Download Glycomics Aims To Interpret the Third Molecular Language of Cells

Glycomics Aims To Interpret the Third
Molecular Language of Cells
Some microbes exploit carbohydrate receptors to enter
and disrupt particular host cells
Alison A. Weiss and Suri S. Iyer
eciphering the chemical codes used
by cells provides valuable insights,
many of which will help toward
ameliorating diseases. In contrast
to impressive progress decoding
genomics and proteomics, efforts to crack the
code for the third major class of biopolymers,
carbohydrates, continue to prove frustrating.
Glycomics, the systematic study of carbohydrate
polymers and oligosaccharides, lags a distant
third compared to genomics and proteomics—
despite the ubiquitous presence of carbohydrates at cell surfaces, where cells engage in
important molecular-based signaling.
Glycolipids and glycoproteins play important
but poorly understood roles in regulating processes in mammalian cells and maintaining human health. At the molecular level, pathogens
exploit carbohydrate structures that decorate
D
Summary
• The many ways in which nine common
monosaccharide units are linked complicates
efforts to describe—and to understand the biological impacts of—sugar-containing polymers.
• Some viruses and bacterial toxins depend on
carbohydrates that decorate mammalian cellsurface receptors to gain entry.
• The preferential recognition by influenza virus
of specific forms of sialic acid has significant
implications for viral transmission and pathogenesis.
• In each case, the interactions of botulinum neurotoxins, Shiga toxin, and pertussis toxin with
particular carbohydrate receptors helps to explain tissue specificity and potency.
mammalian cell surfaces, often using those
structures as a means to infect such cells. By
focusing on the glycolipids and glycoproteins
that microbial pathogens target, we can begin to
decipher the complex language of cell surface
carbohydrates. Glycomics is providing us a
more comprehensive approach to understanding how microbes interact with such carbohydrate receptors and could lead to the development of novel therapeutics and diagnostics.
Nucleic Acids, Proteins, and Sugars
Embody Distinctive Coding
DNA, RNA, and proteins embody linear
codes—arrays of four nucleotides in the case of
nucleic acids, and 20 amino acids for proteins.
In general, mammalian carbohydrates are composed of nine distinct types of monosaccharides along with some variations (Fig. 1).
Moreover, the ways in which carbohydrates
are linked further complicates efforts to describe sugar-containing polymers—let
alone try to understand the biological effects of those different linkages. For example, any pair of six-carbon monosaccharides
can be linked in 11 different ways (Fig. 2).
Assuming that nine distinct types of
monosaccharides are available, there are
9 ⫻ 9 ⫻ 11, or 891, 2-letter words for these
sugars. In comparison, two amino acids can
be joined in only one way, creating only
20 ⫻ 20, or 400, 2-letter words.
With each additional unit, the number of
potential combinations of sugars rapidly
outpaces those for comparable peptides and
proteins. For example, although six amino
Alison A. Weiss is
Professor in the
Department of Biochemistry, Molecular Genetics and
Microbiology and
Suri S. Iyer is Assistant Professor in
the Chemical and
Biosensors group,
Department of
Chemistry, University of Cincinnati,
Cincinnati, Ohio.
Volume 2, Number 10, 2007 / Microbe Y 489
containing materials serve disparate
functions and some are remarkably
multifunctional. For example, the
carbohydrate-based molecule, hyaluronan, plays important roles in ovulation
and fertilization, morphogenesis, tissue
remodeling, atherosclerosis, restenosis,
and other cellular processes. Some
streptococcal strains also contain hyaluronan in their capsular material; its
nonantigenic quality enables these microbes to avoid the host immune system.
FIGURE 1
4
5
1
2
3
6
4
O
HO
HO
OH
OH
OH
6
5
1
HO
HO
OH
2
1
2
3
NH
O
Glucose (Glc)
O
5
HO
OH
3
OH
OH
6
4
O
OH
OH
Galactose (Gal)
N-acetyl glucosamine
(GlcNAc)
OH
4
OH
6
4
O
5
1
HO
2
3
OH
O
5
HO
HO
2
6 OH
O
O
5
HO
HO
OH
3
NH
O
4
1
2
3
OH 1
Xylose (Xyl)
OH
Glycoproteins and Glycolipids
OH
Glucuronic acid (GlcA)
N-acetyl galactosamine
(GalNAc)
HO
OH
4
6
5
HO
HO
OH
O
5
6
1
2
3
Mannose (Man)
O
4
OH
OH
HO
OH
3
1
OH
2
Fucose (Fuc)
HO
9
8
1 COOH
OH
7
6
O
AcHN
4
5
HO
2
OH
3
N-acetylneuraminic acid
or sialic acid (NeuAc)
The nine common sugar “letters” of glycomics. The color symbol for each sugar is
pictured next to the sugar name.
