3-D Nano Images
IBM researchers have developed a nuclear MRI technique that can see features as small as 90 nanometers.
Understanding the functions of proteins often requires knowing their 3-D structures. But deciphering a protein’s structure is a time-consuming and difficult task, typically requiring crystallizing the proteins and bombarding them with x-rays. What’s more, scientists have not been able to crystallize thousands of proteins, so their structures remain unknown.
A far better option would be an analytic method that allowed biologists to directly determine protein structures. So Dan Rugar, manager of nanoscale studies at IBM’s Almaden Research Center, in San Jose, CA, and his colleagues are developing an imaging technique, based on magnetic resonance imaging (MRI), that allows scientists to see a protein’s structure in three dimensions, just as MRI scans show 3-D snapshots of the human body. Conventional MRI can make out features down to three micrometers. In an April 22 Nature Nanotechnology paper, the IBM researchers demonstrate a resolution of 90 nanometers, a milestone toward their eventual goal of imaging individual protein molecules, which are roughly three to ten nanometers in size. “For the first time, [we’re] moving an MRI imaging technique into the nanoscale,” says Rugar.
To image a protein or other biomolecule in three dimensions, the researchers hope to bring the resolution of the MRI method down to less than one nanometer so that they can pinpoint the location of individual atoms in a protein. Scientists could then reconstruct the protein’s structure. “In our dream, we’d be taking MRI 3-D images of all the atoms in a molecule,” Rugar says.
Magnetic imaging takes advantage of atoms such as hydrogen and fluorine whose nuclei act like tiny magnets because of a property called nuclear spin. In traditional MRI, a strong magnetic field forces the hydrogen nuclei in the body to align along the direction of the magnetic field. Then a coil applies a radio-frequency pulse to the body, which makes a few of the nuclei wobble, creating a voltage that is picked up by the coil and then analyzed by a computer to generate an image. “The difficulty with MRI,” Rugar says, “is that magnetism from nuclei is very weak. The weak signals make it hard to detect a small amount of material, which limits the spatial resolution of MRI.”
The IBM technique, on the other hand, can detect a small sample because it measures magnetic forces instead of measuring voltage. The technique sensitively measures the extremely small attraction and repulsion forces between magnetic nuclei in a sample and an external magnet, says John Mamin, a member of the IBM research team.
The researchers detect the small magnetic forces using a 100-nanometer-wide silicon cantilever. They carve four pillars at the free end of the cantilever, load the pillars with calcium fluoride, and then suspend the cantilever over a small, conical magnetic tip. Then they use a special method to flip the magnetic fluorine nuclei back and forth, which attracts and repels the magnetic tip. The force created as a result of the magnets’ interaction makes the cantilever vibrate. The researchers can measure the vibration with a laser, and this signal gives an indication of where and how much fluorine is around the magnetic tip.
Moving the tip underneath the cantilever, the researchers can scan the whole calcium-fluoride sample. “As you move it along, you get more or less signal depending on the shape and size of the sample,” Mamin says. Finally, the researchers use these signals to reconstruct the sample’s image on a computer.
The image in the paper shows the calcium-fluoride samples on the cantilever pillars and the distance between the pillars with a resolution of 90 nanometers. The volume of the calcium fluoride is 60,000 times smaller than the volume that conventional MRI microscopy can detect.
While the images created so far are two-dimensional images, making 3-D images is a matter of making more scans, Rugar says. The researchers would have to move the magnetic tip up and down to image slices of the sample at different depths, and then simply put the slices together and create a 3-D image.
The method could just as easily be used on molecules containing other atoms with magnetic nuclei, including hydrogen, Rugar says. To achieve their ultimate goal of viewing a protein’s structure in 3-D, the researchers would need to precisely detect the locations of single hydrogen atoms in the protein. For this, the researchers would have to detect the magnetism, or spin, of a single nucleus, a resolution of about 0.1 nanometers. This is a challenge, says Chris Hammel, a physicist who does magnetic resonance research at Ohio State University. But, adds Hammel, the IBM group has made significant strides toward this goal.
The results in the new paper are promising because they show that the imaging technique is robust and that the IBM group’s ideas to improve imaging resolution are working out well, Hammel adds. “Single nuclear spin detection is just an amazing thing to even contemplate, and several years ago it was hard for most people to imagine that it was achievable, but this is starting to seem possible now,” he says. “There’s no indication that it cannot be done. This paper is a significant milestone in that quest.”
Become an MIT Technology Review Insider for in-depth analysis and unparalleled perspective.Subscribe today