In late 2014, just a month after learning he had won that year’s Nobel Prize in Chemistry for superresolution microscopy, Eric Betzig and colleagues described a technique that has taken the microscopy world by storm.
Lattice light-sheet microscopy (LLSM) projects ultrathin, low-intensity planes of light into a biological sample, boosting image clarity while reducing phototoxicity and photobleaching, thus allowing researchers to image live cells at high resolution for extended periods. Still in development at Zeiss, and sublicensed to 3i, which has released a commercial version of its own, LLSM has been a hit with the microscopy community, says Betzig, a group leader at the Howard Hughes Medical Institute’s Janelia Research Campus. Besides the systems sold by 3i itself, Betzig estimates his team has guided 70 or so groups in setting up their own versions of the system, and another 40 have gotten time on the system in Janelia’s Advanced Imaging Center in Ashburn, Virginia.
But for all that, the original LLSM had a significant drawback, Betzig says: It was incapable of imaging deep within a sample, because the deeper light penetrates the tissue, the more it ‘warps’ or bends. As a result, though it could image cells on cover slips, it could not do so, say, deep within a growing embryo. Now, that limitation has been lifted.
In the 20 April 2018 issue of Science magazine, Betzig and his team describe a new version of the LLSM system that uses adaptive optics (AO) — a strategy designed to overcome optical distortions in astronomy — to peer deep within living samples. Using that new design, called AO-LLSM, the team capture startling video of endocytosis, organellar dynamics, and neurite growth in developing zebrafish embryos, often for hours at a time.
According to Betzig, the roots of this project extend back to 2006, when he first joined Janelia. “We always clearly suffered a big hit to the resolution and signal as we moved deeper into trying multicellular systems,” he says.
The problem, he explains, is not one of focus, but aberration. It’s like trying to take a picture through a car windshield in the rain without the windshield wipers on: The image is in focus, but blurry.
In 2014, around the time the original LLSM paper was published, his team described in Nature Methods a strategy to circumvent this problem for multiphoton microscopy. The method employs a ‘guide star’ — basically, a bright point of light within the sample — to work out how light passing through the sample is distorted. It then nullifies that aberration by tweaking the properties of a ‘spatial light modulator’ in the excitation beam path, sharpening the image.
“Basically the latest paper is doing nothing more than combining those two papers from 2014,” Betzig says. But, because the LLSM has two independent beam paths, one for excitation and one for detection, the system actually incorporates two complete adaptive optics setups alongside the original LLSM, meaning it’s effectively three microscopes in one. “It’s one microscope, but the complexity is about as bad as if you had three,” he says. (The AO system in the detection path uses a deformable mirror instead of a spatial light modulator, the better to capture photons with high efficiency.)
Indeed, the paper’s supplementary figure 1 shows just how complicated that setup is. Among other things, the parts list includes 35 lenses, 29 mirrors, 7 galvanometers for moving mirrors, not to mention two lasers, three objectives, and four cameras, all arranged on a vibration-dampened optical table.
“It isn’t exactly turnkey,” Betzig allows. “But that doesn’t mean to say that it can’t get there.” Over the past six months or so, he says, his team has been working to redesign the system to make it simpler, more user-friendly, and less expensive. The published system costs probably upwards of $800,000 in parts alone, Betzig estimates, but he hopes to get the next-gen system under $400,000. “It’s gonna take a lot of trimming fat … to get it to that point.”
Ultimately, Betzig hopes to reproduce the success of the original LLSM with its AO-modified cousin, first by simplifying the design, and then by focusing on extensive outreach and documentation. With the original system, he says, his team hosted workshops and posted extensive videos online — more than 16 hours worth — showing how to disassemble and reassemble the system to make it work. Similar efforts are planned for the AO-LLSM. “That’s the big focus of the group right now,” he says.
Jeffrey Perkel is Technology Editor, Nature
Correction (7 May 2018): Eric Betzig’s affiliation was incorrectly listed as ‘HHMI Janelia Farm Research Campus’. It should be ‘Janelia Research Campus’.