Slashdot Mirror
Microbe Processors
gpmap writes "Smart microbes are closer to reality than you might think, as described in an interesting article on Boston Globe Online. Ron Weiss, a Princeton biochemist, has already programmed E. Coli bacteria cells that release a fluorescent protein when they're exposed to certain chemicals. Now that a team at Stanford University has found a computer-based way to make cells react to any known chemical, the idea of weapons-detecting microbes looks even more promising. That is just one of the more modest applications of a remarkable new engineering discipline -- the science of programming cells. Imagine thousands of preprogrammed cells coursing through your bloodstream, checking cholesterol levels and patrolling for cancer. Or an army of bacteria powerful enough to suck the unwanted contaminant out of a whole lake, but smart enough to turn themselves off when no longer needed."
I want microbes that combine to taste like a twinkie or Oreo cookie, but after I swallow them, they dutifully form themselves into some form of dietary fiber that will not give me the runs (I'm thinking of things like Olean(TM) or olestra) nor will it constipate me. If they can do that, the world will be a wonderful place indeed.
It's not enough to bash in heads, you've got to bash in minds. - Captain Hammer
I disagree. It just has to function at some minimally acceptable level the first time. Life is inherently squishy and flexible, there really is no "right" answer. Take an enzyme, for example. It might be 300 amino acids in length, but the number of amino acids that are explicity required for catalysis or proper folding is a tiny fraction of that number. An enzyme I've worked on has two main isoforms, sequence identity within an isoform is about 50%, between isoforms drops down to 10%. They all have comparable levels of activity. Then you can add to it some of life's error checking/avoiding. For example, if you look at the genetic code, of course you've got three bases A, T, C, or G in a row making up a codon, for 64 possible codons but only 20 standard amino acids. Except for the amino acids methionine and tryptophan, each amino acid can be encoded by at least two different codons. The redundancy isn't higgily-piggily either, alanine is encoded by GCA, GCC, GCG, and GCT--the last position is variable instead of just randomly varying across all three positions. Further, similar amino acids are often encoded by codons that are similar. For alanine, if we change for example the first position we can get either serine, proline, or threonine. These amino acids are not entirely unlike alanine and depending on location within the protein the difference may be negligible. Similarly, if we change the first position of the alanine codon, we can get valine, aspartic acid, glutamic acid, and glycine. While aspartic acid and glutamic acid are unlike alanine, valine and glycine are similar. Still, that's a layer of protection in the event of a screw up. If only computers were 1% as robust as life, then the appearance of the Blue Screen of Death (TM) would be so rare that it would warrant an article in the friggin newspaper instead of an almost daily occurance for those of us cursed with Windows 98 boxes.
Combinatorial, baby.
So, if coli make an error in DNA replication 1/10^6 bases, and we make a huge guess at the probability of this error being somewhere important and doing something bad being 1/10^9, and we replace the cells living in the bloodstream once a year, and we have the safety features mentioned above, this becomes a very long sentence. There is little chance of something going wrong.
I'd still like to see how they are going to get these bacteria to live happily within the human body and not be noticed by the immune system, and not produce too much trash, and not grow uncontrollably.
An interesting idea, but ... give 'em a few grants and wait a few years.
...to see the interface between engineers of two different fields. I would have expected to see more knee-jerk "OMFG" posts speaking about the supposed dangers of this, and it's refreshing to see that there's few. It's also nice to see a post get into the nitty of nucleotides, Watson-Crick base pairing and the base pair wobble rules.
For those concerned about unleashing a transgenic bacterium into the wild that could have horrible consequences, consider the following (wrt. academic research): The E. coli used in research labs are most often "recombinant deficient," and viruses are most often "replication deficient." What this essentially means for E. coli is that genes coding for proteins that integrate foreign DNA into the bacterium's genome have been "knocked out" by one method or another (gross mutation, removing significant chunks of promoter sequence and coding sequence, etc). Plasmids carrying the ampicillin resistance gene that are subsequently transfected into the bacteria confer ampicillin resistance to those that carry it, and even replicate to a high copy number within the individual bacteria, but the ampicillin resistance gene does not get integrated into the host genome. Unless the culture is grown in a media that has a constant selection pressure (ampicillin), the bacteria tend to lose their plasmid. Keep in mind that Amp is heat labile and generally has a short half-life at 37 degrees C.
Likewise, replication deficient virus vectors have certain necessary virus sequences (certain packaging genes, promotor sequences, and long terminal repeats), but they also lack the crucial coding sequences that would give them the ability to become harmful and fully replicative. Example: A lentivirus (a subset of the retrovirus) vector can be made to be self-inactivating once it integrates into the host's genome. Additionally, for a researcher to make virions, they must transfect the virus vector into a transgenic "packaging" mammalian cell line that produces the viral coat proteins. Once you have enough of a virus titer, you use these virions to infect other cell lines (which are minus the genes that code for the coat proteins).
In a lot of cases, removing any coding sequences for proteins that facilitate DNA recombination in a host makes this sort of research quite safe when carried out under Good Lab Practice(tm).
While scientists are generally held closely to the ideals of the scientific method by their peers, there are protocols for minimizing risk (using recombinant minus host organisms), and these must also be followed. No scientist wants to be known throughout the world as the one who unleashed the monster.
With respect to the original story poster's comment about "turning themselves off:" It may have caused some confusion. The action of a cell responding to a chemical presence is usually through an inducible promoter system. The presence of a chemical catalyzes the recruitment of transcription mechanisms to the necessary location so they transcribe mRNA that codes for whatever the response is. In this case (as in indicator), it's the enhanced green fluorescent protein (EGFP). If you remove the chemical, the transcription machinery isn't mobilized to the promoter, and ultimately, EGFP isn't produced. Quite elegant. Do some searches for the lac operon or the trp operon if one is curious.
To end this and tie back into engineers of 2 different fields, a molecular bio grad student around our lab often wears the "Code Poet" shirt from ThinkGeek. I tend to think it's as in context as in CE/CS, wouldn't you? DNA is code of the most ancient sort. After looking at lines of ATG GTC CCA CGT CAC... for a while, one could liken it to assembler.
We haven't figured out our compiler fully, though.
Cheers!
-- I'd say your post was about 3 monkeys, 18 minutes.