acids can be combined to form 6.4 ⫻ 107 distinct hexapeptides, six monosaccharides theoretically yield 1.44 ⫻ 1015 distinct hexasaccharides. Positional isomerism, conformational
changes, and hydrogen bonding can further alter oligosaccharide structures, and each variant
has distinctive physicochemical and cellsignaling properties.
While nucleic acids and proteins can be unambiguously represented through the use of oneletter code words, there is no comparably simple
way to represent combinations of complex sugars. To deal with the greater complexity of carbohydrates, Ajit Varki of the University of California, San Diego, devised a way to represent
glycan structures (see www.functionalglycomics
.org/glycomics/molecule/jsp/carbohydrate/carb
MoleculeHome.jsp). Each sugar type with the
same mass uses the same symbol. Thus, for example, hexoses are circles; glucose is a blue circle, galactose a yellow circle, and mannose a
green circle. In nature, some carbohydrate-
490 Y Microbe / Volume 2, Number 10, 2007
In many instances, specific oligosaccharides are attached to particular proteins, making them glycoproteins and
adding further overall complexity. Carbohydrates typically are linked to proteins in one of two ways. If a protein is
O-glycosylated, the hydroxyl of a
serine, threonine, or hydroxyproline
amino acid residue becomes linked to
the anomeric carbon (carbon number 1
of the carbohydrate residues in Fig. 1)
of N-acetylgalactosamine. If N-glycosylated, the sugar, typically N-acetylglucosamine, becomes directly linked to
the amide nitrogen of an asparagine
residue.
These two types of glycosylation differ in
several other ways. O-linked sugars typically are
added one at a time to a protein, and generally
contain only a few residues, typically no more
than four. When a protein is N-glycosylated,
however, oligosaccharide structures already
containing as many as 14 residues are assembled
separately before being linked to aspargine residues on the recipient protein. Once attached,
the sugar unit can be further modified by enzymes that either add or delete sugar residues.
There is more diversity in O-glycosylation than
N-glycosylation because of the different biochemical pathways that lead to these products.
The current consensus is that sugars promote
proper folding of proteins by stabilizing particular conformations and, in some cases, protecting the protein portion of the glycoprotein
against degradative enzymes. The semiflexible
nature of the sugars in a glycoprotein allows for
many more conformations than would their
Weiss Studies both “Northern,” or Respiratory, and “Southern,”
or Diarrheal Pathogens
“I find it hard to accept a future
where death can lurk in your spinach,” says microbiologist Alison
Ann Weiss of the University of
Cincinnati, who describes herself
as an “avid gardener.” Doubtless
her hands-on interest in food
safety was a factor in her adding
Shiga toxin to the set of bacterial
toxins that she studies.
One of Weiss’ current projects
focuses on in vivo signals that
promote Shiga toxin expression
and its release from Escherichia
coli O157:H7, as well as “why
one antigenic form, Stx2, is 100
to 1,000 times more potent than
Stx1,” she says. “What is especially interesting is that Stx1 is
more toxic to cells in culture than
Stx2, just the opposite of what is
seen in intact animals.”
Weiss, 54, was born in Milwaukee, the second of six children.
Although mainly a “city kid,”
early in life she learned firsthand
about farming from her grandfather, who owned a dairy farm.
“Sometimes we would graze a
steer on his land, and once we
butchered it in our basement,”
she recalls. Her grandfather’s
farm was at the top of a hill, and
her family had a garden at the
bottom that was “fertilized by the
manure that gently decomposed
on the way down.
“Times have changed,” she
points out. “Unlike people, cattle
are not susceptible to Shiga toxin
and can carry E. coli O157:H7
asymptomatically. Crops exposed
to fecal runoff are a major source
for human infection.”
Weiss grew up during the Vietnam War era “when young people challenged authority and con-
ventional wisdom at every level.”
