Richard Blakemore was a graduate student in microbiology, looking for Spirochaeta plicatilis in marine marsh mud, when he first noticed an oddity. After being stored on the lab shelf for a while, the sulfide-rich muds contained large numbers of fast-moving bacteria that moved across the microscope slide from south to north. Normal bacteria bumble around randomly -- why had so many apparently decided to go north for the winter? And how were they doing it?

There was a window on the northwest wall of the lab; perhaps these bacteria were responding to the light? But shining light on different sides of the microscope didn't make any difference; Neither did putting a box over the microscope, nor turning the microscope around, nor moving the microscope to another room. Each tiny cell, deaf, dumb, and blind, kept doggedly swimming, always north.

Then Blakemore brought a magnet near the microscope. Hundreds of swimming cells promptly swerved away from it in unison. They acted like tiny, self-propelled compass needles, aligning themselves to the local magnetic field and directing their swimming by it. Blakemore was astonished, but he knew an interesting phenomenon when he saw one. His paper in Science (Blakemore, 1975) initiated a field of study that has drawn in biologists, chemists, geologists, and physicists, and that has implications in fields as widely ranging as biochemical synthesis and the planetary history of Mars.

Blakemore examined some of these bacteria under a transmission electron microscope and found that each cell contained dense particles lined up in two chains of between five and ten particles each, something that hadn't been seen in other types of bacteria. He saw that the particles in the cells often seemed to be surrounded by some type of membrane, and proposed that the membranes might be producing the particles somehow. Similar particles could be found in the mud just below the level where these bacteria had been living. Their tendency to clump suggested they were made of some naturally magnetic substance such as magnetite (also known as lodestone -- the mineral from which the earliest human compasses were made). Blakemore proposed that these particles had been released when some of the bacteria died.

Analysis of the x-rays given off when these dense particles were bombarded with electrons showed that the particles contained a lot of iron, which would be expected of magnetite. Magnetite contains iron and oxygen: 3 atoms of iron for every 4 atoms of oxygen, a chemical combination written as Fe₃O₄. If the particles in the cells were magnetite, and hence magnetic, they would tend to align with the local magnetic field. If the chains of particles were fixed in the cell, so that the cell had to turn with them, the whole cell would align with the local magnetic field, producing the odd behavior that Blakemore had first seen in the light microscope.

It turned out that Blakemore was precisely right.

Single domain magnetism

The dense particles that Blakemore saw do indeed contain a magnetic mineral rich in iron. Furthermore, these magnetic crystals are very unusual.

A normal non-bacterial crystal of magnetite, for instance, just grows every which way as it precipitates out of water or condenses from a vapor phase. It makes crystals that are interlaced with each other as often as separate (single) crystals, and it makes crystals of many sizes.

A crystal of the mineral grows as dissolved molecules of Fe₃O₄ attach to its surface. The attachment of each dissolved molecule is governed by two factors. The first is a magnetic force that tries to align the magnetic moment of the dissolved molecule with the magnetic moment of the crystal. Thus the first factor tends to make the magnetic moment of the crystal larger as the crystal grows. But the magnetic moment of the crystal produces a magnetic field around the crystal -- and the larger the magnetic moment, the more energy in its magnetic field. The second factor is a tendency to reduce the energy in that magnetic field by allowing dissolved molecules to attach with their magnetic moments aligned opposite to the magnetic moment of the crystal, which tends to reduce the magnetic moment of the crystal.

This is complicated by the fact that no molecule wants to be aligned very differently from the molecules around it. These factors combine so that solid magnetic materials contain areas (called magnetic domains) where the molecules are all aligned with each other, surrounded by boundaries (called domain walls) where the alignments of the molecules are gradually changing to match the alignment of the neighboring domain. The magnetic moments of the domains tend to point opposite to each other, which is why most iron doesn't act like a magnet. Domain walls have to have a certain thickness -- in magnetite that thickness is about 150 nm (150 nanometers, or 150 billionths of a meter). Particles of magnetic material that are too small to hold a domain wall have to be a single magnetic domain.

On the other hand, particles that are really small don't have a permanent magnetization of their own -- their magnetic moment is constantly being altered by the thermal motion of their atoms (the slight motion, even in solids, of every atom as a result of its temperature). Magnetotactic bacteria synthesize magnetic particles between 30 and 100 nm in every dimension, big enough to have a permanent magnetic moment, but small enough to be a single domain. This makes the magnetic particles as strong as they can possibly be for their size.

