To see something that no one has ever seen before is one of science's most fulfilling rewards. It's a home run, a trophy, a gift of knowledge. It's a CV highlight, a string of research papers, a story to tell over and over.
Late last year, a pair of then-unemployed Bell Labs expatriates, Eric Betzig and Harald Hess, shipped a prototype microscope they had designed, financed in part with $50,000 of their own money, and built in Hess's living room to collaborators at the National Institute of Child Health & Human Development (NICHD), in Bethesda, Md. They all knew from the first test runs in January that they had in their hands a new window on the biological world.
With PALM, the soothing acronym for photoactivated localization microscopy, Betzig, Hess, and collaborators have been able to discern the precise cellular locations of thousands of individual protein molecules. With conventional fluorescence microscopy techniques, particularly at high magnifications, these molecular multitudes appear as smeared fields of luminous blur. But PALM focuses those blurs, which have served biologists mightily despite their lack of clarity, into subcellular "star fields," each point of light due to the here-I-am beacon of a single cleverly labeled protein.
"It's amazing," says Jennifer Lippincott-Schwartz, cell biologist at NICHD, who, with labmate George Patterson, designed the fluorescent "optical highlighters," a type of photoactivatable light-emitting tag for cellular studies that Betzig and Hess needed to make PALM work. Lippincott-Schwartz is especially excited about combining PALM, which produces images that identify the locations of specific individual molecules in a sample, with electron microscopy (EM), which gives images with fine structural detail but usually without clearly specifying the molecules involved in the structures.
"With PALM EM, we will be able to identify not only structure at high resolution, but also molecular identity," she says. "It is totally an exciting time to be in the imaging field."
Biophysicist Xiaowei Zhuang of Harvard University seconds that sentiment. This month, she and colleagues unveiled their own so-called super-resolution method, which she says is based on the same concept as PALM. Called STORM, or stochastic optical reconstruction microscopy, it too produces super-resolution images by localizing light-emitting fluorophores in each of many excitation cycles. In each cycle, only a small number of photoswitchable fluorophores emit light, but collectively these cycles enable the researchers to achieve an overall image with 20-nm resolution or less (Nat. Methods, DOI: 10.1038/nmeth929). Using a photoswitchable fluorophore that they discovered two years ago, the researchers have used STORM, for example, to produce structurally revealing images of small DNA molecules studded with fluorescent tags. Since the fluorophores they use can be activated and deactivated hundreds of times, Zhuang notes, the system should be useful for tracking processes on nearly molecular scales over time.
In the STORM imaging process, some 20 photoswitchable fluorophores within a single diffraction-limited spot now can be resolved in just a matter of minutes. "This can be pushed to seconds," Zhuang predicts. "STORM has the potential to transform biomolecular and cellular imaging." A standout application for her, she says, will be to investigate how viral particles invade cells.
Adding to the buzz is that these techniques are among a wave of innovations that are poised to usher optical microscopy into a new era of finer vision that no one had suspected possible. The once seemingly ineluctable limit on the resolving power of optical microscopes is proving to have been more a state of mind than an unbeatable constraint founded on theory.
To their own surprise and delight, the techniques' inventors and collaborators already have used their tools to help settle some long-standing questions in molecular biology. In each case, a crisper look was mostly what was needed to adjudicate among different models of how particular protein systems work. "It is kind of breathtaking to see these optical images," says Zhuang, who, like Betzig and Hess, is a Howard Hughes Medical Institute (HHMI) investigator. "Seeing things that are even smaller and finer than was possible before is truly exciting."
"I personally believe this is going to transform cell biology," adds Stefan W. Hell of Max Planck Institute for Biophysical Chemistry, Göttingen, Germany, who has been taking a very different tack to super-resolution microscopy. Once such super-resolution tools get into the hands of the many instead of just the few who have home-built them, Hell and others say, the effect on biology should multiply.