So it was in her household. “My
father and I loved each other, but
we also loved to argue, a trait he
also shared with his father,” she
says. “We would have at it over
‘conventional wisdom’ every night
at the dinner table. Later, I learned
our arguments gave my siblings
indigestion.”
In those days, she continues,
“Women were not encouraged to
pursue careers in science. When I
told my mother I didn’t want to
go to college, she was pragmatic
and told me to take typing
courses, which she knew would
provide exactly the motivation I
needed.” Weiss received her A.B.
in 1975 from Washington University in St. Louis and her M.S. in
1981 from the University of
Washington in Seattle. She received her Ph.D. in 1983 from
Stanford University, where her
thesis advisor was Stanley Falkow.
“Stanley has the extraordinary
ability to inspire everyone to perform beyond expectations,” she
says.
Falkow’s wife, Stanford University microbiologist Lucy
Tompkins, “once told me she attributed the massive entry of
women into science to the Vietnam War,” Weiss says. “Drafteligible men could get a deferment
for undergraduate education, but
not a deferment to pursue graduate studies. Suddenly, women
comprised the bulk of the pool of
applicants to postgraduate programs, likely due to challenges to
the idea that women shouldn’t
pursue higher education—and a
reduction in the pool of male applicants.”
Weiss wryly refers to bacterial
pathogens that cause diarrhea as
“‘Southern hemisphere’ agents . . .
that hit you ‘below the belt,’” and
to pathogens that cause respiratory diseases as “‘Northern hemisphere,’ agents.” In Falkow’s lab,
“mostly everyone was working
with a below-the-belt agent,” she
says. “Whether it was psychosomatic or not, everyone seemed to
get a bit ‘loose’ immediately after
they started working in Stanley’s
lab.
“I bucked the trend and worked
with a Northern hemisphere
agent, the respiratory pathogen
Bordetella pertussis,” she says.
This pathogen, which causes pertussis, or whooping cough, makes
a toxin that is “essential for virulence.” For example, it “disrupts
cellular communication mediated
by GTP-binding proteins, and has
a profound influence on the ability to mount an effective immune
response,” she points out. “Currently we are looking at how pertussis toxin alters T cells. While I
haven’t stopped working with B.
pertussis, I feel like my studies
with E. coli O157:H7 are returning me to my ‘Southern hemisphere’ roots.”
Weiss and her husband, an
immunologist, have two teenage
children, a 17-year-old son and a
14-year-old daughter. “My son
has been doing graphic arts for
me; he generated schematics for
this piece, and he made the logo
for the Midwest Microbial Pathogenesis Conference,” she says.
Marlene Cimons
Marlene Cimons is a freelance writer
in Bethesda, Md.
Volume 2, Number 10, 2007 / Microbe Y 491
Some Pathogens Key in on
Complex Carbohydrates on Host
Cells
FIGURE 2
OH
OH
O
OH
OH
HO
OH
O
O
HO
HO
OH
HO
HO
OH
O
HO
HO
HO
HO
HO
O
HO
OH
OH
OH
OH
OH
HO
OH HO
OH
HO
OH
HO
OH
OH
OH HO
O
OH
OH
HO
OH
OH
OH
OH
OH
OH
OH
HO
OH
OH
OH
O
O
HO
O
OH
OH
O
OH
OH
OH
O
O
OH
OH
O
HO
OH
O
HO
HO
O
HO
O
O
OH
O
O
HO
O
OH
OH
O
OH
OH
O
O
O
OH
HO
OH
O
OH
HO
OH
HO
OH
OH
HO
O
O
O
OH
OH
O
OH
O
OH
HO
HO
O
O
O
OH
OH
OH
OH
OH
OH
OH
O
HO
OH
OH
Eleven possible disaccharide combinations using one monosaccharide.
purely protein amino-acid counterparts. One of
the biggest challenges in glycobiology is the lack
of analytic means for delineating the effects of
individual molecular components on downstream signaling mechanisms.
Glycolipids bring in another set of complexities. One or more of the hydrophilic sugars (Fig.
1) can be directly attached to hydrophobic lipids
embedded in the outer leaflet of lipid bilayers in
mammalian membranes. Glycolipids can form
clusters within such membranes, called lipid
rafts. Such rafts consist of glycosphingolipid
microdomains that are enriched in cholesterol.
