IN PREHISTORIC times, when our predators were large and woolly, we fought for
our lives with knives and spearheads fashioned from stone, copper or bronze.
Today, when most of our assailants are disease-causing microbes, pointy weapons
like these are obsolete.
Or are they? Perhaps we just need to reduce the scale and stab our attackers
with nano-weapons. Researchers are now doing exactly that by designing protein
fragments that self-assemble into 鈥渘anotubes鈥� in the cell membranes of bacteria,
poking holes through them that let the microbe鈥檚 insides leak out. The bugs die
in less than a minute鈥攁s fast as the investigators can measure the effect.
The tiny spears have even cured mice of two of the deadliest
antibiotic-resistant pathogens that plague modern hospitals. That suggests a
welcome new option for doctors and pharmaceuticals companies, who are scrambling
to find novel ways to kill microbes that are outpacing our ability to design new
drugs.
As usual, nature perfected the seek-and-stab strategy long ago. Take, for
example, a natural antibiotic called gramicidin, which the bacterium
Bacillus brevis secretes as a weapon against microbial competitors. Two
gramicidin molecules鈥攅ach a string of 15 amino acids鈥攅nter a
microbe鈥檚 membrane, hook up head-to-tail, and coil into a helix like a telephone
cord, with hydrogen bonds gluing each turn of the cord to the next. In effect,
the peptide pair 鈥渄rills鈥� a hole that drains its victim of vital ions.
Advertisement
At least seven families of defence molecules鈥攐ne called,
unsurprisingly, 鈥渄efensins鈥濃�攈ave the same modus operandi. Moths
manufacture an antimicrobial compound called cecropin, bee venom contains a
toxin called melittin, and frogs ooze magainins through their skin. Each
molecule carries a characteristic positive charge, which draws it to the
negatively charged surface of bacterial membranes.
Researchers have already put such defence molecules to use in humans.
Gramicidin, for instance, is an active ingredient in Neosporin and some other
antibiotic ointments. But many of these molecules are incredibly promiscuous,
binding to or inserting themselves into all kinds of membranes, and that makes
them dangerous. A dose just three times that needed to thwart bacteria will
quickly rupture human red blood cells, says Robert Hancock of the University of
British Columbia in Vancouver. So you can daub a gramicidin-based ointment on a
skinned knee, but you can鈥檛 take most defensins internally to battle a bacterial
infection.
In an effort to control which cells get speared, chemists have tried since
the 1970s to construct synthetic versions of these deadly daggers. But they have
all been stymied by technical problems. The protein fragments clumped together,
refused to dissolve in water or otherwise resisted all efforts to force them to
form pores.
But in 1993, an adept organic chemist came at the problem from a different
angle. Reza Ghadiri from the Scripps Research Institute in La Jolla, California,
who was studying how molecules behave in confined spaces, had caught the
nanotechnology bug and was thinking of ways to make protein nanotubes to use as
tiny reaction vessels. He was stumped, however, by the usual problem of how to
synthesise a string of amino acids that would stay dissolved until just the
right time to coil into a tube.
On layover in New York returning from a meeting overseas, Ghadiri, an
architecture buff, decided to visit one of his old haunts from his postdoc days
at Rockefeller University, the Guggenheim Museum. There on its front steps,
gazing at the cylindrical, 鈥渋nverted ziggurat鈥� conceived by Frank Lloyd Wright,
Ghadiri found his inspiration. 鈥淚t occurred to me that instead of making a
helix, I ought to make circles and stack them,鈥� recalls Ghadiri. 鈥淚 said to
myself, 鈥榯his is how we should make a nanotube.'鈥�
When his flight reached the West Coast, Ghadiri made a beeline for his lab,
found his postdoc student Juan Granja and described the idea by stacking slide
carousels on top of each other to model the design: a pile of rings chemically
glued together. 鈥淎t first, I didn鈥檛 believe it would work,鈥� admits Granja, who
is now at the University of Santiago de Compostela in Spain.
Still, buoyed by Ghadiri鈥檚 enthusiasm, Granja was game to try. The two
chemists began with a string of amino acids of two different kinds, each a
mirror-image of the other鈥攕o-called right and left-handed (d and l)
molecules. Naturally occurring amino acids are all of the l variety. The 鈥渉ands鈥�
alternate, right, left, right, left, like people standing in a line gripping
each other鈥檚 hands with fingertips locked together, rock-climber style . In this
set-up, the 鈥渢humbs鈥濃�攃lusters of atoms that will ultimately glue the rings
together鈥攁lternately protrude up and down. If several of these alternating
鈥渉ands鈥� join into a circle, they form a zig-zag-style ring in which thumbs
alternately point up and down.
