal promoter
P2 had the GG and GC motifs recognized by the alterna-
tive sigma factor sigma N. These promoters were shown
to be under the control of the stationary-phase sigma
factor, sigma S, and the function of azurin was believed
to involve electron transfer in response to redox stress,
although no effect of azurin was discernible to heat and
cold shock, nutritional shift-down or for growth on various
carbon sources. Thus, the physiological function of azurin
in P. aeruginosa is as yet not clarified. It is interesting to
note in this context that a mutation in the pseudoazurin
gene, encoding a cupredoxin also believed to be involved
in anaerobic respiration of nitrite in Paracoccus denitrifi-
cans, has no physiological effect on respiration, although
2005 Blackwell Publishing Ltd, Cellular Microbiology, 7, 14181431
Entry domain of azurin 1427
a copy of cytochrome c551 is essential under such condi-
tions (Pearson et al., 2003). An intriguing aspect of the
azu promoter region is the presence of a CpG island
(Fig. 7A) characteristic of those present upstream of many
human genes including tumour suppressor genes. The 5
CpG islands in the promoter regions of about 60% of
human genes, including the silenced alleles for selected
imprinted genes and the silenced genes of the inactive
X-chromosome of females, which has recently been
sequenced, are susceptible to methylation by DNA meth-
yltransferases that add the methyl groups to the carbon-
5 position of cytosine to generate methyl cytosine. Exten-
sive hypermethylation of the CpG islands upstream of
many tumour suppressor genes or genes involved in DNA
repair or cell cycle, leading to their silencing, is a charac-
teristic of many forms of human cancer (Esteller, 2005;
Wang et al., 2005). Whether the presence of a CpG island
(Fig. 7A), corresponding to Homo sapiens genome
sequence surrounding the NotI site, clone JNB-106R (Kut-
senko et al., 2002) with 88% sequence identity (Fig. 7B),
in the upstream region of the P. aeruginosa azu gene
allows its methylation and modulates the expression of
this gene either during growth in the environment or dur-
ing P. aeruginosa infection in human remains unknown. It
would be interesting to see whether deletion of this CpG
island from the azu promoter region in an in vitro tran-
scription assay will demonstrate any role of this island in
azu gene expression or whether either prokaryotic or
eukaryotic DNA methyltransferases will methylate such an
island to modulate azu gene expression. It is interesting
to note that this CpG island is unique in P. aeruginosa
present only in the upstream of the azu gene and is
absent in other Pseudomonas genomes and in the
genomes of other prokaryotes described in Fig. 6D.
A second intriguing aspect of azurins physiological
function, particularly in pathogenic bacteria, is its pre-
sumptive role in virulence. P. aeruginosa azurin appears
to preferentially attack cancer cells, thus allowing cancer
regression in vivo. Many bacteria are known to target
cancer cells preferentially to normal cells, but how they do
it is poorly understood (Sznol et al., 2000; Yu et al., 2004).
Whether preferential attack to cancer cells by various
microorganisms (Chakrabarty, 2003; Thomas-Tikhonenko
and Hunter, 2003) results from elaboration of azurin-like
proteins that might be surface-exposed is also unknown.
Most interestingly, members of the pathogenic Neisseria
species including the meningococcus and the gonococ-
cus produce two outer membrane proteins: a lipoprotein
called Lip which expresses an outer membrane antigenic
epitope known as H.8 which is common to all pathogenic
Neisseria species; and a second lipoprotein, called Laz.
Laz not only contains in its N-terminal a slightly modified
H.8 epitope but also an azurin domain in the C-terminal
that is highly homologous to P. aeruginosa azurin
1428 T. Yamada et al.

A
(-679)cagcagcgggtcgtagaaaccgttcacttcgagcaggcccagcggcttggcgtggtagcccagttggccccaggtccaga

cctcgaacagctcttccagggtgccgagcccgccaggcagggCGatgaaggCGtCGgccagttCGgccatgCGtgccttg

CGCGCGtgcatgcCGtctaccacttccaggcgggtcaggcccttgtggccgatctcggcctcctgcaggctctgcgggat

gatcccgatcacttcgccgccggcggccaatgcggcgtccgccacggtgcccatcagaccgaccgcgccgccaccgtaga

ccagggtcaggccgcgctcggccaggtgccggccgagggccacggcggcttcctggtagaccggggaagcgccggggctg

gcgccacagaatacgcagacggaacgcaaggtcatgatcgactcctgtcgggggtggaaaaaggcgcacagggtagcggc
P2
tgggagcgcttcgaccaagccgtgcgaagcgttgccggacgttgcgtcgcaggcgcgaagcggcacatctgtgctaaaac
P1
aggagTtccccgtagtaaacgccgggcagatcccgctcgatgccccgccacgtccggttcgggtttgacctgaatcagtg