These domains facilitate protein-protein and
lipid-protein interactions at cell surfaces and
play a critical role in communications between
cells. An important class of glycolipids includes
the gangliosides, which are composed of a glycosphingolipid (oligosaccharide and ceramide)
containing one or more sialic acids residues on
the sugar chain. Gangliosides are abundantly
present on neurons, making them a target for
toxins.
492 Y Microbe / Volume 2, Number 10, 2007
OH
Not only eukaryotes, but also bacteria
make both O-linked and N-linked glycoproteins. For example, Campylobacter jejuni and Campylobacter coli
coat their flagella with acetamidinosubstituted analogs of N-acetyl neuraminic acid, and this coating is required
for cell motility. Even though bacterial
pathogens apparently do not use sugars
to communicate among themselves, microbial cells and their toxins impinge on
the carbohydrates that decorate mammalian cell surfaces—for example,
gaining entry into mammalian cells by
binding to different sugar-containing
receptors. Researchers are beginning to
identify the broad specificities of some
pathogen-carbohydrate
interactions
and toxin-carbohydrate interactions
(see table).
Prominent examples include influenza virus, and bacterial toxins such as
botulinum toxin, Shiga toxin, and pertussis toxin. Each example illustrates
different facets of the biology of sugars
on mammalian host cells and the roles
that they can play in infectious diseases.
Carbohydrates Affect Tissue Tropism,
Virulence, Transmissibility of Influenza
Virus
Understanding the crosstalk between host carbohydrates and pathogens such as influenza virus could prove especially valuable from a public health standpoint. With highly pathogenic
strains of avian influenza now circulating, one
major concern is whether strains will emerge
that can transmit freely and efficiently among
humans. This concern is further amplified because some strains of avian influenza are resistant to available antiviral drugs.
Influenza viruses, including the highly pathogenic avian H5N1 species, carry two glycoproteins on their surfaces— hemagglutinin (HA)
and neuraminidase (NA) that bind to N-acetyl
neuraminic acid (sialic acid) residues on host
cells. These interactions with sialic acid are es-
sential for influenza viruses to enter
FIGURE 3
mammalian cells. Indeed, sialic aciddeficient cells are impervious to this virus.
Influence of viral tropism on transmission and virulence of influenza
Each influenza viral particle possesses approximately 500 copies of HA
HO
COOH
OH
1
HO
trimers and 100 copies of NA tetramers
7
6
8
9
2
on its surface. The HA trimers mediate
O
AcHN
4
5
cell recognition and entry. Viral host
3
OH
O
HO
range strongly depends on the ability of
6
4
O
5
HA to recognize specific structural fea1
HO
tures of sialic acid derivatives and also
2
3
OH
the density and presentation of these
sugar residues on host cells (see table).
For example, influenza C virus infects
animal cells that display 9-O-acetyl
sialic acids on their glycoproteins and
glycolipids, while avian influenza
strains preferentially target sialic acids
linked to the 3 position, and influenza
strains that infect humans preferentially
target sialic acids linked to the 6 positions of galactose or lactose.
This preferential recognition of specific forms of sialic acid has significant
OH
OH
OH
6
implications for viral transmission and
HOOC
OH
4
O
1
HO
5
7
6
pathogenesis. For instance, the upper
8
1
9
O
2
O
2
AcHN
respiratory tract of humans is rich in
3
4
OH
5
3
␣-2,6 sialic acid-linked glycans (Fig. 3,
HO
denoted in blue), whereas cells in the
lower respiratory tract are rich in termiInfluence of viral tropism on transmission and virulence of influenza.
nal ␣-2,3 linkages (Fig. 3, in green).
When influenza infects the upper respiratory tract, high concentrations of viral partiwith birds, and human to human transmission is
cles in respiratory secretions enhance human to
rare. However, mortality is extremely high. The
human transmission via coughing and sneezing.
deadly 1918 influenza pandemic is believed to
In contrast, viral replication in the lower respihave been due to an avian virus that mutated
ratory tract hinders transmission. This greater
from an ␣-2,3 sialic acid binding preference to
efficiency of transmission helps to explain why
␣-2,6, while retaining its high pathogenic poteninfluenza strains that preferentially bind ␣-2,6
tial, making this virus both highly contagious
sialic acids of the upper respiratory tract are
and highly virulent. Whether the currently cirpredominant among humans.