Even a decade ago, almost any scientist would have waved off talk of an optical microscope capable of resolving molecules as nonsense. For generations of scientists, it had been common knowledge that you just can't use an optical microscope, the kind with glass lenses and objectives, to visualize an object that is even a few times smaller than the 300- to 700-nm wavelengths of visible light, let alone tens of times smaller or more. It's been taken for granted that for visualizing viruses, proteins, and other biological molecules, even as vague spots or blips, it would take electrons or X-rays, which have shorter wavelengths, or entirely different imaging strategies, such as the almost tactile approach to "seeing" employed by scanning-probe microscopes. Trouble is, unlike optical microscopy, those techniques entail killing life to see it, and that's not always the way to go for biology studies.
Way back in 1873, the bespectacled German mathematician and physicist Ernst Abbe-whose lens designs were then helping the young Jena-based Zeiss Optical Works earn a top-notch global reputation-proclaimed with what seemed like bulletproof physics that the way light diffracts on small spatial scales means it cannot be used to resolve anything too much smaller than a micrometer. Known in microscopy circles as the diffraction limit or the Abbe limit, this means that shining visible light on a protein spanning just a few nanometers would be about as revealing as probing the surface texture of a poppy seed with your finger. The best microscopists could hope for, Abbe's calculations showed, was an instrument capable of resolving two tiny objects about 400 nm apart. Objects closer than that would appear as a single blob. And indeed they have for almost anyone who has tried to push optical microscopy to those edges of resolution.
"When I was a student, I always thought the Abbe barrier would be there," Hell says, referring to his graduate days during the 1980s at the University of Heidelberg, in Germany. Hell later began to question Abbe's sobering bit of optics gospel and started to steadily publish papers demonstrating super-resolution, that is, resolution beyond the diffraction limit, with stimulated emission depletion (STED) microscopy.
STED employs a pair of overlapping laser beams: a regular spot to excite the fluorescent tag on the biomolecule to be imaged and a doughnut-shaped beam to quell fluorescence around that spot. By tuning short pulses of the doughnut to the tag's fluorescence wavelength, the tags in the annular illumination zone are stimulated to go dark, leaving a doughnut-hole region of excitation. The more the power in the doughnut zone is cranked up, the narrower the hole becomes, Hell says. "This renders continually sharper spots with the size of a molecule being the conceptual limit," he adds. By scanning these illumination beams over, say, part of a cell or organelle, an image can be built up from what effectively are sub-diffraction-scale pixels, a resolution not possible with conventional fluorescence microscopy. Last year, the researchers reported a fluorescent spot of 16 nm, or about one-fiftieth the wavelength of the excitation light (Phys. Rev. Lett. 2005, 94, 143903).
Among other things, Hell and his colleagues have used STED microscopy to, in the wry words of Purdue University chemist Garth J. Simpson, "help settle a question that has been exercising cell biologists for some time." That question has to do with how the protein components of the roughly 40-nm-diameter synaptic vesicles, which are brain cell structures that store and release neurotransmitter molecules after fusing with the brain cell's membrane, are internalized by the cell and then recycled and refilled with new batches of neurotransmitter. "Do their components diffuse on the plasma membrane, or do they remain together?" the researchers asked in one of their papers (Nature 2006, 440, 935).
To get an answer, the researchers labeled a synaptic vesicle membrane protein with antibodies carrying fluorescent tags. Then they trained their STED microscope onto the preparation. Whereas conventional fluorescence microscopy shows only blobs of light with this same standard preparation, STED images showed clusters of bright spots, a strong indication that at least some vesicle components stay together in the initial stages of recycling.
Now Hell is pushing the super-resolution of STED further by playing around with the pulse rates of the illumination lasers in a way that closes down some of the energy-releasing pathways available to the fluorescence tags. The pathways to be shut down begin with an excitation of these dyes' so-called triplet states, which refers to a specific quantum mode of electronic excitation. If unchecked, these triplet-state excitations end up hastening the demise, or photobleaching, of the dye molecules.
Hell's souped-up version of STED, which the researchers refer to as T-rex (for triplet relaxation) STED, results in a roughly 3,000% increase in the tag's fluorescence signal. This increase in intensity brings with it the possibility of reducing the focal spot area almost 150 times smaller than Abbe would have predicted was possible. Hell and eight colleagues in Göttingen recently used the technique to obtain several macromolecular-scale images of individual tagged proteins in synapses and cell nuclei, suggesting a resolution of about 20 nm (Proc. Natl. Acad. Sci. USA 2006, 103, 11440).
