One out of a thousand children in the United States is born deaf; ten percent of all people living in industrialized nations suffer from severe hearing loss — 30 million in the U.S. alone. These are pressing clinical reasons to learn just how hearing works and why it fails.

The hair cells of the inner ear (below) are what make hearing possible.

"Hearing in humans is a remarkable faculty," says Manfred Auer of Berkeley Lab's Life Sciences Division. "It works over six orders of magnitude, from a whisper to the roar of a jet engine. If it were just a little more sensitive, we'd be able to hear the atoms colliding with our eardrums — in other words, our hearing is about as sensitive as we can stand without going crazy."

Hearing is also remarkable for its ability to adapt to constant loud noise yet still manage to pick out barely distinguishable sounds, "like being able to follow a single conversation across the room at a cocktail party, or hearing someone shout at you over the noise of a rock band," says Auer.

And humans can pinpoint the source of a sound to within less than a degree: one ear hears the sound slightly before the other, and the brain calculates the direction from the offset. But the difference in arrival times is less than a millionth of a second, a thousand times faster than most biochemical processes; thus hearing must depend on direct mechanical detection of sounds instantly translated into nerve signals.

The inner ear's hair cells are the key. They convert mechanical responses into electrical signals that trigger adjacent neurons in the brain — a prime example of a phenomenon, fundamental in tissue and cell biology, known as mechanosensation. Hair cells are embedded in the epithelial lining of the cochlea, where they respond mechanically to sound vibrations; others in the nearby vestibular labyrinth move in response to radial and linear acceleration and are the source of the sense of balance.

Bundles of stereocilia bend in response to vibrations in the inner ear. Tension on the tip links (red lines in center diagram) opens channels in the stereocilia to admit potassium and other ions, instantly changing the electrical balance inside the cell and firing nerve cells.

Thus beyond practical concerns lie basic scientific questions about the exact molecular composition and three-dimensional architecture of hair cells and related entities. A uniquely powerful tool for exploring biological structures at this subcellular but supramolecular level is electron microscope tomography — electron tomography for short.

Electron Beams and Zebrafish

"For a long time molecular biologists thought they would be able to understand what goes on in cells and organelles — even whole organisms — if they could understand one protein at a time," says Auer. "But in the late 1990s there was a profound switch in thinking. Researchers realized that most proteins are not floating around inside cells and bumping into each other at random but are organized into molecular machines."
    
Auer compares the cell to an automobile factory whose layout is hidden. "To understand how the factory works you need to know how the machines are put together from smaller parts; at a higher level you need to understand how the machines themselves are organized. So the molecular biologist has to know what proteins look like, how they are assembled into molecular machines, and how the machines work together."
    
Unlike conventional two-dimensional microscopy, electron tomography tilts its target, such as a section of a hair cell, under the microscope's electron beam, yielding a series of projection images of the sample at different angles. From these, a computer constructs a three-dimensional image. 

"Electron tomography bridges the gap between X-ray crystallography and the light microscope," says Auer, who mastered the technique while a member of the laboratory of James Hudspeth at Rockefeller University, before he joined Berkeley Lab in 2004. "X-ray crystallography can reveal the structure of proteins on the nanometer scale, while light microscopy can resolve organelles inside the cell to a couple of tenths of a micron. Most molecular machines fall between those limits. Imaging them with electron tomography is like x-raying that hidden factory from all angles."
      
From Hudspeth, who remains a collaborator, Auer also inherited his interest in the mechanisms of hearing and balance. Hair cells evolved in vertebrates and are similar in animals from fish to humans. Much of the early work on hair cells involved dissecting the inner ears of frogs, "but the frog system was limited by such problems as damage to tissues and sample preservation." Auer now uses a different animal system, wonderfully suited to electron micrography: zebrafish larvae.

	Zebrafish larvae are transparent and their organs are fully formed after three days, helping to make zebrafish an ideal system for electron tomography studies. (Photo (c) 2002 Steve Baskauf)

"The fish are transparent, so it's possible to follow the development of their organs with a light microscope," Auer explains. "The inner ear develops in the first day. All the organs develop in the first three days. They are easy to preserve because the entire larva is tiny and can be frozen instantly under high pressure, so that no ice crystals form to damage the tissue. You don't need prior organ dissection; you can just slice off one organ at a time. In principle, you need only one fish to study every organ!"

It's also relatively easy to introduce foreign DNA into zebrafish. Through this kind of genetic manipulation short stretches of amino acids can be added to specific proteins; fluorescent, nontoxic labels seek out the new sequences and attach themselves to the protein, "so in principle you can localize every tagged protein in any organ or cell," Auer says.

Auer and his colleagues and collaborators have used the bullfrog and zebrafish system to study minute details of hair cell structure and function, in the process uncovering new features — and correcting some misconceptions.

A Hairdo for Hearing

The part of the hair cell that mechanically responds to vibration (or acceleration) is a bundle of fibers called stereocilia, sticking out of the top of the cell like a radical hairdo. In zebrafish the stereocilia are arranged in stair-step fashion. The tallest shaft, made of bundles of cylindrical microtubules, acts like a tent pole to support the development of all the others, which are made of bundles of the protein actin. Each actin-based fiber is shorter than the one next to it, and the tip of each lower fiber is attached diagonally to the side of the adjacent taller fiber by a fine filament called a tip link.

