Introduction to special issue: Biophysics of development

Main text

This special issue of Biophysical Journal focuses on development, a repeatable and reliable transformation of a single cell, the fertilized egg, into a multicellular organism with tissues, organs, physiology, behavior, and reproductive potential. Phenomenological theories of development date back to antiquity and are still, and rightly so, popular among scientists across disciplines. Mechanistic models started to appear only during the 20th century and are increasingly motivated by progress made by cell biologists, geneticists, and biochemists. Modern studies of development monitor and manipulate development at multiple scales, from a single molecule to the whole embryo, and use various mathematical models, from very abstract to very detailed. Articles in this issue offer a glimpse of these efforts by taking the reader through studies of development at the molecular, cellular, organ, and whole-organism scales. In addition to reviews and original research by experimentalists and theorists, this collection includes two historical perspectives on biophysical and mathematical approaches to embryogenesis. The first is the interview with Eric Wieschaus from Princeton University, a geneticist whose work in the 1980s played a key role in launching molecular studies of development. In this interview, which is included below, Eric Wieschaus reflects on his interactions with physicists and theorists. The second historical perspective is by Jeremy Green from King’s College, a developmental biologist who discusses the lasting impact of Alan Turing’s proposal of the Chemical Basis of Morphogenesis. We hope that the readers of Biophysical Journal enjoy this collection and are looking forward to their feedback.

Lessons from model organisms and the biophysical models that describe: an interview with Eric Wieschaus By: Sarah McFann

Earlier this year, I got the chance to sit down with developmental biologist Eric Wieschaus to get his perspective on what biophysics and developmental biology have to offer one another.

Much of Wieschaus’s career has focused on the genetic control of embryonic development. In the late 1970s, Wieschaus and Christiane Nüsslein-Volhard performed a chemical mutagenesis screen that revealed the genetic basis of hundreds of mutant phenotypes (1). In particular, they identified ∼40 genes that affected fly segmentation, the process by which the body of the fruit fly is patterned into repeating segments along the head-tail axis. This work led to their being awarded the 1995 Nobel Prize in Physiology or Medicine, which they shared with geneticist Edward B. Lewis. Many of the genes identified in this screen turned out to be conserved across species, and in the years that followed, the products of many of these genes were characterized and connected to disease and mutant phenotypes in a variety of organisms, including humans.

Wieschaus cites the period directly following the Heidelberg Screen—1980 to 1995—as a major turning point for the field of developmental biology.

“I began studying embryology in the 1960s,” Wieschaus explains. “By that time, Darwinian natural selection had established itself as a general guiding principle of evolution and biology. There was this sense that there might be general principles underlying organismal development, too—a sense that we were getting close to uncovering them.”

Wieschaus says the experiments of Hans Spemann in the 1920s were particularly influential in shaping the opinion that general principles underlay development across species. In a now-famous experiment, Spemann transplanted pieces of tissue from one part of a newt embryo to another. When he transplanted a piece of presumptive skin tissue to a part of the embryo where neural tissue normally forms, he found that the presumptive skin changed its developmental trajectory to become neural tissue. He called the part of the embryo that had converted the tissue “the organizer.” By the 1950s, organizers had also been found in birds and mammals, suggesting that, despite vast differences in the way these organisms develop, the underlying principles might be shared.

Experiments conducted by Wieschaus and his colleagues seemed to confirm that general organizing principles were at play. During the period between 1980 and 1995, several important biochemical pathways underlying development were uncovered. These pathways controlled genetic regulators that determined where specific genes were activated in the embryo. Despite differences in the specific components involved in each pathway, one thing all the pathways had in common was that they appeared to obey relatively simple regulatory logic.

Inevitably, however, as time went on and the identities of more gene products were uncovered, the overall picture became more complicated.

“Around 1995,” Wieschaus says, “the complexity started piling up. The genes we had newly identified provided an essential context for the regulatory logic, but it wasn’t clear to me whether that complexity—the details of such systems-level approaches—really helped in our understanding.”

The desire to return to a more linear way of thinking is what initially drew Wieschaus to biophysics.

“Biophysics seemed to offer a way to get above all these details. Rather than accounting for a large number of molecules, you could use concepts like concentration and force. From an emotional standpoint, biophysics allowed me to get back to the 1980s, when developmental biology was describable as simple linear connections. I actually believe that the early days were correct. What matters ultimately is what varies in a particular cell type—what provides the information.”

