Protein Crystal Growth in Microgravity: What ISS Crystals Show

For more than three decades, researchers have been sending protein solutions to orbit to see whether crystals grow better there. Protein crystal growth in microgravity matters because X-ray crystallography, the method that has revealed most known protein structures, only works if a protein can first be coaxed into a well-ordered crystal, and many proteins refuse. Orbit removes much of what gravity does to a growing crystal. This article covers what the experiments measured, one drug-company case in detail, and why the verdict is more cautious than the headlines suggest.

Key Takeaways

• Gravity drives sedimentation and convection in a crystallization drop; in orbit both are strongly reduced, so molecules arrive at a growing crystal mainly by slow diffusion.
• Some space-grown crystals have diffracted X-rays to higher resolution than Earth-grown ones. In Cryostat flights, satellite tobacco mosaic virus (STMV) crystals reached beyond 1.8 angstroms against 2.3 on the ground.
• A 2015 review by two researchers active in the field described the evidence as resting on only a few, generally nonrepetitive, highly restricted trials.
• Merck's 2017 test of the cancer antibody pembrolizumab (Keytruda) produced uniform crystals in orbit, and the team later reproduced uniform crystals on Earth.
• Competing methods, especially cryo-electron microscopy, do not need crystals at all.

Why do scientists need protein crystals at all?

Proteins are the molecular machines of cells. Each folds into a precise three-dimensional shape, and that shape decides what the protein binds, what it does, and how a drug might interfere with it. Seeing a protein at atomic scale lets a chemist ask where a drug molecule would fit.

The standard way to see it is X-ray crystallography. Researchers coax many identical copies of a protein into a regular, repeating lattice, which is a crystal. They fire a beam of X-rays through it. X-ray wavelengths are about as long as the spacing between atoms, so the lattice scatters them into a pattern of spots. Computers turn the pattern into a map of electron density, and from the map a model of the protein is built atom by atom. The same logic once revealed DNA's double helix; the story is told in the articles on the history of DNA discovery and Rosalind Franklin.

Angstroms are the unit of choice. One angstrom is 0.1 nanometer, about the width of a hydrogen atom. A resolution of 3 angstroms is usually enough to trace the protein chain, and resolutions near or below 1.5 angstroms begin to show individual atoms and ordered water molecules. Lower numbers are better.

How dominant is crystallography?

Very. According to the Wikipedia summary of the Protein Data Bank, as of May 21, 2026, the archive held 254,149 structures.

Method	Structures in the PDB	Share of total
X-ray crystallography	204,327	about 80%
Electron microscopy	34,316	about 14%
NMR spectroscopy	14,761	about 6%

Source: Protein Data Bank statistics as summarized on Wikipedia, May 21, 2026. Shares calculated here and rounded.

Crystals are therefore the bottleneck for most of structural biology. A protein that will not crystallize cannot be solved this way. Many membrane proteins, large complexes, and flexible molecules are stubborn, and every extra bit of crystal order translates into sharper data.

What does gravity do to a growing crystal?

A crystallizing protein is a bit like building a brick wall while someone keeps shaking the table. Molecules have to find their place in the lattice, and anything that jostles them or drops debris on the wall leaves flaws.

According to a 2015 review in npj Microgravity by Alexander McPherson and Lawrence DeLucas, two effects of gravity are central. The first is sedimentation. Time-lapse recordings suggest that nuclei reaching microcrystal size sink to the bottom of their containers, where they touch the container surface and other growing crystals, which can affect shape, perfection, and final size. In microgravity, the same authors note, crystals hold stable positions for long periods, appear roughly evenly spaced, and rarely collide. The second effect is convection, the stirring caused by density differences. Without gravity, convection is absent, and material reaches a crystal by diffusion, which is very slow for large molecules.

The depletion zone: a moat around the crystal

That slowness has a surprising benefit. When a crystal grows, it uses up protein in the liquid right beside it, creating a region of lower concentration called a depletion zone. Think of a food stall at a festival: the people nearest the stall eat first, and the crowd around it thins out. On Earth, convection constantly mixes the liquid and erases the zone. In orbit the zone can persist, and its stability is a central hypothesis in the McPherson and DeLucas paper.