When vibration pushes against the bundle of stereocilia the fibers lean over, stretching the tip-link filaments. This pulls open nearby channels in the fibers (one or two per fiber), allowing potassium ions to flow into the fiber and down to the body of the cell. The electrical balance between calcium and potassium ions in the cell is instantly changed, triggering a signal to adjacent neurons.

If the hair bundle remains bent by persistent noise, a higher level of calcium in the cell signals the structural protein myosin, also present in the stereocilia, to slide down along the actin fibers. By resetting the tension on the tip-link springs in this way, hair cells can adapt to sustained noise levels. 

"There are two ways hearing can be damaged by loud noises," Auer says. "Noise can stress the stereocilia bundle so much that the tip links break. However they usually grow back in 24 hours — this is the rock-concert effect, where hearing loss is temporary. But loud noises can also shear off whole bundles of stereocilia. In mammals these can't regenerate, and the loss is permanent."

Finding a way to regenerate hair cells, says Auer, "is the Holy Grail of research. We're born with just 16,000 hair cells in the cochlea, and every passing subway train kills a few of them." 

Taken individually, the images of stereocilia from which Auer and his colleagues construct electron tomographs don't look much different from the many other microscopic studies of these structures — including blobs near the tips of the fibers that researchers customarily dismissed as "dirt." But, says Auer, "We think there is no such thing as dirt."

The micrograph of the tip link at left is one a series taken at different angles and computer processed to reveal the three-dimensional structure of the tip link, providing valuable clues to the proteins of which it is constituted.

Because electron tomography allows "dissection in silico" Auer's group has been able to analyze these mysterious artifacts, giving rise to provocative hints of unsuspected tip-link structures — including whether there may be more than a single tip link between fibers, how tip links are structured, and what protein or proteins constitute the tip links.

"Until lately, the only protein firmly associated with stereocilia tip structures besides actin was myosin. Now we have 50 candidates — all because we could look at that 'dirt' in 3-D." Auer and his collaborators have developed good evidence for just which proteins are involved in tip-links and in other links among stereocilia. They plan to publish their findings soon. 

Electron tomography studies of hair cells using the zebrafish model also promise to shed light on how this mechanosensory mechanism evolved in the first place. "All cells in tissues can sense whether they are connected to other cells or the extracellular matrix through local contacts, such as the so-called focal adhesion complexes. We are investigating whether hair-cell stereocilia in vertebrates are a modification and evolution of focal adhesion complexes."

And That's Just the Beginning

Auer's investigations with electron tomography reach into many other areas of biological interest, among them:

• the structure of podocytes, highly specialized cells that make up the filtration system of the kidneys;
• how cells in the acini of the breast maintain contact with one another and with the extracellular matrix; loss of contact may contribute to a cell becoming cancerous;
• the essential role of similar contacts in healing wounds and broken bones;
• the fascinating nature of biofilms.

"For years, because they have understandably concentrated on disease organisms, microbiologists ignored the most basic condition of bacterial life, which is that bacteria live in communities," Auer says. Already electron tomography studies have revealed fascinating and unsuspected features of the bacterial communities known as biofilms. Contrary to what most biologists have thought, some biofilms — supposedly made up of independent bacterial cells — have many of the hallmarks of organized tissues.

	Electron tomography of biofilms reveals unsuspected levels of organization among individual bacteria. Resolution is high enough to model distinct, specific ribosomes in the cells (inset).

Indeed, Auer says, "a biofilm is a prokaryotic version of a tissue," and he plans to publish research results soon, demonstrating these similarities in startling detail.

Because electron tomography can bridge the gap between ultrahigh-resolution protein structures and the large-scale organization of cells and tissues available to the light microscope, Auer says, "I would contend that electron tomography will play a major role in investigating all aspects of biology — in structural biology, cell biology, proteomics, biochemistry, physiology, pathology, evolution, everything. Once you have this new toy, you can apply it to all these questions."

Additional information

• Members of Manfred Auer's group include postdoctoral fellow Hildur Palsdottir, research associate Jonathan P. Remis, graduate student W. Jeff Triffo, and Marcin Zemla.
• Auer's collaborators on the hair-cell project include A. James Hudspeth of the Rockefeller University, U. Ziese and A. J. Koster of Utrecht University, Chandrajit Bajaj of the University of Texas at Austin, and Niels Volkmann of the Burnham Institute.
• Auer's collaborators on the Myxococcus biofilms project include Christoph Schaudinn and Bill Costerton of the University of Southern California, and Renate Lux and Wenyuan Shi of the University of California at Los Angeles.
• More about zebrafish

• The Max Planck Institute of Biochemistry in Martinsried, Germany
• The Wadsworth Center in Albany, New York
• Florida State University
• The University of Colorado in Boulder
• The National Center for Microscopy and Imaging Research at the University of California at San Diego
• The University of California at San Francisco
• Lawrence Berkeley National Laboratory
• The Campus de la Universidad Autónoma de Madrid
• And see "Electron Tomography of Biological Specimens," by Roberto Marabini, M.Carmen San Martín, and Jose M. Carazo, published in Contemporary Perspectives in Three-Dimensional Biomedical Imaging