Two flavors of magnets

Though most magnetotactic bacteria that have been studied so far make their magnets of magnetite, a few use the mineral greigite, which is similar to magnetite except that it has sulfur in place of oxygen, so its chemical composition is written Fe₃S₄. Greigite is magnetic, but only a third as magnetic as magnetite.

One unusual magnetotactic bacterium makes both magnetite and greigite magnetosomes. Bazylinski and others found it in the Pettaquamscutt River estuary. Each particle contains only one type of mineral, but both types of particles are arranged in each of its two chains. This 3-micron, cigar-shaped, slow-moving bacterium appears to be unique in producing both kinds of magnetosomes, and since it is found in both microaerophilic and anaerobic environments, the type of magnetic mineral it produces may depend on how much oxygen is in its environment at the time.

The magnetosomal membrane -- more than just packaging

Blakemore observed that the dense particles he saw in the magnetotactic cells seemed to be surrounded by a membrane. This is now called the magnetosomal membrane, and there is a great deal of interest in how it works.

Cells often use membrane compartments to synthesize things. Using a membrane to "wall off" part of its space lets the cell control the concentration of various chemicals in the compartment by using protein "doors" that pump certain molecules in and don't let them out again. Cells can also control the acidity of these membrane compartments, as well as the tendency of compounds inside to accept or give up electrons to other compounds. So the idea that magnetotactic bacteria use membrane-bounded spaces to construct magnetosomes is a perfectly reasonable one.

In addition, magnetosomes are often an unusual shape. Natural crystals of magnetite and greigite tend to be roughly round (cubo-octahedral if you want to get technical), whereas magnetosomes are often anisotropic (not regularly shaped) -- and may be elongated, or even tooth- or bullet-shaped, with their long axes along the axis of the chain, indicating that one face of the crystal has grown faster than the others.

In addition, magnetosomes are unusually free of structural defects -- areas where the layers of material added to the developing crystal didn't line up quite right and grew together any old way. This kind of structural defect occurs more often in inorganic, rapidly growing crystals.

Stephen Mann, along with Richard Frankel and Blakemore, used high-resolution transmission electron microscopy to look at the small, irregular particles sometimes found at the ends of chains of magnetosomes. They found that a single membrane-bounded space contained magnetite on one side and an amorphous (literally "formless" but in this context "lacking order") iron-containing substance on the other. The magnetite appeared to be growing out into the amorphous region, presumably by converting the "pre-magnetite" precursor, thought to be ferrihydrite, a dissolved iron oxide, into magnetite at the surface of the growing crystal. They suggested that a patch inside the magnetosomal membrane had nucleated (started) magnetite crystallization, specifying the orientation of the crystal within the magnetosome. By controlling the direction from which more ferrihydrite could enter the magnetosome, the magnetosomal membrane could control the direction of growth, and thus the shape, of the magnetite crystals. Nucleation might be controlled by special membrane proteins that bind magnetite, organized with a particular spacing on one part of the membrane. Holding a starting layer of magnetite together at one side of the magnetosome would provide a surface for more magnetite to crystallize onto. In addition, the magnetosomal membrane probably controls the final size of the crystals (presumably by stopping the entry of ferrihydrite when the crystal is "done").

Constructing a chain

Many types of magnetotactic bacteria have been observed, of different shapes and sizes, with differently shaped magnetosomes. Some have one chain of magnetosomes, some have two or more, and a few don't seem to have chains at all.

Chains of magnetosomes are probably held together partly by the magnetic interactions of the magnetosomes themselves. Each is positioned so that its north pole points at its neighbor's south pole all along the line, so that they are drawn together by their own magnetic fields. The magnetosomal membrane, and possibly other chain components, separate each magnetosome slightly from its neighbors and provides a little elastic "cushioning," helping make the chain more stable in the face of slight bends that might occur as cells move about.

It looks as though magnetosomal membranes may also have something to do with these chains, even before magnetosomes are made. In one magnetotactic bacterium, grown for some time in the absence of iron, Yuri Gorby, along with Terry Beveridge and Blakemore, saw empty membrane-bounded spaces -- the right size and shape to be empty magnetosomal membranes -- at the end of the magnetosome chain. This suggests that magnetosomal membranes form (and grow new magnetosomes) already aligned with the end of the existing magnetosome chains.

Fixing a chain in a cell

In cells with only one chain of magnetosomes there are a couple of possibilities. There are many types of magnetotactic bacteria, with many chain arrangements. In some cases the cell is long and thin, with the magnetosome chain along its long axis, and the cell is too narrow to allow the chain to turn freely without hitting the walls. A magnetosome chain like this may not need to be fixed. In some cases the cell isn't narrow enough to constrain the chain. In these cases it's suspected that some kind of interior filament in the cell holds the magnetosomes in place, possibly by attaching to the magnetosomal membranes.