The depletion zone may also act as a filter. Large impurities and clumps diffuse even more slowly than single protein molecules, so they are less likely to reach the crystal surface and get built into the lattice. Cleaner growth means fewer defects. The review also notes that protein crystals typically nucleate at supersaturations of roughly 200 to 1,000 percent, compared with a few percent for conventional crystals, which helps explain why proteins are such fussy builders.

In the McPherson and DeLucas review, American investigators on the Mir space station reported that 5 of 21 proteins gave results superior to Earth-grown crystals. That is a real signal, and it is also a reminder that most proteins in that set were not reported as superior.

For more on how crystals grow in ordinary laboratory solutions, see the article on how crystals grow from solutions.

A short history of crystallizing proteins in orbit

Microgravity protein crystallization has been under study since 1985, according to the same review, with support from NASA, ESA, JAXA, the Canadian Space Agency, Germany's DLR, and the China National Space Administration. The hardware changed repeatedly.

Era	What flew	Detail
Before January 1986	Hand-held vapor diffusion apparatus	Flew on four Shuttle missions; trays held 24 chambers, each drop 20 to 40 microliters; directed by Charles Bugg, Fred Suddath, and Larry DeLucas at the University of Alabama at Birmingham
October 26, 1988	Enhanced vapor diffusion apparatus	Shuttle flights resumed after Challenger; 60 experiments per unit
Shuttle Spacelab era	Cryostat (German DARA)	14 temperature-controlled chambers on IML-1
Shuttle Spacelab era	Advanced Protein Crystallization Facility (ESA, built with Dornier)	Two units totaling 96 reactors on IML-2, a 13-day mission
Shuttle USML-1	Glove-box experiments	DeLucas flew as payload specialist and ran hundreds of experiments in nearly two weeks
Mir and ISS	PCAM and DCAM	One PCAM cylinder held 63 experiments; an incubator held 378; DCAM grew centimeter-scale crystals for neutron diffraction
2002	Granada Crystallization Facility	Up to 138 capillaries in a 12.5 × 12.5 × 8.5 centimeter box

Source: McPherson and DeLucas, npj Microgravity, 2015.

The programme also lost hardware to accidents: one interferometry system for observing crystal growth, the Observable Protein Crystal Growth Apparatus, never flew because of the Columbia disaster.

The International Space Station has been continuously inhabited since November 2000, and NASA describes it as orbiting about 250 miles (roughly 400 kilometers) above Earth. The ISS National Lab defines the weightlessness there as the perceived weightlessness of free fall: the station and everything in it are falling around the planet together. Learn more in the article on the ISS and living in space.

How does an orbital crystallization experiment actually run?

Most experiments follow the same pattern: prepare on the ground, fly, grow, return, analyze.

Preparation. Scientists purify the protein and mix it with a precipitant, a chemical such as polyethylene glycol that pushes protein out of solution slowly. The mix is loaded into sealed hardware. Some designs use vapor diffusion, where water slowly leaves a protein drop toward a reservoir; others use batch methods, where everything is mixed up front and temperature does the work.

Flight. The hardware launches on a cargo vehicle and is installed in a temperature-controlled incubator or left in a stowage locker. Merck's 2017 experiment, for example, launched aboard SpaceX's CRS-10 mission in February 2017.

Return and analysis. On the ground, crystals are examined under a microscope, often flash-frozen to preserve them, and exposed to intense X-rays, typically at a synchrotron. Researchers then compare diffraction quality against control crystals grown on Earth from the same stock.

That comparison is the hard part. McPherson and DeLucas list four criteria: visual quality, size distribution, shape, and X-ray diffraction quality. Resolution limits are set where the signal-to-noise ratio drops below a chosen threshold, applied consistently to both sets of crystals. A fair test needs proper controls, which is not trivial when a flight happens once every few months.

