So, in the preceding blogpost I introduced my current crowdfunding project, IndieBB:
..If you navigate to the “updates” panel on that crowdfunding page, I've been busy keeping things current. Among the promises I've made lately was to run through my workflow for total-plasmid-design as I plan to apply to IndieBB.
So, here we go. Before clicking “more”, be aware that you do not need to understand any of the below to complete the IndieBB kit. IndieBB is designed for rank beginners to biotech; this blogpost is intended for ‘advanced’ readers with an interest in my design process.
With that said:
A bit of history-of-Cathal for starters; this isn't my first plasmid-design project. People who know me for long may recall me talking about an “IndieBB” a long time ago, and of which I've spoken little since. Yes: I had a prior project, also called “IndieBB”, which deserves its own blogpost during this campaign so supporters can see where I've made mistakes and learned from them. Suffice at this point to say; I've done this before, and it didn't work well enough to be released. However, there are sound and predictable (in hindsight) reasons why that was the case, and despite those reasons it still did work partially.
Since that project, my first synthetic biology project in fact, I have learned a great deal about DNA design. In fact, I've learned enough to have written my own software to assist in the process, and I've used that software to design genes that worked excellently. Further, while that project was highly speculative in nature, the current project is based upon “modules” which are individually well understood and known to work on their own. And, since they do not interdepend, I see no reason why they should not work when combined.
That said, allow me to defer relating my prior experience of genetic design and plasmid engineering, and cut to how I plan to do it this time around. I promise I'll reveal more about the first IndieBB, and what became of it, in a later post.
How to make a plasmid Pt1: Minimal Plasmids
Plasmid design requires knowledge of how plasmids work, and there are several ways plasmids can work. First, let's define “plasmid”. A plasmid is a piece of DNA, usually found in bacteria, which is separate from the chromosome(s) of the bacteria, has a distinct (usually minimal) method of reproduction, usually does not bear critical genes, and is significantly smaller than a typical chromosome. Contrary to popular belief, plasmids are not always circular, though most of them are as the mechanisms of replication is less complicated for circular DNA.
The prototypical plasmid replication system is “Rolling Circle Mechanism” (hereafter ‘RCM’). There are several other significant forms of plasmid, which I will not go into for brevity. IndieBB will, in all likelihood, use a RCM-type core.
Rolling-circle mechanism requires only a single plasmid-borne protein (usually called “rep” for “replication protein”), plus a target origin-of-replication (ORI) site and a non-coding section of DNA usually called a “single-stranded regeneration sequence”, but more memorably known as a “single-stranded origin” (ssORI). Don't worry: I'll try to avoid alphabet soup here.
In brief, because there's plenty out there if you want to go deeper into how RCM works: the ori target-site is where the rep protein, when available, snips and opens the plasmid for replication. This leaves a loose semi-double-stranded end of DNA, which the bacterium's own DNA replication systems will extend, copying one strand of the plasmid while displacing the other strand with that replacement. When the copied strand returns to the ori site, it is “finished” by an interaction between the rep protein that cut it, and another rep protein, which sort-of neutralise one another and finish the replication process.
Now you have a circular plasmid as before, and a single-stranded copy that's been spooled-off. This is usually made double-stranded once again by the single-stranded regeneration sequence (ssORI), which, when single-stranded, folds up in a special shape that mimics a gene promoter just enough that the cell's protein transcription machinery mistakenly begin making an RNA copy of the DNA, then abort. This RNA copy is used as a primer from which to copy the missing strand, creating a second full plasmid.
Technically speaking, all you require for a functional rolling-circle-mechanism plasmid is your origin cassette, consisting of your ORI site, your rep protein, and the single-stranded regeneration sequence. While wild plasmids may have these elements thrown around, in artificial plasmids they're bundled together. In fact, most plasmid maps just wrap the whole thing and call it “ori”. This can be quite small, less than a thousand base-pairs (‘kilobase’) in length. Especially if, as many plasmids do, it lacks the ssORI sequence element.
However, there are two things about that minimal plasmid that suck. Firstly, it's got no payload; it doesn't do anything for you or the cell carrying it. Secondly, it's got no mechanism of selection. Some words on the latter.
How to make a plasmid Pt2: Artificial Selection
If you're mucking about with Plasmids, you're doing it in-vitro: isolated DNA, snipped and pasted by enzymes and added to other DNA to create a desired final sequence.
Once you're ready to put that DNA into a target cell, you'll be using a process generically called transformation to make the target cells absorb the DNA and begin using it. But here, you have some problems. Firstly, no method of transformation is 100% effective, and most are highly inefficient, meaning that of the pool of cells you provide DNA to, a tiny fraction will actually absorb the DNA. Secondly, those cells that do bear the DNA may have no strong reason to tolerate it; especially man-made DNA that usually forces the cells to do something that doesn't especially suit their evolutionary best interests.
Putting it like this makes a distinction that most researchers and engineers don't bother drawing, but which I prefer to for clarity's sake; there are two problems here, initial selection and ongoing selection.