gaactcggtgcccgatcgggcagtctgctctttcaggatTcatcgcccaacctgcctaggaggctgctccATG

B
P.aeruginosa: -576 ggcagggcgatgaaggcgtcggccagttcggccatgcgtgccttgcgcgcgtgcatgccg -517
|||||||||||||| |||||| ||||| | |||||||||| |||||||||||||||||
Homo sapiens: 159 ggcagggcgatgaaagcgtcgctcagttggtgcatgcgtgccatgcgcgcgtgcatgccg 218
Genomic sequence
Accession No. AJ323090

Fig. 7. Analysis of the Pseudomonas aeruginosa azurin promoter region.

A. P1 and P2 promoter regions described by Vijgenboom et al. (1997) are indicated with the transcriptional start site (T) in bold. The DNA region
underlined matched a genomic sequence of Homo sapiens; the CG motifs within this region are indicated in bold capital letters.
B. Sequence alignment (-576 to -517 nucleotides) of the P. aeruginosa azurin upstream region with the H. sapiens genome sequence
demonstrates a very high sequence identity (53/60 nucleotides, 88% identity).

(Gotschlich and Seiff, 1987; Kawula et al., 1987; Cannon,
1989). The size of the Lip protein can be somewhat vari-
able among the Neisseria species but generally contains
1120 pentameric AAEAP amino acid repeats, no
aromatic amino acids and a lipid-modified N-terminal
cysteine residue. Lip has been shown to be a potent
inflammatory mediator capable of inducing the release of
the chemokine interleukin 8 and the cytokine interleukin
6 as well as activation of the transcription factor NF-kB by
human endocervical epithelial cells and embryonic kidney
cells transfected with Toll-like receptor 2 (Fisette et al.,
2003). The Laz protein, on the other hand, contains imper-
fect AAEAP pentameric motifs at the N-terminal but an
127-amino-acid azurin domain at the C-terminal that is
highly homologous to the P. aeruginosa azurin (Gotschlich
and Seiff, 1987; Cannon, 1989). Similar to P. aeruginosa
azurin, inactivation of the azurin domain in Laz still
allowed the mutant to grow under both aerobic and anaer-
obic conditions in the presence of nitrate (Cannon, 1989),
suggesting a non-obligatory role of this moiety in anaero-
bic respiration of Neisseria species. No function of the Laz
protein is presently known. It would be interesting to clone
the azurin gene with or without the gene for the H.8
epitope and examine their gene products for cytotoxicity
against various normal and cancer cells. The presence of
the azurin moiety on the surface (outer membrane) of the
Neisseria species may provide some insights to how bac-
teria may target cancer cells preferentially, if the azurin
moiety can be shown to bind cancer cells in preference to
normal cells.
Experimental procedures
Cell cultures
The culture conditions of J774 and UISO-Mel-2 cells have been
described earlier (Yamada et al., 2002a,b; Goto et al., 2003).
Human normal fibroblast cells were cultured in MEM with
Eagles salt containing 2 mM L-glutamine, 0.1 mM MEM essen-
tial amino acids and supplemented with 10% heat-inactivated
fetal bovine serum, 100 units ml-1 penicillin and 100 mg ml-1
streptomycin. The culturing of MCF-7 and MCF-10F cells has
been described (Punj et al., 2004). Peritoneal macrophages and
mast cells were isolated as described previously (Melnikov
et al., 2000).
Plasmid constructions
The plasmids expressing fusion GST-truncated azurin derivatives
were constructed by a polymerase chain reaction (PCR) using
proof-reading DNA polymerase. For pGSTazu 36128, an
amplified PCR fragment was introduced into the BamHI and