The power of precise measurements

Wieschaus’s first foray into biophysics began as a collaboration with Steven Block, an expert in single molecule biophysics. Their first paper together, published in 1998, focused on vesicle trafficking in fruit fly embryos (2). In order to establish polarity, embryos need a way of asymmetrically distributing their internal components. One way to accomplish this is by using vesicles to actively transport cargo along cytoskeletal tracks. In flies, large-scale vesicular transport happens twice in early development: once right after fertilization and again when it’s time for the basic body plan to be established. Wieschaus and Block wondered if they could better understand regulation of vesicle transport by identifying what causes individual droplets to move. With the help of lab members Steve Gross and Michael Welte, they used optical tweezers to stall droplets and thus measure the forces powering them.

Wieschaus likens their approach to the way people naturally learn about the macroscopic world.

“As children, we put our hands into a river and see how the water flows around and through our fingers. Such childhood experimentation gives an intuition for how flowing rivers move generally, the physics of them. We also grab objects to see how they respond to the forces we apply. What would happen if we could grab onto something microscopic in the embryo and pull on it? Does the microscopic world parallel the macroscopic?”

Gross and Welte’s optical tweezers experiments opened up this microscopic realm to interrogation (Fig. 1).

Figure 1.

Steve Gross (left) and Michael Welte (right) in Guyot Hall at Princeton University. To see this figure in color, go online.

“With the optical traps,” says Wieschaus, “we could grab on and measure the stalling forces. We could bring biology into the realm of high-school physics. I wanted to get my hands in there and pull on things.”

Wieschaus says this study highlights one of the key strengths of a biophysical approach: precise measurements paired with physical modeling can provide profound mechanistic insights.

“When we measured the stalling forces required to stop droplets, we noticed they came in discrete intervals. That clued us in to what might be happening at the molecular level.”

Stalling a droplet requires applying more force to it than what is applied by the embryonic components propelling the droplet forward. The fact that the required stalling forces were quantized pointed to molecular motors as the driving force behind vesicular transport. Differences in stalling forces among vesicles could be explained by differences in the number of motors attached to each vesicle.

“Dynein motors are far too small to see with a confocal microscope,” Wieschaus says. “You can’t count them that way. Precise force measurements, however, allowed us to resolve the process of vesicular transport to the point where we were able to count the number of motors powering individual droplets. It was amazing getting so close to the molecular components at work.”

Models that fit, and models that didn’t

Wieschaus pointed to his 2011 Biophysical Journal paper on the Bicoid gradient as another example of what a biophysical approach can lend to developmental research (3). In 1952, mathematician Alan Turing hypothesized that some chemicals are capable of patterning tissues in a concentration-dependent manner. Turing called these chemicals “morphogens.” It took until the late 1980s, however, for the first morphogen-like factor to be discovered in Bicoid, a protein that is distributed in a gradient along the fly embryo’s head-tail axis and that specifies cell fates in a concentration-dependent manner.

At a glance, Bicoid appears to form an exponential gradient, easily described by a reaction-diffusion model that balances Bicoid production, diffusion, and degradation. Fitting such a model to the Bicoid spatial distribution, one can work out the average lifetime of a Bicoid molecule. Without a way to measure the lifetime of Bicoid molecules, however, there was no way to confirm whether this simple biophysical model truly described the reality of Bicoid gradient formation.

To solve this problem, Oliver Grimm, a postdoctoral fellow in Wieschaus’s lab, and Jeff Drocco, a graduate student working with David Tank, tagged Bicoid with the photoswitchable protein Dronpa (Figs. 2 and 3). By switching a small portion of the Dronpa-tagged Bicoid population to the dark state and tracking it, they were able to infer Bicoid’s lifetime. Then, by using their construct to mimic a change in the Bicoid degradation rate, they were able to confirm that changing Bicoid’s lifetime changed the shape of the gradient as expected.

Figure 2.

Jeff Drocco (left) and Eric Wieschaus (right) in Wieschaus’s Guyot Hall laboratory. To see this figure in color, go online.

Figure 3.

Oliver Grimm in Guyot Hall. To see this figure in color, go online.

“It was a pretty and reassuring paper,” Wieschaus says.

Not all of Wieschaus’s work has played out in such a reassuring manner. In his early scientific career, Wieschaus became familiar with the models of biological self-organization developed by Turing and later extended by Alfred Gierer and Hans Meinhardt. Turing’s model shows how chemical patterns can spontaneously arise from a near-uniform mixture as long as the mixture’s components possess certain properties. Mixtures containing short-range activators and long-range repressors, for example, can form stripes and spots.