Usually, these are solved by including a gene on the plasmid that provides plasmid-bearing cells with resistance to an antibiotic. This means that initial selection is facilitated by growing the cells on antibiotic, and the majority of non-plasmid-bearing cells will fail to grow, leaving only transformed cells (and mutants). Filter out the mutants by testing them for plasmid, and you've got your initial selection. Then, going forward, you only use antibiotic-supplemented broth and agar to grow the cells, preventing cells from losing the plasmid.
Tidy and all as this sounds, it's not actually that simple. I'll bitterly relate in a later post one of the things that can go wrong here. Put simply, the protein that confers resistance usually does so by destroying the antibiotic. After a few hours of growth, all of the antibiotic may be gone.. and thereafter, cells without the plasmid will grow marginally or significantly faster, so they can come to dominate the culture. This is more of a problem with some antibiotics than others; ampicillin, I'm looking at you. But, it's a general truth that after a few hours, most antibiotics will simply be gone, leaving you without ongoing selection. There be Darwinian dragons.
Finally, since we're talking about a teaching and amateur setting, there are many who feel (as I do) that it's best if people not use antibiotics for their projects. This isn't really because they might spread antibiotic resistance; after all, agriculture's already doing a great job spreading resistance directly to deadly pathogens. But, it's worth being principled here and warding off that possibility anyway. More relevant though is the difficulty acquiring, or the legality of using, antibiotics outside a medical setting.
IndieBB breaks the mould here by producing its own toxic protein, one which is not medically relevant, and which is not harmful or even significant to non-E.coli organisms, and simultaneously providing immunity to that protein. This addresses the question of using antibiotics in an amateur setting, and it saves users from having to prepare special antibiotic broth/agar media. In short, it makes things easier, it sidesteps legal or ethical swamps, and it's cool!
To be specific, IndieBB will take advantage of a class of microbial toxin called a bacteriocin, which bacteria use to kill their close cousins. Unlike medical antibiotics*, which are usually designed to kill distantly related organisms and don't affect the producer by virtue of that distance, bacteriocins ordinarily are as toxic to the producer as the victims. For that reason, they come with resistance genes baked-in. And, in contrast to the resistance genes used against antibiotics which generally destroy the antibiotic, the bacteriocin resistance genes usually merely passively block the toxin. After all, the producer doesn't want to destroy its hard work!
The bacteriocin to be used in IndieBB is called Colicin V, formerly known as Microcin V because it's (genetically speaking) really tiny. Whereas most bacteriocins, in common with antibiotics, require a great deal of DNA to produce (encoding many proteins that perform different steps in synthesis), the Microcin class require relatively little. In fact, the biggest contributor to the size of the Colicin V multi-gene-array (known as an “operon”) is the pair of proteins needed to shove the mature toxin out of the cell. The Colicin V protein toxin and the antitoxin are both very small indeed.
This means the Colicin V system can fit on a plasmid with room to spare for other stuff; projects that make medicines, perfumes, sensors, signals, DNA-literature or fuel. None of which will then require antibiotics to design, create, store or grow onwards.
*Technically, bacteriocins could be called “antibiotics”, and are by some. To avoid confusion, I'm following convention and using the word “antibiotic” to refer to broad-spectrum, non-peptide chemicals used to kill distantly related species, and which are potentially of medical use.
How to make a plasmid Pt3: Identifying Transformants
This is technically “optional”: if your plasmid is of a known size, and can be selected for easily, then you can just test the cells that survive selection for the presence of a plasmid of the correct size. Better yet, if you know the sequence (because you're using an open source plasmid, eh?), you can run a sequence-specific DNA copying reaction known as “PCR” to get a precise answer as to the presence or absence of your DNA in a surviving colony of cells.
However, for ease, many plasmids contain additional DNA that encodes a gene that can be used to identify the bacteria. For example, the plasmid may contain a gene encoding part of a protein called “Beta Galactosidase”, which when used by the bacteria while growing on specially formulated agar, will make the surrounding agar dark blue. Alternatively it may encode a fluorescent or pigment protein such as green fluorescent protein (as IndieBB does) or a derivative.
Because IndieBB is intended for use by people who may not know how to extract DNA, analyse it for size or run a PCR reaction, it encodes a gene for green fluorescent protein, so that transformed cells become fluorescent.
To be specific, it encodes wild type Green Fluorescent Protein (wtGFP), meaning the ‘natural’ form as found in the jellyfish A.victoria. This is because all enhanced forms of GFP remain patented, meaning that you or I are forbidden from using, improving, or sharing them with others without permission from countless hard to identify patent holders, who don't see innovation the same way we do.
This has two unfortunate consequences. Firstly, the absorption and emission spectra of wtGFP are less than ideal; while “enhanced” GFP can be made to fluoresce using a blue or UV-A LED torch, wtGFP only reacts strongly and distinctly under a UV-A strip-bulb, otherwise known as a blacklight. Thankfully, such blacklights are very inexpensive; small ones can be bought as note-checkers for testing for counterfeit money. They are also sold in some petshops as supplementary light for fish or reptiles. I have even noted a distinct green colour to colonies under full-spectrum fluorescent lighting, but I wouldn't rely on it.