2005 Blackwell Publishing Ltd, Cellular Microbiology, 7, 14181431


EcoRI sites of the commercial GST expression vector pGEX5X
(Amersham). The fragment was amplified with pUC19-azu as a
template and primers, 5-CGGATCC CCG GCA ACC TGC CGA
AGA ACG TCA TGG GC-3 and 5-CGGAATTC GCA TCA CTT
CAG GGT CAG GG-3, where the additionally introduced BamHI
and EcoRI sites are underlined, respectively, C-terminus trunca-
tion of azu gene was cumulatively performed by introducing a
stop codon using QuickChange site-directed mutagenesis kit
(Stratagene). For pGSTazu 3650, pGSTazu 3677 and
pGSTazu 3689, stop codons were introduced into Ser51,
Ser78 and Gly90 respectively. The plasmid carrying pGSTazu
36128 was used as template DNA. Three sets of oligonuclotides
for site-directed mutagenesis are shown as follows. For pGST
azu 3650: 5-GGC CAC AAC TGG GTA CTG TGA ACC GCC
GCC GAC ATG CAG-3 and 5-CTG CAT GTC GGC GGC GGT
TCA CAG TAC CCA GTT GTG GCC-3. For pGSTazu 3677:
5-CCT GAA GCC CGA CGA CTG ACG TGT CAT CGC CCA
CAC C-3 and 5-GGT GTG GGC GAT GAC ACG TCA GTC GTC
GGG CTT CAG G-3. For pGSTazu 3589: 5-CCA AGC TGA
TCG GCT CGT GAG AGA AGG ACT CGG TGA CC-3 and 5-
GGT CAC CGA GTC CTT CTC TCA CGA GCC GAT CAG CTT
GG-3. The plasmid pGSTazu 5077 and pGSTazu 6777
were generated by PCR using pGSTazu 3677 as a template
DNA. Amplified PCR fragments, azu 5077 and azu 6777, were
obtained using forward primer 5-CGGGATCC TGA GCA CCG
CCG CCG ACA TGC AGG G-3 and 5-CGGGATCC CCG GCC
TGG ACA AGG ATT ACC TGA AGC CCG-3, where additionally
introduced BamHI site is indicated by underlining. The reverse
primer, 5-CGGAATTC GCA TCA CTT CAG GGT CAG GG-3,
was utilized in both cases. The plasmid carrying gstazu 5077
was used for generating pGSTazu 5066 by introduction of a
stop codon in Gly67 using oligonuclotides as follows: 5-GAC
GGC ATG GCT TCC TGA CTG GAC AAG GAT TAC C-3 and 5-
GGT AAT CCT TGT CCA GTC AGG AAG CCA TGC CGT C-3.
The gfp gene encoding the GFP was also amplified by PCR.
Forward and reverse primers used were 5-CGGGATCC CCA
TGG TGA GCA AGG GCG-3 and 5-CGGAATTC CTT GTA CAG
CTC GTC CAT GCC G-3 containing BamHI and EcoRI sites at
the 5 end of each oligonuclotides. The resultant PCR fragment
was ligated into pGEX5X vector for creating pGSTGFP. For
plasmid DNA carrying gstgfpazu 5077, azu 5077 gene was
amplified by PCR with pGSTazu 5077 as a template and a set
of primers 5-CCGCTCGAG CCT GAG CAC CGC CGA CAT
GCA GGG-3 and 5-TTTTCCTTTTGCGGCCGC TCA GTC GTC
GGG CTT CAG GTA ATC C-3, where the introduced XhoI and
NotI sites are underlined respectively. Purified azu 5077 frag-
ment was introduced into pGSTGFP at XhoI and NotI unique
restriction enzyme sites.
Purification of proteins
Wild-type azurin was purified as described before (Yamada et al.,
2004a). Plasmid DNA, pET9a carrying the rus gene encoding the
cupredoxin rusticyanin from A. ferrooxidans, was kindly provided
by Dr Kazuhiko Sasaki, Central Research Institute of Electric
Power Industry, Chiba, Japan. Rusticyanin was isolated from
Eschirichia coli BL21 (DE3) harbouring the rus gene as reported
previously (Sasaki et al., 2003) with some modifications. We
used acetic acid buffer (pH 4.0) and CM-Sepharose (Sigma)
instead of b-alanin buffer (pH 4.0) and TSK-gel CM-650 column.
All recombinant GST fusion derivatives were purified as follows:
2005 Blackwell Publishing Ltd, Cellular Microbiology, 7, 14181431
Entry domain of azurin 1429
E. coli BL21 cells were used as the host strain. After induction
with 0.4 mM IPTG at early log phase of growth in L-broth, GST
fusion proteins were purified from cell extracts by using Glu-
tathione Sepharose 4B affinity chromatography and Sephadex
75 gel-filtration column with phosphate-buffered saline (PBS)
(Amersham). Purified proteins, WT azurin and GST derivatives
or other cupredoxins, labelled with Alexa fluor 568 (Molecular
Probes), were isolated according to manufacturers instructions.
Unbound free fluorescent chemical was removed by gel-filtration
column.
Azurin internalization in UISO-Mel-2 and fibroblast cells
UISO-Mel-2 and fibroblast cells were cultured on coverslips indi-
