4 Engineering with proteins
What are the prospects for designing and making new proteins for specific purposes? The technology exists to build polypeptide chains unit by unit in a test tube, but this is time-consuming and expensive. Often a more practical approach is to find ways of working with nature to produce useful substances in a form that we can use. This might involve extracting a naturally occurring protein and chemically modifying it in some way, or using genetic engineering to produce a particular protein in controlled conditions. An example of how this has been applied to spider silk is described in Box 5 Genetic engineering: spider silk from goats?
Box 5 Genetic engineering: spider silk from goats?
Proteins are made in cells by following coded instructions carried by the genetic material, DNA. A gene is a portion of DNA that holds the instructions for a particular protein. If these instructions are changed, or added to, then different proteins can be made. Genetic engineering is the term given to a range of new research techniques that allow us to manipulate the DNA in cells.
Gene-splicing is a method for transferring selected genes from one species to another, by cutting a fragment from the DNA of one organism and joining it to a DNA molecule from another organism. Bacteria are often used as the host and can be used to produce large quantities of artificial proteins in the lab.
Spider silk can't easily be made in bulk by farming spiders, because they have an unfortunate habit of eating each other. Transferring the gene to bacteria has not proven successful either, producing silks that are impossible to spin or have low strength. Consequently, some researchers have turned to mammal cells as a possible alternative host and, in 2002, a Canadian company announced that it had succeeded in breeding goats carrying a modified version of the spider-silk gene. The protein can be extracted from the goat's milk, and spun into yarn in a process that mimics conditions in the body of the spider.
The technology is still a long way from practical production, but if development continues we could see spider silk rivalling synthetic polymers like nylon and Kevlar in specialist applications where high toughness is required, from bulletproof clothing to reinforced composites for aircraft panels.
Using naturally occurring proteins in new ways is one thing, but the ultimate goal of designing new protein-based materials or devices from scratch is much, much more difficult. I chose spider silk as an illustration of structure-property relationships within proteins because it has been widely studied and the links between the conformation of the protein chain and the mechanical properties of the material are, at least at a superficial level, relatively easy to spot. However, for most proteins, particularly those with more complex functions, such links are much harder to establish.
Designing a protein does not just involve choosing the right ingredients to achieve a desired structural element: it must also take account of possible unforeseen interactions which might lead to unwanted folded-chain conformations. This aspect is often termed negative design. The goal of negative design is to ensure that only one folded-chain conformation is formed, by ensuring that all other possible structures are energetically unfavourable. Not a trivial task!
In one simple example of protein design, scientists are working to produce improved materials for making water-based gels (hydrogels), in a similar way to that in which denatured collagen associates to form jelly. Hydrogels have many important uses, from highly absorbent dressings and water-retaining granules for agriculture, to structural devices such as contact lenses and components for microfluidic analysis systems.
The polypeptide chains designed to make the gels in my example consist of two different modules. The central portion is a long, disordered chain that cannot fold into a regular structure, but that includes hydrophilic side groups that can associate with water. At the ends are stretches of chain with a row of hydrophobic side groups. The groups from one chain can line up side by side with those from another in a leucine zipper. These form the cross-links, which are weak enough to break apart when heated. By varying the length and the precise composition of the different sections, a range of proteins with different gelation properties can be made.
However, in most cases we are a very long way indeed from being able to design a protein from scratch for a particular function. At the moment, the best we can typically hope to do is to identify the function of naturally occurring protein-based molecular devices and try to identify the key parts of the molecule and how they work. Then, if we are very lucky, we might get away with a tiny bit of tweaking in order to make the molecule fit in better with our system. Many of the current practical applications of bionanotechnology are aimed at doing just that.
(a) What are the principal design requirements for a protein chain if it is to form a random, cross-linked network capable of trapping water to form a gel?
(b) Which aspects would you classify as positive design and which as negative design?
(c) What extra feature is required if the gel is required to break up on heating?