culating strains of avian influenza will mutate to
In contrast, the preference of avian influenza
recognize ␣-2,3 and ␣-2,6 linkages to become
for ␣-2,3 sialic acids suits the H5N1 virus for
highly transmissible remains an open concern.
infecting the intestinal tract of birds, but makes
Understanding these molecular details involvthem not so effective with humans. However,
ing sialic acids is crucial for the development of
when H5N1 viruses reach the lower respiratory
antiviral drugs. Viral neuraminidase cleaves
tract, they tend to cause severe, life-threatening
sialic acid residues from infected cells, releasing
infections, particularly because gas exchange
viral progeny that infect new cells. Influenza
becomes compromised.
drugs, such as Relenza威 and Tamiflu威, are anaThis ␣-2,3 sialic acid binding preference helps
logs that bind and inactivate neuraminidase.
to explain what happens when humans become
However, some avian influenza strains now are
infected with avian influenza. Thus, infection
resistant to Tamiflu威, making it important to
rates are low, usually requiring close contact
pursue alternative means for inhibiting these
OH
OH
Volume 2, Number 10, 2007 / Microbe Y 493
Saccharide specificities for select bacterial and viral
pathogens and toxins
Bacterial pathogen, virus,
or toxin
Escherichia coli
Type 1 pili
P
S
CFA/1
K99
Shigella dysenteriae
Neisseria gonorrheae
Bordetella pertussis
Viruses
Foot and mouth virus
Avian influenza virus
Human influenza
Influenza C
Bacterial toxins
Botulinum neurotoxins
Shiga toxin 1 and 2
Broad saccharide specificity
Man(␣1,3)Man(␣1,6)Man
Gal (␣1,4)gal
NeuAc(␣2,3)Gal(␤1,3)GalNAc
NeuAc(␣2,8)NeuAc
NeuGc(␣2,3) Gal(␤1,4)Glc
GM1, asialo GM1
Gal(␤1,4)GalNAc
Gal(␤1,4)Glc-ceramide
Heparin sulfate
2,3 linked sialic acids
2,6 linked sialic acids
9-O-acetyl sialic acids
Pertussis toxin
Cholera toxin and LT-I
E. coli enterotoxin, LT-IIa
gangliosides GD1a, GD1b, GT1a
Gal (␣1,4) Gal(␤1,4)Glc-ceramide (Gb3) and
analogues
Sialic acid, Gal(␤1,4)Glc-ceramide,
gangliosides
Ganglioside GM1
Gangliosides GD1b, GD1a, GM1
viruses, including by blocking their binding to
sialic acid residues, as potential new drugs.
Bacterial Toxins Such as Botulinum Toxin
Bind Carbohydrate Receptors
Although many bacterial toxins enter cells via
protein receptors, other toxins, including several on the biothreat list such as the botulinum
neurotoxins and Shiga toxin, depend partly or
fully on carbohydrate receptors (see table). Even
though protein receptors confer exquisite selectivity, cells generally express few molecules of a
particular protein, whereas carbohydrate residues typically are abundant. Moreover, the
same glycosylation pattern may decorate many
different proteins on a single cell type, while also
decorating each protein at several sites. Thus,
the extremely high potency of some biothreat
toxins may reflect their binding to high-density
carbohydrate receptors.
The family of seven botulinum neurotoxins
(BoNT) produced by Clostridium botulinum, a
gram-positive, spore-forming bacterium, are
among the most toxic substances found in na-
494 Y Microbe / Volume 2, Number 10, 2007
ture. The A subunit of BoNT is a protease
that cleaves components of the secretory machinery in mammalian nerve cells. Cleavage
then prevents acetylcholine-containing vesicles from fusing with the terminal membranes of motor neurons, leading to flaccid
muscle paralysis and, ultimately, death due
to respiratory failure. One gram of aerosolized BoNT serotype A is enough to cause
more than 1 million deaths. A single lethal
dose of about 1010 molecules, or 10 ng,
happens to be the same as the number of
neurons in a human.
BoNT gains entry into cells via two distinct types of receptors— ganglioside carbohydrate receptors, such as GT1b or GD1a,
and a protein receptor, synaptotagamin. The
ganglioside coreceptors are highly expressed
at nerve terminals, and aid in defining such
cells as targets for this toxin. The gangliosides provide a high-density, low-affinity receptor— enabling high concentrates of
BoNT to build on nerve cell surfaces, increasing the probability that the toxins will bind
synaptotagamin, its low-density, highaffinity receptor.