Such rings, Ghadiri and Granja realised, should spontaneously stack into
perfect cylinders, drawn and held by the hydrogen bonds between a protruding
thumb of one ring and the heel of a hand on the next ring
(see Diagram). The
beauty of the scheme was its versatility: they could broaden or narrow the
diameter of a tube by simply adding or removing amino acids from each ring, and
they could target it to specific sites by carefully choosing the amino acids
used to form the rings.
But Ghadiri did not know at the time that others had tried the approach
before and failed. Their rings had tended to clump together instead of stacking
neatly into tubes. 鈥淲e were lucky,鈥� Ghadiri says. 鈥淚f we knew, we may not have
even begun.鈥�
Ghadiri鈥檚 team faced the same problem: if they engineered their rings with
water-loving amino acids鈥攕o they would dissolve in water鈥攚ater
molecules clustered around them and prevented hydrogen bonds from holding the
rings together into stacks. So Ghadiri tried a clever trick. He built rings
containing some amino acids that carry positive or negative charges at normal
pH, so they would dissolve in water, and some uncharged amino acids
that prefer to dissolve in oily solvents, such as the lipids that make up a
cell鈥檚 membrane. He reasoned that if he got the proportions right, he could
dissolve the rings in water and add them to a suspension of cells, whereupon the
rings would lodge within their membranes. There, having lost their cloud of
water molecules, the rings would self-assemble into perfectly shaped tubes held
tight by hydrogen bonds.
The plan worked. The researchers added the amino-acid rings to test tubes
containing artificial membranes, then checked the resulting structures using a
variety of techniques such as shining beams of polarised light at molecules and
using the pattern of scattered light to determine their orientation. When all
the numbers had been crunched, Ghadiri鈥檚 team knew they had indeed created a
channel through the membranes that could allow the passage of ions and larger
molecules such as glucose.
Next, Ghadiri turned to the problem of building tubes that would lance
pathogens but not human cells. He began with an important group of bacteria
called gram-positive bacteria, whose cell membranes are more negatively charged
than those of mammals. To target the bacterial membranes, the researchers
therefore built their rings using positively charged amino acids along with the
lipid-loving ones.
They tested their strategy on methicillin-resistant Staphylococcus
aureus, a notoriously nasty gram-positive bug that currently defies all but
one antibiotic, vancomycin. When squirted into test tubes loaded with the deadly
microbe, the nanotubes are as effective as vancomycin at killing the bacteria,
says Ghadiri. Perhaps they are even better, because most
antibiotics鈥攊ncluding penicillin and vancomycin鈥攁re bacteriostatic,
merely slowing or stopping bacterial growth long enough for the patient鈥檚 immune
system to gain the upper hand. The nanotubes, on the other hand, actually kill
bacteria within a minute of the rings being added. 鈥淲e find that our peptides
are bactericidal,鈥� says Ghadiri. Human red blood cells, however, do just
fine.
The 鈥渘anobiotic鈥� even works in live mice. In pilot experiments, Ghadiri鈥檚
team injected various doses of nanotube rings into mice infected with S.
aureus, four mice per dose, leaving four other infected mice untreated.
Within about 48 hours, all the untreated animals had died, but the animals given
the nanotubes survived. When the researchers later examined the tissues of the
nanotube-treated animals under the microscope, says Ghadiri, even those given
the highest dose seemed normal. 鈥淥ur studies have shown that we can cure mice of
the infection,鈥� he says. 鈥淎nd there is minimal toxicity with up to five times
the dose that we need.鈥� At last, the goal of a synthetic defensin looks within
reach.
Indeed, Ghadiri may have beaten nature at her own game. Because his nanotube
rings contain both the usual L amino acids and their synthetic d counterparts,
the body鈥檚 protein-digesting enzymes, called proteases, can鈥檛 degrade the rings.
What鈥檚 more, when Ghadiri鈥檚 team added their structures to human
serum鈥攚hich is teeming with natural proteases鈥攖he tubes remained
intact for days, a feat unheard of for most biologically based proteins and
drugs.