(a) Random implies no crystallisation. An irregular sequence of groups with little capacity for forming secondary bonds is required. The two protein chains must be able to link together at certain points to form a network. The chains must contain some hydrophilic groups to associate with the water molecules and keep them trapped within the network.
(b) Choosing an irregular sequence of groups that won't form secondary bonds is negative design. The linkages, and the need for some hydrophilic groups, are positive design.
(c) If the gel is required to break up on heating then the cross-links must not be permanent, but formed by weak bonds that can break when heated.
As an example of how a naturally occurring protein can be put to a new use, and modified for the purpose, let's look at green fluorescent protein (GFP), which is found in certain types of jellyfish. As its name suggests, GFP emits green light, using energy absorbed from ultraviolet light. Controversially, it has been introduced into animal cells through genetic engineering, to create fluorescent varieties of plants and animals, such as the ‘GFP bunny’ described as a ‘transgenic artwork’ by its creator, the artist Eduardo Kac. Such uses have sparked important public debate on both the safety and the morality of genetic engineering: issues that are likely to increase in significance as our knowledge and capabilities advance further.
A much less controversial application of GFPs is their use in scientific and medical research to ‘label’ an object of interest and to study its behaviour. For instance, GFP can be attached to another protein, whose movement around a cell can then be tracked simply by watching the green glow move around under a microscope. Another recent development is to modify GFPs to act as biosensors, which fluoresce in a characteristic way if a particular substance is present.
The structure of GFP is illustrated in Figure 16. This type of picture is known as a ‘ribbon representation’ as it uses ribbons to pick out key structural features. You should be able to spot a β-sheet, here folded up to produce a ‘can’. There are also short sections of a-helix, coloured red in the diagram, one of which passes down the middle of the can. In the middle of the β-can, safely protected from polar water molecules that would interfere with its action, is the light-emitting centre, shown in yellow on the diagram. This is made up from a particular sequence of three amino acids, which interact together to form a structure in which the band gaps are suitable for the absorption and emission of light.
We can modify this structure by making slight changes in the amino acid sequence or by adding portions to the end of the chain, while keeping enough of the original arrangement to ensure that the basic design is not compromised. How might we want to change it, to increase its usefulness? One possibility is to tweak the colour of the light emitted, by making slight adjustments to the environment of the light-emitting structure. Variations of GFP that fluoresce blue, cyan and yellow have all been made, allowing different substances to be tagged and monitored at the same time. Another common modification is to provide attachment sites for different molecules by adding flexible linkages to the ends of the polypeptide chain, so that the range of substances that can be labelled is increased. In another example, a blue fluorescent protein has been developed to detect the presence of zinc ions; zinc ions interact with the light-emitting centre in a way that increases the level of fluorescence, so that the level of zinc present can be measured from the brightness of the signal.
This is just one example of how we can learn from nature by taking a naturally occurring molecular machine and extending its function to suit our own goals. As we work our way through the cell we will see many more biological devices and learn further principles for working at the nanoscale, which we may be able to adapt and apply to new situations. One day we will surely know enough to design novel nanomachines of our own.
(a) Describe aspects of the primary and higher-order structure of the GFP protein that make it particularly suited for its purpose as a light emitter.
(b) Which parts of the structure would you expect to be most open to modification, without compromising the function of the molecule?
(a) The key feature of the primary structure is the sequence of three amino acids that forms the light-emitting centre. The key features of the higher-order structure are the β-can, which protects the light-emitting centre from the polar water molecules that surround the structure and would interfere with its operation, and the α-helix, which positions the light-emitting centre in the right place in the centre of the can (both of these would, of course, require a suitable primary sequence of amino acids to form, but different variations are possible).
(b) The loose loops at the top and bottom of the structure, and the chain ends, would be the obvious places to attempt to modify the structure, as their precise shape is likely to be less crucial to the functioning of the device.
Suggest an engineering application that could benefit from the manufacture of protein-based structural materials.
The example I thought of was tissue engineering, where biological materials are being developed to provide substitutes for living tissue – for instance, to replace skin or to repair damaged blood vessels. An ultimate aim is to find ways to prompt the body to grow its own replacement parts.