Synaptotagamin, which is part of the neurotransmitter secretory machinery, further defines the cell as a BoNT target. Synaptotagamin
is incorporated into secretory vesicles and is
expressed transiently on cell surfaces immediately after secretory vesicles fuse at presynaptic
nerve terminals and before endocytosis recycles
membrane components. Molecules of BoNT
bind to gangliosides at the cell surface, then bind
to molecules of synaptotagamin once they become surface exposed. As endocytosis internalizes synaptotagamin, any BoNT that is bound to
those molecules also is internalized.
The seven serotypes (A-G) of botulinum neurotoxin differ in toxicity. Most (52%) human
botulism cases result from BoNT/B, followed by
BoNT/A at 34%, and BoNT/E at 12%. The
other serotypes do not appear to be toxic for
humans. It appears that affinity for the ganglioside receptor dictates susceptibility, and minor
structural differences can determine the difference between life and death following exposure
to these toxins. For example, GT1b and GD1b,
gangliosides that differentially bind BoNT/B,
differ from one another by a single sialic acid
moiety. The differences are even more subtle
between GD1a, which binds BoNT/B with lower
efficiency, and GD1b. GD1b and GD1a
differ only in position of one sialic acid
present in both gangliosides. More detailed information about the effects of
structure on activity might help medicinal chemists develop better inhibitors
and diagnostics for this deadly toxin.
FIGURE 4
A
Site 1
Site 2
B
Site 3
Multivalency of Shiga Toxin
Ensures Affinity for Receptor
Shiga toxin is a major virulence factor
of several types of gram-negative bacteC
D
ria, including Escherichia coli O157:
SA1
S1
H7. Shiga toxin belongs to the AB5
family of toxins (Fig. 4). The A subunit
S2
TYR
is an N-glycosidase, which cleaves sinSA2
SA3
gle adenine residues from 28S riboS3
S2
S3
somal RNA molecules, rendering the
LAC
GAN
ribosome incapable of synthesizing proSA2
LAC
SA3
GAN
teins. Curiously, this same activity is
Structures of two diverse AB5 toxins, Shiga toxin and pertussis toxin. (A,B) Structure of
associated with the select agent from
Shiga toxin 1. Stx1A subunit, orange; and Stx2 B-pentamer, blue. Amino acid regions
castor bean plants, ricin, which preferassociated with binding to the mammalian receptor, Gb3, shown in shades of red; Site
entially binds to both galactose and N1 (red), Site 2 (magenta), and Site 3 (salmon). A, Bottom view of the B-pentamer; B, Side
view. (C,D) Structure of Pertussis Toxin. S1, orange; S2, blue; S3, green, S4, white; and
acetylgalactose. However, the B, or
S5, gray. Amino acid regions associated with binding to mammalian cells are highlighted
binding subunit of Shiga toxin, binds to
in red; SA1-SA3, sialic acid binding sites 1–3; LAC, lactosylceramide binding site; GAN,
terminal sugars on globotriaosylceramganglioside binding site; and TYR, tyrosine-associated binding site. C, Bottom view of the
B-pentamer; D, Side view.
ide (Gb3), delivering the A catalytic
subunit to its cytoplasmic target.
The B subunit of Shiga toxin consists
geting enables B subunits to activate Src family
of five identical polypeptides, each with multiple
kinases within minutes of binding to cells, probinding sites for Gb3 (Fig. 4A). While the bindmoting rapid toxin entry. Furthermore, this siging affinity for individual molecules of Gb3 is
naling may determine the fate of the toxin,
very low, the affinity of the toxin for its receptor
which is not toxic to all cells. For example,
increases more than 1,000-fold when that recepalthough Shiga toxin enters the cytoplasm and
tor is embedded in a membrane in the form of a
damages ribosomes in susceptible cells, toxinlipid raft on mammalian cell surfaces. The sugar
resistant bovine cells shuttle Shiga toxin to lysoresidues in such membranes are tightly clustered
somes to be degraded, allowing cattle to carry E.
and thus can simultaneously engage multiple
coli O157:H7 without developing symptoms.
binding sites on the toxin. The low-affinity interactions are synergistic, amounting to binding
forces that rival antibody affinities in strength.