Hell describes STED as a top-down approach to super-resolution. It works by shrinking as much as possible the diffraction-limited focal spots of one laser using the fluorescence-depletion effect of another laser, he says. Betzig, Zhuang, and Hess are taking what might be considered a bottom-up approach. Their tack is to build up super-resolution images by activating small subsets of fluorescently tagged proteins in each one of thousands of illumination cycles. During each cycle, a computer calculates the few fluorescing molecules' precise positions from dim signals impinging on the instrument's sensitive charge-coupled device light detector. All of these localizations ultimately are combined to generate a full image showing the distribution of tagged molecules with a spatial resolution measured in nanometers. At the moment, it is a time-consuming process, lasting for hours.
Last month, Betzig and colleagues unveiled PALM in a Science paper with images showing pointillist distributions of target proteins at cellular locations ranging from mitochondrial membranes to cytoskeletal fibers (DOI: 10.1126/science.1127344). One set of images shows the detailed distribution of fluorescently tagged Gag, a viral protein from human immunodeficiency virus, on the surface of a cultured monkey kidney cell.
Betzig's interest in developing techniques for seeing what no one had seen before goes back to the 1980s and early '90s when he was in Murray Hill, N.J., at AT&T Bell Laboratories helping to pioneer so-called near-field scanning optical microscopy. In this technique, the diffraction limit can be considerably surpassed by doing away with a lens altogether. Instead, light from a sample is slowly and painstakingly gathered through a tiny aperture scanned just above a sample but at a distance shorter than the normal focal length. The technique is difficult and time-consuming, and it's limited to surface features of a sample. That leaves a lot of biologically interesting territory out of view. Additionally, Betzig notes, the fluorescent labels available at the time did not have adequate specificity. "A lot of the fluorophores were going on to things other than what you were targeting," he says. Betzig, who happily admits to having an intellectual restlessness that inspires him to shift projects and work venues every so often, ultimately lost interest in pursuing near-field techniques.
After a several-year stint in Michigan working for his father's machine tool business, Betzig started getting itchy again a few years ago to make a mark in super-resolution microscopy. The trick, he says, was to find a way to get only those molecules of interest within a minuscule field of view to send out enough photons in such a way that would enable an observer to precisely locate the molecules. He also hoped to figure out how to watch those molecules behave and interact with other proteins. After all, says Betzig, "protein interactions are what make life."
Betzig, who at the time was a scientist without a research home, knew also that interactions with other researchers almost always are what it takes these days to make significant scientific or technological contributions. Yet he was a scientist-at-large spending lots of time on a lakefront property in Michigan, often in a bathing suit. Through a series of both deliberate and accidental interactions in the past two years with scientists at Columbia University, Florida State University, and the National Institutes of Health, Betzig was able to assemble a collaborative team and identify the technological pieces that he and Hess needed to realize what would become known as PALM.
One pivotal realization for the two tool builders happened last year during a trip to Florida State. There they met Michael W. Davidson, a prolific and creative optical microscopist who runs the expansive website Molecular Expression. "He was a catalyst," Betzig says. From Davidson, they learned about the expanding portfolio of fluorescent tags that chemically inclined biologists were designing and making. Some of these tags, they learned, can be selectively turned on and off with specific wavelengths of light. "In the world of biology, there is a new generation of fluorescent proteins that you can switch on at will with a little bit of violet light," Hess says. When he and Betzig found out about those molecular tools, they sought to learn more about them from Lippincott-Schwartz and Patterson. They soon convinced the NICHD folks to become partners in the quest to realize a new type of optical microscope for visualizing biological molecules.