When Wieschaus began his investigations into early fruit fly development, he expected to find spontaneous self-organization at work. Instead, he found redundancy and over-specification.

The early fly embryo is patterned by maternal gradients that provide outside information about where cells are located in the embryo. Often, more than one gradient is available to help provide this information. For example, in addition to the Bicoid gradient, flies also use spatial patterns of the proteins Nanos and Hunchback to help cells determine their position. This over-specification of positional information biases the embryo toward one specific pattern, leaving no room for spontaneous pattern formation.

Wieschaus finds this redundancy fascinating.

“In theory, the Bicoid gradient is sufficient to inform cells of their location along the embryo’s anterior-posterior axis. And yet genes like nanos and hunchback have stuck around. It makes one wonder exactly what sort of robustness these redundant systems are providing.”

Evolutionarily speaking, Bicoid is a relatively new source of positional information. While Nanos and Hunchback are well-conserved among insects (4,5), Bicoid patterns only a small subset of species related to fruit flies. Despite its niche utilization, Bicoid is good at what it does. In a 2007 study, Wieschaus and co-authors Thomas Gregor, David Tank, and William Bialek employed information theory to assess how much information the Bicoid gradient is capable of communicating to the embryo (6).

“The amount of information a morphogen gradient can provide is limited by how reproducible the gradient is from embryo to embryo,” Wieschaus explains. “The amount of information a cell can interpret, on the other hand, is limited by variability at the gene enhancer level. What are the chances that a gene will fire when presented with morphogen? It’s unclear at this point where the information bottleneck is.”

To gain insight into the reason for patterning redundancy, Wieschaus has taken the approach of removing the fly’s natural patterning gradients. Removing these gradients one-by-one can reveal how much information each morphogen contributes to positional patterning. Alternatively, this approach can be taken to an extreme in which all the patterning sources are removed, transforming the fly embryo into a “blank slate” absent of any external patterning instructions. Wieschaus is particularly interested in what these blank-slate embryos can tell us about the fly’s capacity to self-organize. Beneath all those layers of redundant, external patterning cues, is the capacity for spontaneous, Turing-like pattern formation still present? It’s an interesting question given that spontaneous pattern formation does occur in other early developmental contexts, including the symmetry breaking event of early mammalian embryos.

Removing all of an embryo’s external patterning sources also raises interesting questions about the nature of what remains. Is a fly embryo with no external patterning sources still a fly embryo? Can what is learned from blank-slate embryos be said to be true of fly embryos generally? Wieschaus notes that this type of research speaks to a more widespread recent trend in developmental biology. Since the late 1990s, the embryo has been treated by many researchers more as a sophisticated test tube than as something mysterious worth studying for its own sake.

“Biophysicists and molecular biologists alike have begun using the embryo as a testing ground for studying their favorite biological phenomena,” says Wieschaus. “Whether you’re interested in transcription, cytoskeletal rearrangement, phase separation—whatever it is, the embryo is more likely than not a great place to study it. In a way, the embryo really isn’t being used to study developmental biology anymore. It’s being used to better understand biological phenomena in general.”

The question of generalizability

When asked what most differentiates biophysicists and molecular biologists in their approach to developmental questions, Wieschaus returned to the idea of generalizability.

“It often comes up over beers or coffee. Midway through conversation, the biophysicist will lean over and ask me whether I think the problem we’re working on is truly general. The idea that it might not be concerns them. Personally, I don’t worry about generalizability. I don’t think developmental biology is ready for it.”

The history of physics is distinct from that of developmental biology, and Wieschaus thinks this difference is partly responsible for the difference in expectation regarding generalizability.

“Some of the earliest physics was concerned with understanding what governed the trajectories of stars, but humans had already been watching those stars for thousands of years. They had seen the orbital patterns repeat many times. The story of gravity is similar. You see enough apples fall from trees and it occurs to you that the same general principle might account for every fall. We haven’t seen enough developmental biology to make these sorts of conclusions yet. We haven’t fully characterized enough of the components or fully explored the biochemistries of enough mutant gene products.”

In a 1994 article entitled “Do we understand development?” (7), another famous developmental biologist, Lewis Wolpert, posed the question: Is the egg computable? Specifically, he wondered if biochemistry and biophysics would ever get to the point where it would be possible to predict an organism’s developmental trajectory merely from its DNA.

I asked Wieschaus whether we’re there yet.

“No,” he said. “Will we get there eventually? It’s a difficult question. I think it remains to be seen.”

^∗Note: Some quotes have been condensed for length and clarity.