In some cases there is more than one chain of magnetosomes. In cases like these the chains of magnetosomes will repel each other, like two aligned bar magnets placed side by side. Calculations of the forces involved show the chains will press against the sides of the cell, holding each other in place by friction, so that when they orient in an external magnetic field, so does the cell. Cells have been observed with up to five chains of magnetosomes -- at this point the chains of magnetosomes may be acting like boning in a bodice, stiffening the wall of the cell and holding the cell "open."

How Swimming North Helps Them

Bacteria are simple creatures. If they have a behavioral trait, like following magnetic field lines, it's probably a survival advantage. But what advantage is there to a bacterium in swimming north? The answer seems to be that they swim north to get away from a toxic substance.

Earth's magnetic field has a downward component everywhere except on the equator. Swimming north doesn't matter to a bacterium, but swimming down may be quite another matter. A bacterium's world is one where gravity has little effect, and being able to follow magnetic field lines "reduces a three-dimensional search problem to a one-dimensional search problem" as Richard Frankel put it. Environments such as ocean marshes and soft mud lake bottoms tend to be disturbed now and again, by wading children or the rising tide. Magnetotaxis is a simple way for the bacterium to find its way home.

Anaerobes -- oxygen is poison!

We think of oxygen as vital to life, and it is, for us. But oxygen is also a very harsh, corrosive element, and living things have to be specially adapted to undo the harm it causes in order to be able to tolerate it for any length of time, much less to use it in metabolism. Many types of bacteria are not adapted to oxygen. Such bacteria would have been much more common when Earth was much younger, before some clever cell learned the trick of using light to rip carbon dioxide apart and eat it. Even now, when the oxygen by-product of photosynthesis has thoroughly polluted the atmosphere, many bacteria find places to hide away from it, in deep water or mud or under the soil. Such bacteria are called anaerobes. The magnetotactic bacteria that produce greigite tend to be anaerobes, though there are a few examples of anaerobes that produce magnetite.

Microaerobes -- fresh air is too much of a good thing

Some bacteria are adapted to use a little bit of oxygen, but only about a twentieth as much as is found in air; they are called microaerobes. Most of the bacteria that produce magnetite are microaerobes, though the oxygen in their magnetite has been shown to come from water, not from air. Microaerobes live in the same sorts of environments anaerobes do, but usually a bit closer to the surface and the air.

Magnetotactic bacteria don't have the same kind of run-and-tumble motion as E. coli. Since they're oriented to the magnetic field, they can't really tumble very well anyway. Instead, when their flagellar motors reverse direction, the cells also reverse direction and travel the other way along the magnetic field line.

There are two ways magnetotactic bacteria sense and respond to changes in oxygen. One is to compare the way things are with the way they were a few seconds ago. If things are getting better -- i.e., if the cell is moving toward its desired concentration of oxygen -- it keeps going in the same direction. If things are getting worse, the cell changes direction. This is called axial magnetoaerotaxis. The second way is also very simple, but involves quite a different mechanism. The cell compares the oxygen it's experiencing to the oxygen it wants. If there's too much oxygen, it swims north, no matter which direction it was going before. If there's too little oxygen, it swims south. This is called polar magnetoaerotaxis.

And in Australia they stand on their heads

As you can imagine, swimming north to get to areas of low oxygen concentration would pretty much finish off any magnetotactic bacterium living south of the equator. If magnetotactic bacteria benefit from magnetotaxis because it keeps them away from oxygen, then one would expect magnetotactic bacteria found in the southern hemisphere to swim south in order to swim down. In 1981 Blakemore and Frankel went to Tasmania and New Zealand and found that sediments there contained magnetotactic bacteria that swam south, as expected.

On a stranger note, sediments from the equator also contain magnetotactic bacteria. It's a bit harder to imagine why magnetotaxis would be an advantage at the equator, where the magnetic field is horizontal. Frankel has suggested that following magnetic field lines at the equator would at least prevent bacteria from wandering up toward high oxygen concentrations by mistake. However, if the organisms are strongly held to the magnetic field lines, a bacterium that swirled off the bottom mud during a disturbance would actually be less able to find its way back down to the mud than a simple aerotactic bacterium that used the run-and-tumble motion we see in E. coli. But the organisms may not be as strongly held to the field lines as they are in temperate latitudes; the magnetic field at the equator is only 1/2 to 1/3 what it is at the poles.

