There was a fascinating comment added to the recent discussion here on ammonia synthesis. It was pointed out that the amount of man-made Haber Process available nitrogen is outclassed by the amount fixed biologically. The legumes do their share, but a lot more is handled by free-living single-celled organisms. What's really startling is the estimate for the total amount of nitrogenase enzyme, by weight, that is responsible for the production of at least 100 million metric tons a year of reduced nitrogen: about twelve kilos.
It's important for us, as chemists, to contemplate figures like that lest we forget how unimpressive our own techniques are in comparison. Not all enzymes are that impressive, but many of them are extremely impressive indeed. One of Clarke's laws gets quoted a lot, the one about any sufficiently advanced technology being indistinguishable from magic. But there's no magic involved - these are things that we could do, if we just knew enough about how to do them.
Enzymes use a variety of effects to work these wonders, but a lot of it comes down to holding the reacting species in one place and lining everything up perfectly. It isn't as important to hold on to the starting materials or products, as it is to interact with and stabilize the highest-energy species in the whole process, the fleeting transition state. Various chemical groups can be brought to bear that activate or deactivate specific bonds, and everything works, at its best, with near-perfect timing. If you want molecular level-nanotechnology, this is it, and there's absolutely no reason why it has to be done inside a peptide backbone. If we understood enough, all sorts of other polymers, with all sorts of new functionality built into them, could presumably do things that Nature has never needed to do, under conditions that we could select for.
But we're unfortunately a long way from that. There's still a tremendous amount of argument about how even model enzymes actually work, with some rather exotic mechanisms being proposed. And if we don't understand what's going on, we sure can't design our own imitations. Making enzymes from scratch brings together a whole list of Very Hard Problems, from protein folding to femtosecond reaction dynamics, and making enzymes out of something other than proteins will be even harder. We're going to need to be a lot smarter, as a species, to figure out how to do it.
But learning more about such stuff is one of the things we do best. At least for the last few centuries it has been, and if we keep it up, there seems to reason why we shouldn't be able to figure out this one, too. Then, at long last, human ingenuity will have pulled even with blue-green algae, the fungi that live in rotting logs, and various sorts of pond scum. The little guys have had a big head start, but we're gaining fast.