vidually. After overnight incubation, cells on coverslips were
washed with fresh media and all three cell lines on coverslips
were placed on a culture dish containing 7 mM WT azurin conju-
gated with Alexa fluor 568. The cells were then incubated for
varying periods at 37C under 5% CO2.
Competition assay
UISO-Mel-2 cells (1 104) were seeded on coverslips and
allowed to grow overnight at 37C. Cells were washed with pre-
warmed fresh media and then incubated with either 7 mM Alexa
fluor 568-conjugated GSTazu 5077 or a mixture of 7 mM of the
Alexa fluor-conjugated GSTazu 5077 and 7, 14 or 56 mM unla-
belled WT azurin. In separate experiments, competition for the
entry of 7 mM GSTazu 5077 was examined in the absence or
in the presence of 6, 12 or 48 mM rusticyanin. Controls in the
absence of Alexa fluor-conjugated GSTazu 5077 or the com-
peting proteins were maintained. After 1 h incubation with such
protein mixture at 37C, cells were washed with PBS to remove
excess proteins and nuclei were stained by DAPI. Entry of GST
azu 5077 was observed under confocal microscopy.
Confocal microscopy
For preparation of microscope samples, cells were cultured on
coverslips overnight at 37C. Cultured cells were placed at 37C
or 4C for 2 h before protein treatment. Pre-warmed 37C fresh
media or chilled 4C fresh media were mixed with red fluores-
cent-labelled (with Alexa fluor 568) cupredoxins or GST fusion
derivatives, and incubated with the cells for indicated times. The
cells were washed with PBS, and fixed with methanol at -20C
for 5 min. After washing with PBS twice and the addition of
mounting media containing 1.5 mg ml-1 DAPI for staining nuclei
(VECTASHILD, Vector), images were taken by a confocal
microscope.
Azurin internalization in J774 cells as measured by flow
cytometry
To analyse GSTGFPazu 5077 internalization by flow cytom-
etry, 1 105 J774 cells were seeded into six-well plate. After
overnight incubation, old media were replaced with fresh media
and cells were incubated with varying concentrations of GST
GFPazu 5077 at 37C. After GSTGFPazu 5077 treatment
for specified periods, cells were washed with PBS twice, and
1430 T. Yamada et al.
collected by FACS (BD FACS Calibur) using CellQuest software
(Becton Dickinson).
Structural similarity between azurin and other cupredoxins
The crystal structures of azurin from P. aeruginosa (1JZG_A),
auracyanin B from C. aurantiacus (1QHQ_A), rusticyanin from
A. ferrooxidans (1A3Z), plastocyanin from Ulva pertusa (1IUZ)
and pseudoazurin from Achromobacter cycloclastes (1BQK), as
shown in Fig. 6C, were aligned using the VAST algorithm (Gibrat
et al., 1996), available at http://www.ncbi.nlm.nih.gov/Structure/.
The amino acid sequence alignment displayed in Fig. 6D was
generated using CLUSTAL X (Higgins and Sharp, 1988).
Acknowledgements
This investigation was supported by a Public Health Service
Grant ES-04050-19 from the National Institute of Environmental
Health Sciences. A.M.F. acknowledges sabbatical grants from
Fundacao para Ciencia e Tecnologia (FCT) BSAB-457, Portugal,
and Fundacao Calouste Gulbenkian.
References
Appia-Ayme, C., Guiliani, N., Ratouchniak, J., and Bonnefoy,
V. (1999) Characterization of an operon encoding two c-
type cytochromes, and aa(3)-type cytochrome oxidase,
and rusticyanin in Thiobacillus ferrooxidans ATCC 33020.
Appl Environ Microbiol 65: 47814787.
Bond, C.S., Blankenship, R.E., Freeman, H.C., Guss, J.M.,
Maher, M.J., Selvaraj, F.M., et al. (2001) Crystal structure
of auracyanin, a blue copper protein from the green ther-
mophilic photosynthetic bacterium Chloroflexus aurantia-
cus. J Mol Biol 306: 4767.
Cannon, J.G. (1989) Conserved lipoproteins of pathogenic
Neisseria species bearing the H.8 epitope: lipid-modified
azurin and H.8 outer membrane protein. Clin Microbiol Rev
2: S1S4.
Chakrabarty, A.M. (2003) Microorganisms and cancer: quest
for a therapy. J Bacteriol 185: 26832686.
Chaloin, L., Bigeye, P., Loup, C., Marin, M., Galled, N.,
Piechaczyk, M., et al. (2001) Improvement of porphyrin
cellular delivery and activity by conjugation to a carrier
peptide. Bioconjug Chem 12: 691700.