Furthermore, variations in the structure of the
glycolipid can lead to selectivity between serotypes, and small structural changes in the host
receptor can lead to high specificity between
closely related Shiga toxins.
In targeting glycolipids in lipid rafts, Shiga
toxin gains another advantage. Because lipid
rafts play a critical role in cell-signaling processes and Gb3 associates with lipid rafts, bound
toxin molecules do not depend merely on receptor turnover to carry them into cells. This tar-
Pertussis Toxin: Lectin Activity
and a Second Layer of Toxicity
Pertussis toxin is the major virulence factor of
Bordetella pertussis, which infects humans and
causes whooping cough. Whooping cough has
the dubious distinction of being the only vaccine-preventable disease that is increasing in incidence in the United States. Its toxin is the most
complex bacterial toxin described to date.
Like Shiga toxin, pertussis toxin is an AB5
toxin, with an enzymatically active A subunit
and a 5-member B (binding) subunit (Fig. 4).
Volume 2, Number 10, 2007 / Microbe Y 495
The A subunit catalyzes transfers of the ADPribosyl group of nicotine adenine dinucleotide
(NAD) to regulatory GTP-binding proteins of
mammalian cells, ultimately disrupting G-protein-mediated intracellular signaling. Unlike
Shiga toxin, the B subunits of pertussis toxin are
not identical. Instead the pertussis toxin B-pentamer (Ptx-B) is composed of four different subunits termed S2, S3, S4, and S5 in a 1:1:2:1 ratio
(Fig. 4C-D).
The B subunit not only transports A subunits
into the cytoplasm of target cells, but also possesses a toxicity of its own. Thus, Ptx-B binds to
and promotes signaling through diverse receptors on different classes of mammalian cells,
including glycoprotein 1B expressed by platelets, Mac-1 (CD11b/CD18) expressed on monoblasts, CD14 expressed by macrophages, TLR-4
expressed by dendritic cells, and the T-cell receptor expressed by T cells. This ability to affect
so many different receptors on such a diverse
array of cells suggests that these cells and receptors share some property. In fact, all these receptors are glycosylated, and all of them are sensitive to activation by clustering. Specifically,
Ptx-B activates these receptors by clustering
them, and Ptx-B clusters them by binding to
their sugars.
This activity of Ptx-B is analogous to that of
plant lectins, which contain two or more carbohydrate binding domains and activate cells by
clustering receptors expressing appropriate
binding sugars. For instance, concanavalin A
(ConA) and Phaseolus vulgaris lectin, PHA, activate T-cells of the immune system. While lectins typically possess identical binding sites,
PTX-B has five binding sites that recognize different sugars, including three sialic acid binding
sites, a lactosylceramide-binding site, and a ganglioside-binding site. Ptx-B targets so many different cell types and receptors because it binds
such diverse sugars.
While Ptx-B is a mammalian cell toxin, it also
has limited antiviral activity. It can prevent replication of HIV in peripheral blood leukocytes in
vitro, in activated primary T-cells, monocyte-
derived macrophages, and in ex vivo cultures of
human lymphoid tissues. In vivo, Ptx-B has activity against simian immunodeficiency virus
(SIV) in chronically infected macaques, and
against HIV in SCID mice reconstituted with
human peripheral blood leukocytes. Thus, this
lectin-like activity of pertussis toxin could lead
to development of novel therapies for treating
HIV.
Concluding Remarks
Several novel methods have been developed to
probe the interactions of cell-surface carbohydrates with their receptors. Metabolic engineering allows for the creation and study of cells
deficient in glycosylation or directed to display
unnatural sugar patterns on the cell surface.
Glycomicroarrays have been generated, but are
limited by the lack of pure preparations of complex sugars. Recent advances in synthetic tools
such as enzymatic, programmable one-pot and
solid-phase synthesis have alleviated the bottleneck for some compounds. However, there are
still challenges to be met. Finally, the study of
glycolipids and glycoproteins targeted by microbial pathogens provides yet another research
tool to investigate the basic biology of cellsurface carbohydrates.
The Consortium for Functional Glycomics
(CFG—http://www.functionalglycomics.org/
static/index.shtml) is a large research initiative
composed of more than 300 Participating Investigators and seven scientific core laboratories.