And Their Sisters and Their Cousins and Their Aunts

A large number of types of magnetotactic bacteria have been found. In his first paper on the subject Blakemore noted that he'd seen at least five kinds in the mud from around Wood's Hole, Massachusetts. Unfortunately, partly because of their microaerophilic or anaerobic metabolisms, magnetotactic bacteria are very difficult to grow in pure culture (i.e., a lab flask that contains bacteria that all descended from a single bacterium -- so that you know they're all the same species). Microbiologists refer to them as "fastidious" -- meaning they're very picky about their living conditions.

Since they're so hard to grow in pure culture, it's very difficult to compare a particular bacterium to all others to be sure it's a new species. Part of the reason for this is that different bacteria may look alike, and can only be distinguished by their metabolism: what they will use for a "carbon source" (which is to say what they will eat) and what they will use as an "electron acceptor" (which is to say what they will "breathe" -- a substance bacteria use the way we use oxygen), and whether there are other things they must get from the environment, the way we must get certain amino acids and vitamins. It's hard to know much about the dietary requirements of a wild bacterium when you have no control over its diet.

Mug shots

A few types of magnetotactic bacteria have been grown in the laboratory. There is Magnetospirillum magnetotacticum, the first magnetotactic bacteria to be cultured. (It is referred to as Aquaspirillum magnetotacticum in older papers.) It is also referred to as "strain MS-1." A closely related species, Magnetospirillum gryphiswaldense, was described in 1991.

There is also a strain of anaerobic sulfate reducing bacteria called RS-1 that produces magnetite magnetosomes.

Bacteria that have been observed in the wild but not cultured include Magnetotacticum bavaricum, a large (8-10 microns long) slow-moving (40 microns per second) rod-shaped bacterium with five chains, of about 200 particles of magnetite each, that run the length of the cell. This is far more particles than other magnetotactic cells have, but they are needed to provide enough torque to efficiently orient this very large cell. This bacterium was found in the sediments of Lake Chiemsee in Upper Bavaria, Germany.

Greigite producing bacteria include a widespread "multicelled magnetic procaryote" referred to as MMP, which is found as clusters of about 20 cells. The MMP cell cluster aligns to the local magnetic field and swims along field lines, but attempts to pull it apart into single cells result in non-swimming (probably dead) cells.

Another greigite producing bacterium has been observed in samples from the Sweet Springs Nature Preserve in Morro Bay, California -- this is a rod about 8 microns long with two chains of magnetosomes running the length of the cell.

Kith and Kinship

Edward F. Delong, along with Frankel and Bazylinski, showed that the magnetite producing magnetotactic cocci (round bacteria) were closely related, and that the vibriod magnetotactic strain MV-1 was related to Magnetospirillum magnetotacticum and Magnetospirillum gryphiswaldense, but that the odd multicelled magnetotactic procaryote called MMP (which produced greigite crystals) was more closely related to the sulfate-reducing bacteria (bacteria that are anaerobes, and use sulfate to "breathe") than to any other magnetotactic bacterium they studied. They suggested on this basis that the type of magnetic mineral produced (magnetite vs. greigite) depended on the phylogenetic affiliation of the bacteria, and that the two types of magnetosomes may have separate evolutionary origins.

Unfortunately, DeLong examined these issues seven years ago, and some of the more intriguing bacterial strains (like the Pettaquamscutt Estuary bacterium that produces both magnetite and greigite, and RS-1, the anaerobe that produces magnetite, and Magnetotacticum bavaricum) were not included in the study.

Their Mortal Remains

Back in the first paper on the subject, Blakemore described how particles resembling the bacterial magnetosomes appeared in the mud below where the magnetotactic bacteria lived. He proposed the magnetosomes were released from a cell when it broke apart after death, and built up in the sediment below. Since then, people have gone looking for these "magnetofossils," partly because they might provide information about how long magnetotaxis has existed, and partly because the natural magnetism of soil and rock might be due, to some degree, to the buildup of bacterially generated magnetic particles.

Nikolai Petersen and Tilo von Dobeneck sifted through sediments from the bottom of the South Atlantic ocean and found that bacterial magnetite was the main source of natural magnetism in sediments dating back 50 million years. They washed magnetite of the right size range (less than 100 nm -- see the discussion of single domain magnetism above) out of the sediment cores and showed that most of the magnetite particles were perfect crystals, or other unusual shapes that had previously been observed in magnetosomes from magnetotactic bacteria. Some of them were even found in chains like the chains of magnetosomes often observed in bacteria.

