Gas vesicles are found in a huge diversity of aquatic microorganisms, from green sulfur bacteria and cyanobacteria to methanogens and haloarchaea, indicating an ancient evolutionary lineage stretching back in time more than three billion years, when these two domains of life — bacteria and archaea — diverged from a common ancestor, an age when life was confined to Earth's oceans.

Structure of Gas Vesicles

What are gas vesicles made of?

Though the term “vesicle” conjures up images of phospholipid bilayers, gas vesicles are unique: they are composed entirely of protein! Shaped like a cylinder capped by cones, these protein nanostructures consist of “ribs” made of gas vesicle protein A (GvpA), a small, hydrophobic protein that is conserved across all gas-vacuolate species. By forming linear crystalline arrays — the “ribs” — GvpA monomers become the major structural units of gas vesicles.

Another gas vesicle protein, GvpC, is often bound to the external surface of gas vesicles: by interlocking between the GvpA “ribs”, GvpC strengthens the gas vesicles and increases their critical collapse pressure dramatically. However, gvpC is not an essential part of the gas vesicle gene cluster and several organisms synthesize gas vesicles lacking GvpC.

Biogenesis

One major question remains: the mechanism of synthesis of gas vesicles by gas-vacuolate microbes.

How do gas vesicles form?

Current evidence suggests that gas vesicle polymerization begins from a bicone stage which gradually extends in length to form the cone-capped cylinder structure. Lengths and widths of gas vesicles vary dramatically between species — these dimensions affect the mechanical strength of gas vesicles — but remain fairly constant between cells of the same species. Electron micrography reveals that any gas-vacuolate cell includes gas vesicles at all stages of growth.

How are they filled with gas?

The answer is surprising: surface tension at the hydrophobic inner surface of the gas vesicle (remember, GvpA is a very hydrophobic protein) excludes water molecules from the interior and only allows the free diffusion of gases through the vesicle walls! As a result, no special mechanism is required to fill up the gas vesicle; by virtue of their biochemical composition, they fill with gas even as they are synthesized.

Bioengineering Gas Vesicles

How have gas vesicles been modified?

In recent years, labs around the world have bioengineered gas vesicles for diverse purposes — from ultrasonic molecular imaging to vaccine delivery — and this versatility comes from one simple structural feature of the gas vesicle: GvpC. Expressed on the gas vesicle surface, GvpC is a hydrophilic protein that accommodates insertions near its C-terminus, enabling proteins of interest to be fused to it. Antigenic peptides, fluorescent proteins and even enzymes have been successfully expressed on the surface of gas vesicles by fusion to GvpC!

How do we do this?

This outcome can be accomplished by two distinct methods, in vivo and in vitro. The in vivo method relies on genetically modifying the gvpC gene of gas-vacuolated microbes and culturing these GMOs to produce modified gas vesicles with the GvpC-fused protein of interest expressed on the surface. The in vitro method involves chemically stripping native GvpC off the surface of the gas vesicles, leaving only the GvpA shell, and refolding recombinantly-produced GvpC-fused protein of interest back onto the stripped gas vesicle surface. The in vivo method has been demonstrated in Halobacterium salinarum NRC-1 by Prof. DasSarma, University of Maryland, while the in vitro method has been used in Anabaena flos-aquae by Prof. Mikhail Shapiro, Caltech.

Isolation of Gas Vesicles

Isolating gas vesicles from the microbes which synthesize them is no trivial task; even getting the gas vesicles out of the cells intact proves to be a difficult, painstaking challenge. The first limitation is that gas vesicles have a critical pressure beyond which they collapse irreversibly due to mechanical failure — flattening like a crushed soda can — ruling out cell lysis by sonication or French press, methods which produce pressures that gas vesicles cannot survive. Cells now can only be lysed enzymatically, a far more expensive method that can't easily be scaled up!

There's a more fundamental problem...

Once the cells are lysed and the gas vesicles are released into suspension, another hurdle awaits us — separating the gas vesicles from the remaining cell fractions is not as easy as it sounds... Intuitively, the solution seems obvious: simply allowing the suspension to stand for enough time should make the gas vesicles float to the surface so that they can be skimmed off; after all, the gas vesicles have an incredibly low density.

But that does not work.

In fact, the gas vesicle suspension must be centrifuged in a swinging-bucket rotor at very low speeds of ~60 g (to avoid hydrostatic pressure induced collapse) for long periods, often overnight, before gas vesicles rise to the surface as a milky-white suspension that must be carefully skimmed off. To ensure sample purity, this tedious process must be serially repeated by diluting the suspension, centrifuging, and skimming off the surface over and over again.

Gas vesicles don't float in vitro.

For some reason, gas vesicles don't float well at all in vitro, unlike gas-vacuolate cells. No bioengineered gas vesicle — despite all its remarkable acoustic, mechanical and surface characteristics — can ever be exploited for its most fundamental property: buoyancy.

And so we decided on our project: to make gas vesicles float again — iFLOAT!

References