Reading the numbers: resolution, mosaicity, and rocking curves

Two quality measures appear repeatedly in this literature, and both are easier to grasp with an analogy. Resolution is like the megapixel count of a photograph of the protein: a lower number in angstroms means finer detail. Mosaicity is like how neatly a stack of floor tiles lines up. A perfect crystal is a single block in which every unit cell faces the same direction; a poor one is a mosaic of slightly tilted blocks. Mosaic spread shows up when a crystal is rotated through the X-ray beam and the diffracted spot is measured across angles. A narrow "rocking curve" means the blocks are well aligned. The review reports that Helliwell's lysozyme crystals grown in microgravity had narrower rocking curves than ground controls, which points to better internal order even where resolution alone does not change much.

What have the measurements shown?

Early space results were striking enough to launch a field. Selected comparisons reported in the McPherson and DeLucas review:

Protein or system	Space result	Earth comparison
Satellite tobacco mosaic virus (STMV), Cryostat	Diffracted beyond 1.8 angstroms	Over 15 times smaller in volume; 2.3 angstroms
STMV, cubic form, APCF	About 4 angstroms resolution; over 30 times the crystal volume	About 6 angstroms
Canavalin, rhombohedral form	2.3 angstroms	2.8 angstroms
Canavalin, hexagonal form	2.2 angstroms	2.7 angstroms
Mir investigators' test set	5 of 21 proteins gave superior results	The other 16 were not reported as superior

Source: McPherson and DeLucas (2015). Lysozyme crystals studied by Helliwell showed narrower rocking curves in microgravity, indicating better internal order; the paper reports no single number for the improvement.

Turnip yellow mosaic virus, 280 angstroms across and 3.5 million in molecular weight, was described as the largest macromolecule crystallized in space at the time of the review.

Notice what these numbers are. They are improvements in specific cases, measured by specific teams. They are not a general rule that orbit yields crystals 40 percent or 100 percent sharper, a claim that circulates widely online and that the primary literature cited here does not support.

Why STMV improved so much

STMV is a good case for the depletion-zone idea. Microscopy work summarized in the review shows nuclei and microcrystals settling onto growing STMV and lysozyme crystals and being incorporated, sometimes with visible defects. Remove the settling and the crystal grows by adding single units, which is slower but tidier.

Case study: Merck, pembrolizumab, and a crystal suspension

The best-documented modern experiment is not about structure at all. It is about drug formulation.

Pembrolizumab (Keytruda) is a monoclonal antibody, a Y-shaped immune protein engineered to help the immune system attack tumors. The article on antibody structure explains the shape. It has been given to patients by intravenous infusion in a clinic. A concentrated injectable form would be more convenient but is hard to make, because very concentrated antibody solutions become thick and difficult to push through a needle.

Merck researchers tried crystallizing the antibody into a suspension of many tiny uniform crystals, which flows better than a dissolved solution at the same dose. According to the paper by Paul Reichert and colleagues in npj Microgravity (December 2, 2019), they used batch crystallization in 35-bottle handheld hardware, with a temperature ramp from 4°C to 30°C over 48 hours, and sent samples to the ISS on CRS-10.

Measurement	Flight sample	Ground control
Particle size distribution	One peak, near 39 micrometers	Two peaks, near 13 and 102 micrometers
Viscosity at 50 mg/mL	3.66 centipoise	5.48 centipoise
Viscosity at 75 mg/mL	4.80 centipoise	6.83 centipoise
Time to sediment	About 57 minutes	Not fully settled after several hours

Source: Reichert et al., npj Microgravity 5:28 (2019). The vehicle (liquid without crystals) measured 3.14 centipoise.

The team reported that dissolved crystals retained more than 94 percent of their binding activity and that crystallization did not increase aggregation, which are important checks for a medicine.

The twist: Earth caught up

The paper's most useful finding is about Earth. Having found that sedimentation and temperature gradients seemed to be the key variables, the team reproduced uniform crystals in ordinary laboratory conditions by rotating samples at 24 revolutions per minute to limit sedimentation and by controlling the temperature gradient. An inverted temperature gradient, from 50°C down to 22°C over 24 hours, produced even smaller uniform crystals with a geometric mean of about 1.3 micrometers. The ISS National Laboratory's account of the paper also stresses that the team reproduced uniform crystalline suspensions on Earth using rotational mixers.

That makes the flight a diagnostic tool: it showed which variables mattered, and engineers then copied the microgravity conditions on Earth. This pattern, in which orbit teaches and the ground applies, appears repeatedly in the literature.