Petersen later brought together another group, which compared fresh and fossil magnetosomes and noted that in stable marine environments one shape of fossil magnetosome dominated in a given sample, while in less stable environments several types of magnetosomes coexisted. Since most known strains of magnetotactic bacteria tend to produce only one shape of magnetosome, this tends to suggest more species of magnetotactic bacteria inhabit the less stable environments, which would be consistent with known theories of species diversity in other ecological niches.

Magnetosomes are somewhat sensitive to corrosion under some conditions, so failing to find fossil magnetosomes is like failing to find any other fossil -- it doesn't tell you whether or not the animal existed in the area. So far, all the magnetofossils described have been magnetite.

The Microbe from Mars?

In 1996, David S. McKay and 8 other researchers made a big splash with an article in Science about a meteorite called ALH84001, believed to have been knocked off the surface of Mars 16 million years ago by the impact of an asteroid. It landed in the Allen Hills of Antarctica (hence the ALH in the name) about 13,000 years ago and was found in 1984 (hence the 84 in the name -- the 001 means it was the first meteorite found in 1984).

The 1996 article pointed out that the rock contained 1) polycyclic aromatic hydrocarbons (a kind of tar-like substance that can be formed by biological processes but is also known to form in space, completely inorganically), 2) carbonate globules from 1 to 250 microns in diameter found along fractures and in pores in the rock which looked similar to formations that result from bacterial activity in fresh water ponds on earth, 3) small ovals near the carbonate globules which were proposed to be fossilized Martian "nanobacteria" (though the proposed nanobacteria were only 100 microns long -- that is, as long as one or two magnetosomes) and 4) single domain magnetite crystals with no structural defects, similar to those found in terrestrial magnetotactic bacteria. Response to the article was mixed. Any of the "biomarkers" mentioned could be produced by inorganic processes as well as biological ones. John P. Bradley, Ralph P. Harvey, and Marry Y. McSween, Jr. wrote an article that also came out in 1996, which pointed out that much of the magnetite in ALH84001 was in the form of whiskers and plates -- forms associated with deposition from the vapor phase (similar to the way frost crystals grow from water vapor, only much hotter). Many of the whiskers had a central spiral dislocation -- a type of structural defect that produces rapid growth in one dimension, giving rise to the whisker form -- that has not been seen to date in magnetosomes.

In late February of this year the debate heated up again. Kathie L. Thomas-Keprta, David S. McKay, and 8 other researchers brought out a paper in Proceedings of the National Academy of Sciences describing the subset of magnetite crystals in ALH84001 that they thought resembled magnetosomes. They pointed out that about a quarter of the magnetite in ALH84001 was in the single domain size range and had a shape similar to, but not quite the same as, the "truncated hexa-octahedral" shape of the magnetite produced by MV-1. In addition, the single-domain magnetite in ALH84001 was chemically pure (consisting of iron and oxygen only, no cobalt or copper or zinc), like most bacterial magnetite. They didn't comment on the chemical purity of the rest of the magnetite (the whiskers and plates) in the sample. The single-domain magnetite they examined in ALH84001 had few structural defects. Magnetosomal magnetite also has few structural defects. The single domain magnetite they observed in ALH84001 was not organized into chains. Some fossil magnetosomes have been found in chains in terrestrial samples, but not to the degree that they are organized in chains in many magnetotactic bacteria. And some fossil magnetosomes aren't found in chains at all. On the basis of these observations they suggest that a quarter of the magnetite in the meteorite shows signs of being produced by magnetotactic bacteria.

One question that pops to mind is: what good would magnetosomes do a bacterium on Mars? There's reason to think that Mars was warmer and wetter once, so there may have been water for them to live in, but there's no oxygen to speak of to swim away from -- the Martian atmosphere is mostly CO₂. Unfortunately, Mars has no magnetic field to speak of -- its present-day magnetic field is about 1/800th of Earth's. A magnetotactic bacterium would need a lot of magnetite to be oriented by that. Nevertheless, interest in the subject remains strong.

Cat Faber lives in Oregon, where she writes songs, performs (as part of the duo Echo's Children), and is looking for work as a science writer. Her previous appearance in Strange Horizons was "Plastic That Comes Alive."

References

Further Reading:

R. Blakemore (1975). "Magnetotactic Bacteria." Science 190: 377-379.

R.P. Blakemore and R.B. Frankel (1981). "Magnetic Navigation in Bacteria." Scientific American 245: 58-65.

Richard A. Kerr (2000). "GEOLOGICAL SOCIETY OF AMERICA MEETING: Geologists Pursue Solar System's Oldest Relics." Science 12-22-2000.

R.P. Williams (1990). "Iron and the origin of life." Nature 343: 213-214.

Other References