Elliott, G., and OHare, P. (1997) Intercellular trafficking and
protein delivery by a herpesvirus structural protein. Cell 88:
223233.
Esteller, M. (2005) DNA methylation and cancer therapy: new
developments and expectations. Curr Opin Oncol 17: 55
60.
Fisette, P.L., Ram, S., Andersen, J.M., Guo, W., and Ingalls,
R.R. (2003) The Lip lipoprotein from Neisseria gonor-
rhoeae stimulates cytokine release and NF-kB activation in
epithelial cells in a Toll-like receptor 2-dependent manner.
J Biol Chem 278: 4625246260.
Fittipaldi, A., Ferari, A., Zoppe, M., Arcangeli, C., Pelligrini,
V., Beltram, F., and Giacca, M. (2003) Cell membrane lipid
rafts mediate caveolar endocytosis of HIV-1 Tat fusion pro-
teins. J Biol Chem 278: 3414134149.
Gibrat, J.F., Madej, T., and Bryant, S.H. (1996) Surprising
similarities in structure comparison. Curr Opin Struct Biol
6: 377385.
Goto, M., Yamada, T., Kimbara, K., Horner, J., Newcomb, M.,
Das Gupta, T.K., and Chakrabarty, A.M. (2003) Induction
of apoptosis in macrophages by Pseudomonas aeruginosa
azurin: tumour-suppressor protein p53 and reactive oxygen
species, but not redox activity, as critical elements in cyto-
toxicity. Mol Microbiol 47: 549559.
Gotschlich, E.C., and Seiff, M.E. (1987) Identification and
gene structure of an azurin-like protein with a lipoprotein
signal peptide in Neisseria gonorrhoeae. FEMS Microbiol
Lett 43: 253255.
Gough, J., and Chothia, C. (2004) The linked conservation
of structure and function in a family of high diversity: the
monomeric cupredoxins. Structure 12: 917925.
Grossmann, T.G., Hall, J.F., Kanbi, L.D., and Hasnain, S.S.
(2002) The N-terminal extension of rusticyanin is not
responsible for its acid stability. Biochemistry 41: 3613
3619.
Hallbrink, M., Floren, A., Elmquist, A., Pooga, M., Barfai, T.,
and Langel, U. (2001) Cargo delivery kinetics of cell-
penetrating peptides. Biochim Biophys Acta 1515: 101
109.
Higgins, D.G., and Sharp, P.M. (1988) CLUSTAL: a package
for performing multiple sequence alignment on a micro-
computer. Gene 73: 237244.
Hiraoka, Y., Yamada, T., Goto, M., Das Gupta, T.K., and
Chakrabarty, A.M. (2004) Modulation of mammalian cell
growth and death by prokaryotic and eukaryotic cyto-
chrome c. Proc Natl Acad Sci USA 101: 64276432.
Kawula, T.H., Spinola, S.M., Klapper, D.G., and Cannon, J.G.
(1987) Localization of a conserved epitope and an azurin-
like domain in the H.8 protein of pathogenic Neisseria. Mol
Microbiol 1: 179185.
Kreitman, R.J., and Pastan, I. (1995) Importance of the
glutamate residue of KDEL in increasing the cytotoxicity of
Pseudomonas exotoxin derivatives and for increased bind-
ing to the KDEL receptor. Biochem J 307: 2937.
Kreitman, R.J., and Pastan, I. (2003) Immunobiological treat-
ments of hairy-cell leukemia. Best Pract Res Clin Haematol
16: 117133.
Kutsenko, A.S., Gizatullin, R.Z., Al-Amin, A.N., Wang, F.,
Kvasha, S.M., Podowski, R.M., et al. (2002) NotI flanking
sequences: a tool for gene discovery and verification of the
human genome. Nucleic Acids Res 30: 31633170.
Lindgren, M., Hallbrink, M., Prochiantz, A., and Langel, U.
(2000) Cell-penetrating peptides. Trends Pharmacol Sci
21: 99103.
Lippincott Schwartz, J., Donaldson, J.G., Schweizer, A.,
Berger, E.G., Hauri, H.P., Yuan, I.C., and Klausner, R.D.
(1990) Microtubule-dependent retrograde transport of pro-
teins into the ER in the presence of brefeldin A suggests
an ER recycling pathway. Cell 60: 821836.
Liu, Y., Jones, M., Hingtgen, C.M., Bu, G., Laribee, N., Tanzi,
R.E., et al. (2000) Uptake of HIV-1 tat protein mediated by
low-density lipoprotein receptor-related protein disrupts the
neuronal metabolic balance of the receptor ligands. Nat
Med 6: 13801387.
Melnikov, A., Zaborina, O., Dhiman, N., Prabhakar, B.S.,
Chakrabarty, A.M., and Hendrickson, W. (2000) Clinical
and environmental isolates of Burkholderia cepacia exhibit
differential cytotoxicity towards macrophages and mast
cells. Mol Microbiol 36: 14811493.
Morris, M.C., Depollier, J., Mery, J., Heitz, F., and Divita, G.
2005 Blackwell Publishing Ltd, Cellular Microbiology, 7, 14181431