What this did not produce

It is easy to blur the Merck story with a real product, so a clear line is worth drawing. On September 19, 2025, the FDA approved pembrolizumab and berahyaluronidase alfa-pmph (Keytruda Qlex) for subcutaneous injection. The FDA announcement describes a combination of the antibody with an enzyme, hyaluronidase, and does not mention crystals. Merck's own description of the ISS work used forward-looking words such as "could" for the benefits it hoped crystal formulations would bring. Nothing in the sources reviewed here shows that the approved product came from space-grown crystals.

What does the honest scorecard look like?

The McPherson and DeLucas review is candid, which is rare in a field with commercial interests. Its main conclusions:

• The evidence rests on only a few, generally nonrepetitive, highly restricted trials.
• Some European (INTOSPACE and Cosima) and Mir results were indeterminate; some proteins improved and others did not.
• Many studies lacked detailed quantitative analysis of crystal size, shape, defects, or diffraction.
• A National Research Council review found results inconclusive, improvements largely incremental, and control experiments difficult, and urged focus on challenging biological targets.
• Adoption by structural biologists has been limited by restricted access, launch and retrieval delays, heavy paperwork, and the uncertainty of spaceflight.

Put together, the review's verdict is that microgravity shows potential for improved crystal quality, but the evidence base is thin and the practical barriers are high. That is a much weaker claim than "the ISS cures diseases," and a more honest one.

Is the trip worth it? Cost, access, and the ground alternatives

The review's reference list includes a title that captures the cost question: "Mir for the crystallographer's money," a 1992 Nature article by Stoddard and colleagues that addresses the cost-effectiveness of crystallization on the Mir station. The question has not gone away. Flight opportunities are scarce, researchers often wait a year or more for a flight, and a failed run cannot be repeated next week. The authors project that commercial transport to the ISS could give roughly bimonthly access, which would help, and the ISS National Laboratory's program is meant to make crystallization on the station easier and more economical.

Some Earth-bound alternatives mimic what orbit offers. Merck's rotating mixers, for example, slowed sedimentation enough to reproduce its flight result. Each alternative removes one effect of gravity, rarely all of them at once, which is part of why orbit still attracts experiments.

The competition: crystals are not the only route

Two other developments changed the calculus. First, cryo-electron microscopy, which won Jacques Dubochet, Joachim Frank, and Richard Henderson the 2017 Nobel Prize in Chemistry, images flash-frozen molecules without crystals. By 2020, according to the Wikipedia summary, the best cryo-EM resolution recorded was 1.22 angstroms, though most single-particle structures were at 3 to 4 angstroms. Second, machine learning now predicts many structures directly from sequence; see the articles on AlphaFold and how AlphaFold solved protein folding.

Neither makes crystallography obsolete. Drug designers still want experimental structures of proteins bound to drug candidates, and crystals of such complexes are routine in industry. But for the most stubborn targets there are now more ways around the problem than launching samples to orbit.

What is running on the station now?

The ISS National Laboratory runs a Microgravity Molecular Crystal Growth program intended to make crystallization research easier and cheaper to do on the station. In a June 12, 2018 article, it listed several investigations:

• Eli Lilly and Company was running crystallization experiments aimed at structure-based drug design.
• Oak Ridge National Laboratory planned to crystallize the enzyme acetylcholinesterase, to help develop an improved antidote for nerve agent and pesticide poisoning.
• The Michael J. Fox Foundation was growing crystals of the LRRK2 protein, a key target in Parkinson's disease research.

These were announced plans, and the 2018 article reports no numbers on crystal size, quality, or yield. This article found no published outcomes for them in the sources checked, so it does not claim any.

Other ISS biology is covered in the articles on Alzheimer's research in space, cancer cell behavior in zero gravity, genetic expression in microgravity, organ development in microgravity, and pharmaceutical development in space. For orbital manufacturing of non-biological materials, see advanced materials processing in space. And for the biology of what proteins are and how cells build them, see how proteins are made.

Common claims that need correcting

Some assertions about space-grown crystals appear often online and are not supported by the sources above.