"So now we had our team, but not a lot of money," Betzig says as he recalled an exciting and scary time last year when it seemed to him that the PALM technique that was now gelling in his and Hess's minds was up for grabs. "It was ridiculously simple," he says, not necessarily thinking of the man on the street. Indeed, the researchers' paranoia was justified. Not only was Zhuang then cooking up STORM, but Samuel T. Hess of the University of Maine was working on a nearly identical technique with almost the exact same name: F-PALM, short for fluorescence photoactivation localization microscopy.
"We didn't waste time with writing grants, we just started building" the prototype PALM microscope, says Betzig. "We designed it last summer, built it in Harald Hess's living room, and shipped it in October" to Lippincott-Schwartz's lab. Meanwhile, the biologists on the team had used genetic engineering techniques to create cells that would express specific proteins specially altered to incorporate the photoactivatable tags. "By January," Betzig continues, "we were popping out image after image that was knocking our socks off."
PALM isn't easy to use just yet. For one thing, it takes up to 12 hours to assemble a star field image from literally thousands of optical cycles during which as few as 10 labeled proteins in a cell blink on each time. This long observation time also means that the cells can't move, Lippincott-Schwartz notes. And that means the cells need to be fixed, which means they are dead as doornails during the observations.
Even so, Betzig says with an infectious air of engineering bravado, the observation time can and will be brought down to minutes, even seconds. In time, he says, PALM instruments will get model numbers and price tags and become available to anyone who might want to determine the locations and distributions of specific proteins in cells as they live and breathe. When this tool all comes together as he envisions it will, he adds, "it finally will link molecular biology to cell biology." Moving the tool in that direction will be on the agenda in the coming years for Betzig and Harald Hess, who are both about to begin jobs at HHMI's brand-new Janelia Farm research campus in Ashburn, Va.
Meanwhile, even as STED and the PALM- and STORM-type techniques mature, other scientists are developing their own super-resolution techniques, each one with its own strengths and weaknesses. Mats G. L. Gustafsson of the University of California, San Francisco, for example, has devised and demonstrated a super-resolution scheme called saturated structured illumination microscopy (SSIM) that requires only one standard-issue laser and a wide-field microscope that doesn't need scanning mechanisms as STED techniques do.
Despite this burst of activity, "single-molecule fluorescence is still in its infancy," cautions Davidson. In time, the development of ever-more-sensitive light detectors, new types of fluorescent tags, improved specimen preparation techniques, and more powerful mathematical and computational data processing techniques should help to make super-resolution microscopy part and parcel of what molecular biologists do in their labs, he says.
To be most useful for biologists, the tool developers also need to find ways of reducing the time it takes to get images to the seconds range, Zhuang notes. She is looking forward particularly to the beautiful and telling imagery that will come when super-resolution studies begin to use tags of multiple colors, each one associated with a different molecule. "Then we will be able to determine the locations of several different species and see their interactions. You will be able to see it with your eye," she says, adding with gravity that "you can learn a lot just by seeing."
Over the years, researchers have devised and employed a dazzling colorfest of fluorophores-chemical entities that passively fluoresce or can be activated to do so-for detecting and studying proteins in both dead and living cells.
The members of one huge class of these fluorescent agents are themselves proteins, most notably green fluorescent protein (GFP), which plays a key role in the bioluminescence of some jellyfish. GFP in turn has inspired the discovery or development of a diversity of similarly functioning fluorescent proteins. Through genetic engineering techniques, these fluorescent tags are appended to target proteins, a versatile ability that enables researchers to optically highlight specific proteins in cells.
Another large class of fluorophores is in the form of commercially available organic dye molecules that researchers often first attach to antibodies that will zero in on a target protein. More recently, scientists have been using fluorescent quantum dots, most often nanoscale inorganic particles, which they also attach to antibodies to target specific proteins.
Meanwhile, investigators continue to engineer or discover fluorophore variants that are brighter or smaller, that emit different colors, or whose fluorescence characteristics change in response to biochemical realities, among them enzyme activity, protein-protein interactions, and changing metabolite concentrations. Among the advances in the past few years are photoactivatable fluorescent proteins, which can be switched on in as arbitrarily small a focal region as is technically possible. These developments have made possible the detection of individual tagged protein molecules in cells using optical microscopes.