鈥淩eza really has pushed the envelope,鈥� says chemist Jeff Moore at the
University of Illinois in Urbana-Champaign. Although chemists have long dreamed
about making nanobiotics, he notes, no one before has come up with a working
product. 鈥淩eza has taken something that was talked about and actually put it
into practice,鈥� he says.
Meanwhile, chemist Sam Gellman鈥檚 team at the University of Wisconsin has had
similar success with a somewhat different approach. Gellman and his colleagues
decided to use helices instead of stacks, fashioning them from so-called beta
amino acids. These carry two carbon atoms in their centres instead of the single
one found in the alpha amino acids of natural proteins. The extra carbon atom
gives beta amino acids more flexibility, so they are easier to contort into
predetermined shapes simply by adding on bulky side groups of atoms. Ring-shaped
side groups tend to force the chain into just the right helical configuration,
Gellman finds. And beta amino acid chains, just like Ghadiri鈥檚 d-and-l
alternating rings, are immune to proteases.
In April, Gellman鈥檚 group took beta amino acid chains carrying an overall
positive charge and induced them to form pores in four strains of gram-positive
bacteria, including two that were resistant to common antibiotics
(Nature, vol 404, p 565). The chains, which often begin coiling even before
reaching the membranes, work as well as natural defensins in test-tube assays.
And, while they killed bacteria, they left human red blood cells untouched.
Infectious-disease specialists鈥攚ho are engaged in a running battle with
bugs that continually outsmart the current antibiotic arsenal鈥攚ill be
delighted if pore-forming compounds make it to pharmacies. 鈥淭he concern about
resistance, especially for certain types of bacteria that are gram-positive,
has reached a feverish pitch,鈥� says John Bartlett, chief of infectious disease
at Johns Hopkins University Medical School in Baltimore, Maryland. 鈥淲ith every
drug, antibiotic resistance is going to occur. It鈥檚 just a matter of when and
飞丑别谤别.鈥�
So the more weapons doctors have, the better. 鈥淲ith these diseases we have to
have multiple approaches,鈥� says Alice Clark, director of the National Center for
Natural Products Research at the University of Mississippi. 鈥淚t鈥檚 not going to
happen with a single drug.鈥�
Nanobiotic peptides may have another advantage in the war against resistance.
Unlike antibiotics such as penicillin that bind and block crucial bacterial
enzymes, nanotubes lodge in the cell membrane. So to develop resistance, the
bacteria would have to change not merely the shape of a single enzyme, but the
structure of their entire membrane.
But there鈥檚 a dark side to nanobiotics. Even Ghadiri, who calls himself
鈥渙ptimistic鈥�, cautions against high hopes at this stage, because researchers
know so little about these nanotubes and how they behave. For example, a team
led by Bill DeGrado at the University of Pennsylvania in Philadelphia made beta
chains similar to Gellman鈥檚 and added them to both bacteria and red blood cells.
The microbes died, but so did the human cells鈥攁 clear warning. No one
knows why DeGrado鈥檚 tubes killed human cells while Gellman鈥檚 didn鈥檛, though
Gellman suggests one reason might be that his beta amino acids had ringed side
chains, while DeGrado鈥檚 did not. The ringed amino acids may stabilise the
molecule鈥檚 helical shape, Gellman thinks, exposing the positive charges that are
so critical for targeting.
Natural defensin-like compounds鈥攎any of which are active ingredients in
bee, wasp and scorpion venom鈥攈ave the potential to damage a patient鈥檚
cells as well as bacteria, says Hancock. It鈥檚 no accident, then, that all the
ongoing clinical trials of defensins are for external use. Add to that the fact
that nanotubes tend to stick around much longer than natural peptides and
toxicity looms as an even greater risk.
So it should come as no surprise that most experts have a wait-and-see
attitude towards nanotubes. 鈥淭hat鈥檚 a beautiful dream,鈥� says organic chemist
Bert Meijer at the Eindhoven University of Technology in the Netherlands. 鈥淏ut I
think we have a number of huge steps to take before they really work,鈥� he
says.
Indeed, even with his apparent success, Ghadiri says his team is focusing on
designing libraries of ringed peptides with various applications rather than
preparing for preclinical trials.
鈥淚鈥檒l leave the later phases to people in the pharmaceutical industry,鈥� he
says. 鈥淭hey have the know-how, and it really takes a lot of people to do this.鈥�
But certainly if researchers do eventually hit their mark with bacteria-stabbing
nanospears, we鈥檒l all be dancing with our ancestors to celebrate the conquest of
yet another dangerous, albeit tiny, beast.