(2001) A peptide carrier for the delivery of biologically
active proteins into mammalian cells. Nat Biotechnol 19:
11731176.
Murphy, L.M., Dodd, F.E., Yousafzai, F.K., Eady, R.R., and
Hasnain, S.S. (2002) Electron donation between copper
containing nitrite reductases and cupredoxins: the nature
of proteinprotein interaction in complex formation. J Mol
Biol 315: 859871.
Pearson, I.V., Page, M.D., van Spannig, R.J.M., and Fergu-
son, S.J. (2003) A mutant of Paracoccus denitrificans with
disrupted genes coding for cytochrome c550 and pseudoa-
zurin establishes these two proteins as the in vivo electron
donors to cytochrome cd1 nitrite reductase. J Bacteriol 185:
63086315.
Punj, V., Bhattacharyya, S., Saint-Dic, D., Vasu, C., Cunning-
ham, E.A., Graves, J., et al. (2004) Bacterial cupredoxin
azurin as an inducer of apoptosis and regression in human
breast cancer. Oncogene 23: 23672378.
Ralph, P., and Nakoinz, I. (1975) Phagocytosis and cytolysis
by a macrophage tumour and its cloned cell line. Nature
257: 393394.
Rienzo, F.D., Gabdoulline, R.R., Menziani, M.C., and Wade,
R.C. (2000) Blue copper proteins: a comparative analysis
of their molecular interaction properties. Protein Sci 9:
14391454.
Sasaki, K., Ida, C., Ando, A., Matsumoto, N., Saiki, H., and
Ohmura, N. (2003) Respiratory isozyme, two types of rus-
ticyanin of Acidithiobacillus ferrooxidans. Biosci Biotechnol
Biochem 67: 10391047.
Schwarze, S.R., Ho, A., Vocero-Akbani, A., and Dowdy, S.F.
(1999) In vivo protein transduction: delivery of a biologically
active protein into the mouse. Science 285: 15691572.
Schwarze, S.R., Hruska, K.A., and Dowdy, S.F. (2000) Pro-
tein transduction: unrestricted delivery into all cells? Trends
Cell Biol 10: 290295.
Snyder, E.L., Meade, B.R., Saenz, C.C., and Dowdy, S.F.
(2004) Treatment of terminal peritoneal carcinomatosis by
a transducible p53-activating peptide. PLoS Biol 2: 186
193.
Sofer, A., and Futerman, A.H. (1995) Cationic amphiphilic
drugs inhibit the internalization of cholera toxin to the Golgi
apparatus and the subsequent elevation of cyclic AMP.
J Biol Chem 270: 1211712122.
Suzuki, T., Futaki, S., Niwa, M., Tanaka, S., Ueda, K., and
Sugiura, Y. (2002) Possible existence of common internal-
ization mechanisms among arginine-rich peptides. J Biol
Chem 277: 24372443.
Sznol, M., Lin, S.L., Bermudes, D., Zheng, L.-M., and King,
I. (2000) Use of preferentially replicating bacteria for the
treatment of cancer. J Clin Invest 105: 10271030.
Thomas-Tikhonenko, A., and Hunter, C.A. (2003) Infection
and cancer: the common vein. Cytokine Growth Factor Rev
14: 6777.
2005 Blackwell Publishing Ltd, Cellular Microbiology, 7, 14181431
