I'd like to address some of your points, but first I'd like to suggest
that saying "don't do that, everyone else is doing this" is a poor
reason to ward someone off of a project. E.coli is awesome, it's true,
but B.subtilis has long-recognised advantages in areas where E.coli lags.
So, if I want to do libraries and sequencing, I'll take E.coli. If I
want a high-risk project where I need my variables as minimally-variable
as possible, I'll take E.coli.
If I want long-term shelf-life or easily shipped, or simply tolerant of
abuse and forgetfulness, I'll take B.subtilis. If I want something
that's more accessible to an E.coli-averse public, I'll take B.subtilis.
If I want protein export from the cell (which is sometimes necessary for
correct protein folding, I'll point out), I'll take B.subtilis, because
E.coli protein export beyond the periplasm is pretty damned awkward.
This is biology; relative species advantage is the name of the game, and
for some things B.subtilis is more attractive than E.coli.
For that reason, long back I advocated B.subtilis for amateur biotech
because it's easier to handle and grow (I don't notice a *practical*
difference in growth rate btw; it's there but is rarely an issue), it
survives real life amateur conditions (forgotten, mishandled, stored in
a shoebox for years), and it's naturally competent. As a store of
genetic material, it's great in theory; without a -80, E.coli stocks can
be very unreliable, but B.subtilis can be stored on card on a shelf.
Again, for practicality, B.subtilis is great for amateur use.
Since then, I've found protocols that help make E.coli more practical,
too; better transformation systems for amateur environs, but B.subtilis
still has advantages.
My attempt to make a B.subtilis plasmid that would be practical for
amateurs was a partial success, but I didn't have the time or money to
make it a full success, so it's shelved for now. For reasons you
suggest, I am working on a lower-risk project right now in E.coli, but
that doesn't diminish my interest in providing a B.subtilis plasmid
someday, perhaps soon.
Anyway, by heading:
> 1. *B. subtilis* might be autocompetent but homologous recombination
> from what I read requires overlapping regions of > 500bp [...]
Yes. Homologous recombination in most species requires long tracts, and
this is no shorter AFAIK in E.coli or Yeast. The advantage of B.subtilis
is that it has an *active uptake and recombination* system, rather than
passive, so actually you can make do with shorter tracts and get
reasonable efficiency. This was never a target of mine so I don't
remember how low the homology can go.
Remember also that for gene erasure or replacement, the overlapping
regions are part of the target region, so it's not so burdensome. If you
want to introduce new DNA, then having to add up to 500bp extra is
harsh. For deleting DNA if all you need is 500bp, that's cheap. More on
this below, where deletion is relevant to strain improvement..
> 2. Under optimal growth conditions the doubling time of
> *B. subtilis* is always slower than *E. coli*. *B. subtilis*
> is measured to be somewhere around 50-120min
This is true, and I'm not disputing that for the same *cell mass* this
is a problem. However, I can only say that in my experience, this rarely
impinges upon the ease of working with B.subtilis. Part of the reason is
probably that colonies in B.subtilis are more sparse, so if you're only
seeking colonies and not a large cell mass, B.subtilis effectively
competes with E.coli because colonies as fast by volume if not mass.
Mass is important for protein or DNA yield, so for miniprep DNA yield,
this is still an issue, yes.
Bottom line is that for my purposes, which probably overlap well with
Mega's and many others', growth rate was rarely an issue when working
with Bacilli, for me at least.
> 3. Transformation efficiency as you have stated and as I have read in
> other papers is staggeringly low in *B. subtilis*, on the order of
> 10^5 /ug of DNA at the extreme high end
Pretty sure I recall seeing it up at 10^6/7 sometimes, but yes; getting
high efficiency is harder than with E.coli. IIRC (and I probably don't),
those protocols I saw were chemical methods that involved making
spheroplasts, so let's chalk those up as "not applicable".
So yes, as you well point out, you don't want to do libraries for
sequencing or in-vitro mutagenesis in Bacilli; go with E.coli. Without
taking away from that fact, consider that for sequencing, libraries are
of decreasing relevance, and for mutagenesis there are in-vivo
alternatives if you really want to work with Bacilli.
If you only have a tiny amount of DNA with which to transform, for
example with a quickligate or what-have-you, then I'm with you here;
don't risk your tiny amount of DNA on B.subtilis if you can't afford to
However, again on the practicalities; most use-cases of a competent
bacterium, I find, do not call for 10^6/7/8/9. 10^3 might be a bit on
the low side, but frankly I'm impressed that just throwing DNA at
growing cells worked out for Andreas! B.subtilis 168 (the "type" lab
strain) is widely known to be extremely easy to hack, and I had someone
say to me that you can practically throw DNA at it but I thought they
Anyways, I never quantified my transformational efficiency because of
the particulars of the selection system I was trialling, which was
positive rather than negative selection. I ordered a conventional
negative selection plasmid but failed to transform it into an E.coli
stock prior to use in Bacillus and lost the DNA. So, deduct a point from
E.coli, I guess. Takeaway point; I got plenty of transformants using my
sporulate-then-germinate-with-added-DNA method. Bonus; that's the
protocol, it's really that simple. No steps or nothin', just add spores
and DNA and plate. :)
> 4. According to the NCBI database there are being worked on or have
> been completed 237 projects to sequence *E. coli* genomes
Forgive me, but isn't this an unfair comparison? E.coli genomes will
outnumber probably every other species, so you could use this argument
to suggest that E.coli is a more important organism than Humans, for
That said, let's address a non-obvious factor you ought to consider even
if you apply weighting by publication count, genome size, etc. to figure
out the relative sequence coverage of Bacillus: species-level balkanization.