Funded by the National Institute of General
Medical Sciences in 2001 to elucidate the roles
of carbohydrate-protein interactions in cell
communication at the cell surface, the CFG has
made enormous progress towards its goals. The
scientific cores produce a variety of resources
and services for investigators performing experiments that address the goals of the CFG. Data
produced by the scientific cores are accessible
from their website. Specialty databases for glycan-binding proteins, glycan structures, and glycosyltransferases are also available.
ACKNOWLEDGMENTS
This work was supported by NIH/NIAID R01 AI064893 and AI23695 (AAW), and the Center for Biosensors, Department of
Chemistry, University of Cincinnati (SSI), the University of Cincinnati Research Council (AAW and SSI). The authors
gratefully acknowledge the artistic contributions of Peter Monaco, Scott Millen, Michael Flagler, and David Friedmann.
496 Y Microbe / Volume 2, Number 10, 2007
SUGGESTED READING
Katagiri, Y. U., T. Mori, H. Nakajima, C. Katagiri, T. Taguchi, T. Takeda, N. Kiyokawa, and J. Fujimoto. 1999. Activation
of Src family kinase yes induced by Shiga toxin binding to globotriaosyl ceramide (Gb3/CD77) in low density, detergentinsoluble microdomains. J. Biol. Chem. 274:35278 –35282.
Manes, S., G. del Real, and A. C. Martinez. 2003. Pathogens: raft hijackers. Nature Rev. Immunol. 3:557–568.
Paton, J. C., and A. W. Paton. 2006. Shiga toxin “goes retro” in human primary kidney cells. Kidney Int. 70:2049 –2051.
Paulson, J. C., O. Blixt, and B. E. Collins. 2006. Sweet spots in functional glycomics. Nature Chem. Biol. 2:238 –248.
Rizzi, C., M. P. Crippa, R. E. Jeeninga, B. Berkhout, F. Blasi, G. Poli, and M. Alfano. 2006. Pertussis toxin B-oligomer
suppresses IL-6 induced HIV-1 and chemokine expression in chronically infected U1 cells via inhibition of activator protein 1.
J. Immunol. 176:999 –1006.
Rossetto, O., and C. Montecucco. 2007. Peculiar binding of botulinum neurotoxins. A.C.S. Chem. Biol. 2:96 –98.
Schneider, O. D., A. A. Weiss, and W. E. Miller. 2007. Pertussis toxin utilizes proximal components of the T-cell receptor
complex to initiate signal transduction events in T cells. Infect. Immun. 75:760 –765.
Seeberger, P. H., and D. B. Werz. 2007. Synthesis and medical applications of oligosaccharides. Nature 446:1046 –1051.
Shinya, K., M. Ebina, S. Yamada, M. Ono, N. Kasai, and Y. Kawaoka. 2006. Avian flu: influenza virus receptors in the human
airway. Nature 440:435– 436.
Varki, A., R. Cummings, J. Esko, H. Freeze, G. Hart, and J. Marth (ed.). 2002. Essentials of glycobiology, 2nd edition. Cold
Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
Q: When does
4 x $17 = $48?
Postdoctoral
Membership
Recognizing the unique status of
scientists during the postdoctoral
training phase of their careers, ASM
now offers Postdoctoral Membership.
Now postdoctoral trainees can receive all the
privileges* of Full Membership for $37 a year, for 4
consecutive years – two years longer than the
Transitional Member rate offers.
The 2007 membership fee will be $37 ($35 for
those who join or renew online). Any microbiologist who has earned a doctorate within the past
12 months is eligible. The Postdoctoral rate may
not be combined with the Transitional rate to
allow 6 years of reduced rate membership.
Contact Membership Services at service@asmusa.
org for more information, or join online at
estore.asm.org today!
1752 N Street, NW
Washington, DC 20036
http://www.asm.org
[email protected]
*except the right to
vote or hold office
A: When Professors,
Mentors and
University Department
Chairs purchase 4 or more
student memberships for
their
undergraduate and
graduate scholars!
Buy 4 Student Memberships and receive a discount of $5
per membership!
To take advantage of this offer or to get more
information contact Membership Services at
[email protected]. You can also go to the
Membership Menu to get started
http://www.asm.org/Membership/index.asp?bid+363.
1752 N Street, NW
Washington, DC 20036
http://www.asm.org
[email protected]
Volume 2, Number 10, 2007 / Microbe Y 497