Entry domain of azurin 1431
Tyagi, M., Rusnati, M., Presta, M., and Giacca, M. (2001)
Internalization of HIV-1 tat requires cell surface hepa-
ran sulfate proteoglycans. J Biol Chem 276: 3254
3261.
Vijgenboom, E., Busch, J.E., and Canters, G.W. (1997) In
vivo studies disprove an obligatory role of azurin in denitri-
fication in Pseudomonas aeruginosa and show that azu
expression is under control of RpoS and ANR. Microbiol-
ogy 143: 28532863.
Wang, W., Huper, G., Guo, Y., Murphy, S.K., Olson,J.A.,
Jr, and Marks, J.R. (2005) Analysis of methylation
sensitive transcription identifies GADD45a as a frequently
methylated gene in breast cancer. Oncogene 24: 2705
2714.
Yamada, T., Goto, M., Punj, V., Zaborina, O., Kimbara, K.,
Das Gupta, T.K., and Chakrabarty, A.M. (2002a) The bac-
terial redox protein azurin induces apoptosis in J774 mac-
rophages through complex formation and stabilization of
the tumor suppressor protein p53. Infect Immun 70: 7054
7062.
Yamada, T., Goto, M., Punj, V., Zaborina, O., Chen, M.L.,
Kimbara, K., et al. (2002b) Bacterial redox protein azurin,
tumor suppressor protein p53, and regression of cancer.
Proc Natl Acad Sci USA 99: 1409814103.
Yamada, T., Hiraoka, Y., Ikehata, M., Kimbara, K., Avner,
B.S., Das Gupta, T.K., and Chakrabarty, A.M. (2004a)
Apoptosis or growth arrest: modulation of tumor suppres-
sor p53s specificity by bacterial redox protein azurin. Proc
Natl Acad Sci USA 101: 47704775.
Yamada, T., Hiraoka, Y., Das Gupta, T.K., and Chakrabarty,
A.M. (2004b) Rusticyanin, a bacterial electron transfer pro-
tein, causes G1 arrest in J774 and apoptosis in human
cancer cells. Cell Cycle 3: 11821187.
Yamada, T., Hiraoka, Y., Das Gupta, T.K., and Chakrabarty,
A.M. (2004c) Regulation of mammalian cell growth and
death by bacterial redox proteins: relevance to ecology and
cancer therapy. Cell Cycle 3: 752755.
Ye, Z.M., Miao, X.D., Yang, D.S., Xu, R.Z., Huang, X., and
Ge, F.F. (2005) Selective inducement effect of bacterial
redox protein azurin on apoptosis of human osteosarcoma
cell line U20S. Ai Zheng 24: 298304 (in Chinese).
Yu, Y.A., Shabahang, S., Timiryasova, T.M., Zhang, Q.,
Beltz, R., Gentschev, I., et al. (2004) Visualization of tumors
and metastases in live animals with bacteria and vaccinia
virus encoding light-emitting proteins. Nat Biotechnol 22:
313320.
Zaborina, O., Dhiman, N., Chen, M.L., Kostal, J., Holder, I.A.,
and Chakrabarty, A.M. (2000) Secreted products of a
nonmucoid Pseudomonas aeruginosa strain induce two
modes of macrophage killing: external-ATP-dependent,
P2Z-receptor-mediated necrosis and ATP-independent,
caspase-mediated apoptosis. Microbiology 146: 2521
2530.