"ISS crystals have produced dozens of approved drugs." The primary reviews and program pages cited here describe potential for drug design and specific research projects, not a documented chain from an ISS crystal to an approved medicine. Structure-based drug design is real, but most of the structures it uses come from Earth-grown crystals.

"Space crystals are always 100 times better." The measured improvements above are specific and varied. In the McPherson and DeLucas review, resolutions moved by fractions of an angstrom in some cases and by a few angstroms in others, and many proteins showed no clear gain.

"Convection is the whole story." Convection matters, but so do sedimentation, the depletion zone, and the filtering of impurities, and which dominates differs by protein.

"It is cheap and routine." Access is restricted, launches and returns cause delays, and the review cites heavy paperwork.

What this means

The most defensible reading of three decades of work is that microgravity changes crystal growth in real, measurable ways, sometimes improving the crystals researchers need, and that the benefit is uneven. The clearest modern lesson comes from Merck's antibody: orbit revealed which variables mattered, and engineers then reproduced the result with a rotating mixer and a temperature gradient on the ground. If that pattern holds, the long-term value of orbital crystallization may lie less in producing crystals for every project and more in teaching scientists what gravity gets in the way of, so they can design around it at home.

---

Sources

• McPherson, A., and DeLucas, L. J., "Microgravity protein crystallization," npj Microgravity 1, 15010 (2015).
• Reichert, P., et al., "Pembrolizumab microgravity crystallization experimentation," npj Microgravity 5, 28 (2019).
• ISS National Laboratory, "Crystallizing Proteins in Space for Drug Development," June 12, 2018, and "Published Results From Crystallization Experiments on the ISS Could Help Merck Improve Cancer Drug Delivery," December 2, 2019.
• U.S. Food and Drug Administration, "FDA approves pembrolizumab and berahyaluronidase alfa-pmph for subcutaneous injection," September 19, 2025.
• NASA, International Space Station.

FAQ

Q: Why does microgravity help protein crystals grow? A: On Earth, gravity makes small crystals sink and drives convection that stirs the solution. In orbit both effects are strongly reduced, so molecules reach a crystal mostly by diffusion. A depletion zone can form around the crystal and may filter out large impurities, which can mean fewer defects.

Q: Do space-grown crystals always diffract better? A: No. In a 2015 review by two researchers active in the field, results were mixed. Some proteins improved clearly, such as satellite tobacco mosaic virus, which diffracted beyond 1.8 angstroms against 2.3 on Earth. Others did not. On Mir, 5 of 21 proteins gave results superior to Earth-grown crystals.

Q: Did the ISS Merck experiment lead to an approved drug? A: No source reviewed here shows that. Merck grew crystals of pembrolizumab on the ISS in 2017 to study formulation, then reproduced uniform crystals on Earth. The subcutaneous product the FDA approved in September 2025 combines the antibody with an enzyme, and the FDA announcement does not mention crystals.

Q: Can the same results be achieved on Earth? A: Sometimes. The Merck team identified sedimentation and temperature gradients as key variables and reproduced uniform crystals in the laboratory by rotating samples at 24 revolutions per minute and controlling temperature gradients. Other proteins may not be as easy to fix.

Q: Do scientists still need crystals to see protein structures? A: Often, yes. About 80 percent of the 254,149 structures in the Protein Data Bank (May 2026) were solved by X-ray crystallography. But cryo-electron microscopy needs no crystals and won the 2017 Nobel Prize in Chemistry, and machine-learning tools can predict structures from sequence.

Q: What are researchers growing in space now? A: As of the ISS National Laboratory's 2018 overview, projects included Eli Lilly's structure-based drug design work, Oak Ridge National Laboratory's acetylcholinesterase crystals for a nerve agent antidote, and the Michael J. Fox Foundation's LRRK2 crystals for Parkinson's research. This article found no published outcomes for those projects.

Q: Why is orbital crystallization not more widely used? A: The McPherson and DeLucas review cites restricted access, launch and retrieval delays, heavy paperwork, and the uncertainty of spaceflight, along with a thin and generally nonrepetitive evidence base.

---

Read next:

• How Mass Spectrometry Identifies Molecules
• MRI Technology: Magnetic Resonance Imaging Explained
• Bone Loss in Weightlessness: ISS Research on Osteoporosis