The other consequence is that, fluorescence spectra aside, fluorescent yield can be poor with wtGFP, meaning the signal can be very poor, ordinarily. To solve this, I conducted an experiment using my own gene optimisation software, PySplicer, where I optimised wtGFP for expression in E.coli. The result was a gene that expressed very well in E.coli and gave clearly visible fluorescence; chalk up another win for Synthetic Biology!
How to make a plasmid Pt4: Editability
So by now, we have a plasmid that A) replicates, B) can be selected for, and C) makes cells visibly distinct from untransformed cells. But what's the point? As a beginner's project, completing the task of transforming cells is a worthwhile task in itself, and that's what IndieBB is primarily intended for.
However, that skill, transforming and modifying bacteria, is too cool only to use once. You need to be able to modify the DNA to do something else, to embed your own projects and build upon IndieBB in a freeform, flexible way.
Doing so requires, and will help you to develop, an expanded skillset, but it also requires stretches of DNA designed to facilitate editing. The traditional routes for DNA editing involve cutting, joining and pasting DNA using pre-defined sequences known as restriction sites, which can be cut by restriction enzymes (the use of the word restriction has a good historical reason).
A plasmid that ought to be editable using this traditional method should have such restriction sites in known, unimportant locations of the plasmid where they can be safely cut, additional DNA inserted in, and joined again without interrupting the other plasmid functions. Usually, these sites are clustered together into a location called a multiple cloning site (MCS).
However, there are newer methods for plasmid engineering that hybridise with cutting and re-joining, such as the use of the aforementioned DNA copying reaction, PCR, to copy just the parts of the plasmid or inserted DNA that are desired, sometimes with the addition of small custom bits of DNA to either site, prior to cutting and pasting. Newer still is the omission of restriction enzymes entirely, using flexible and rapid methods like Gibson_ Assembly_ that build upon the PCR reaction with a few accessory proteins, to allow the melting-pot assembly of multiple stretches of DNA together in one step.
To facilitate old and new methods, IndieBB will have as much of the above as practical. A classic (and out-of-patent) multiple cloning site from the venerable plasmid pUC18 will be used, with some added enzyme-cuttable sites where possible. To this will be added regions of DNA designed to be good targeting sites for the PCR reaction, and by extension the Gibson Assembly method. These aren't essential as you can target PCR to target most DNA anyway, but dedicated regions make the task more straightforward to accomplish and offer a de-facto standard for “hot-pluggable” DNA inserts, by me or third-parties, for Gibson Assembly directed use.
Because “Multiple Cloning Site” is so 2005, the dedicated-editable region of IndieBB will be termed a “Multiple Hacking Site” (MHS).
How to make a plasmid Pt5: Suggestions from Contributors
The above describes the original plan for IndieBB. Simple as plasmids go, but by virtue of its proven and well-understood components, it should do something new and useful; self-selection and self-maintenance, with a clearly visible fluorescent trait, using methods that can be accomplished in a kitchen.
During the crowdfunding campaign so far, though, I've had some additional suggestions from supporters that I've elected to include in the final design.
Firstly, from prolific biohacker Andreas in Austria, the suggestion to make each portion of the plasmid (Origin/Replication cassette, Colicin V operon, GFP Gene, Multiple-Hack-Site) entirely modular, so they can be easily separated and reassembled with other plasmids. The original suggestion was to allow use of restriction enzymes to cut them apart, but as I'd rather keep restriction sites unique and only present in the MHS, I think instead I will place dedicated primer sites for the PCR reaction. So, users can rapidly copy out the sections they want using pre-defined DNA primers, and use the Gibson reaction or primer-attached restriction sites to reassemble into their own projects.
Secondly, from Nathan McCorkle with input from Andreas, the suggestion to have a system to deliberately remove the plasmid from cells on-demand. There aren't many cases when you'd actually want to do this, but as we were discussing speculative scenarios anyway, the suggestion that you might “lose” your pre-transformation plasmid-free cells came up. This has actually happened to me! Well, kind-of. Still, it stirred interest on the list from other supporters, and who am I to say no? So, adding a (probably tetracycline-based) system that would switch off the Rep protein and the Colicin V toxin production genes would mean that the plasmid could be switched off on demand, leading to loss of the plasmid. If anyone suggests ridiculous worst-case-scenario cases in an attempt to scaremonger IndieBB, I can now say “it's ok, it's got a suicide system built-in”. Even though it doesn't need one, and it's only a crowd-pleaser!
There's my promised plasmid-design post.
In honour of the launch of FAIL BETTER at Science Gallery this week, I may next write up my previous experience, and how I succeeded and failed in equal measure, and learned a great deal from it. It's a gamble, revealing your failures; we all have them, but we mistakenly believe others don't. So, if I say nothing of my failures, many may feel encouraged, thinking that I'm some sort of perennially successful wonder-genius and my project is a sure bet. But, I'm gambling that revealing the fact that I'm not perfect will instead encourage people to see that I've had a chance to learn what not to do, and to give me extra credit in this project as a result.
Also coming up is a questionnaire for non-supporters; why didn't you find IndieBB interesting? What can I change about the pitch or the plasmid to reach a wider audience without compromising on ease of use, hackability, and freedom to develop?