This was a really annoying issue for me when I was designing my plasmid,
because one of my design constraints (which may help account for your
questions as to my bizarre methodology) was that all genetic material on
the plasmid should be of B.subtilis origin only. This is because EU law
regards cisgenics as "not Genetically Modified", in a very explicit way.
So if my plasmid were useful and only consisted of B.subtilis, I could
sell it to unlicensed people in the EU and they could legally use it.
I discovered that, whereas E.coli is a pretty diverse species name,
B.subtilis had been diced up into lots of separate species in the
nineties. Being honest with you, I never bothered to align E.coli
genomes and pre-balkanisation B.subtilis genomes to see whether they
were justifiably different, but it struck me that there was a
fashionable trend of renaming B.subtilis strains as new species for a
while there, and it hindered my work because some of the more pigmented
strains became legally "new species" and therefore pigments were
off-limits on my plasmid. Boo. Funny how humans can so quickly rewrite
Nature, right? Who needs to actually do biohacking to make new species
when you can just join a classification committee and make new species
with the stroke of a key? :)
> 5. Many proteins don't unfold/refold well, as occurs during
> secretion. Many protein have cofactors or require chaperones
> not found outside cells. This makes secretion only a slightly
> useful tool.
Here I'm going to mostly agree with you. On the one hand, you can't take
the average protein and just shove it out of the cell and expect good
results without some trial and error, mostly error. On the other hand,
B.subtilis' ability to export functional protein is oversold, because
the average strain also exports loads of proteases in late growth/early
Here's where I catch my above reference to strain improvement. There is
prior work in removing the protease genes in order to improve export
capability of Bacilli, and to add in copies or modify-in-place the
export machinery to improve throughput (usually just by increasing the
amount of export machinery in the cell to remove bottlenecks). Without
these improvements, the advantage of export isn't as big a selling point
over periplasmic export in E.coli followed by cell rupture.
However, it's a *nice feature* to have. E.coli is awful for export, as I
learned during my ill-fated protein purification project. There are two
options for periplasmic export, but no single option for export beyond
that; E.coli proteins that are fully exported seem to each have their
own system for export. Some of these systems appear somewhat modular,
but from what I could see the modularity was hit-and-miss. Some proteins
could be exported with a given protein's machinery, but not others.
If, for some reason, export from the cell is critical to your project
(and it might well be!), B.subtilis seems to have the upper hand, and
there's not much in E.coli's toolbox that can compete. This isn't
because B.subtilis is amazeballs at export, it's simply because they
have the same export machinery in common, but E.coli wrapped a periplasm
around it all. So B.subtilis' homologues of E.coli periplasmic export
just drop the protein into the extracellular environment.
> 6. <You mentioned plasmid instability in a prior email I think>
Here's a virtual point I want to address, a partial myth. Yes:
B.subtilis is known to have issues with plasmid stability in the
literature. However, some reading on this yielded that this instability
can be traced to a consensus sequence believed to be associated with
TopoI in B.subtilis. Excluding that consensus sequence from your designs
is easy with tools that accept IUPAC notation (PySplicer, for example!),
and should address the instability problem. RecA homologues etc. may be
a problem for some projects but I hear little complaint about B.subitlis
168's stability outside of the TopoI issue.
And again, I worked with B.subtilis for only one major cloning project,
so I can't pretend to be an expert. Just pointing at the TopoI
connection as an example of a problem remaining more widely known than
On 04/01/14 15:54, Josiah Zayner wrote:
> *E. coli* has so many powerful genetic tools it is a pity not to use it.
> I have seen multiple single papers with more edits in the* E. coli* genome
> If one is going to "express" recombination competence one can do that in *E.
> coli* also. And use "PCR products" to recombine into the host genome. See
> Wanner strains (
> Expressing proteins to purify is not so simple as you say, just secrete
> them from cells. And overexpressing proteins rarely kills cells. Though I
> am sure it might actually happen this is rarely/never a case Scientists
> worry about. Protein folding/aggregation and intrinsic toxicity is of way
> more importance. Expressing something like a restriction enzyme that would
> damage the host DNA would be a case in which *B. subtilis* would be useful
> but general protein expression I have my doubts. Though I have no evidence
> for or against.
> Your enthusiasm is great but if *B. subtilis* is so amazing why don't most
> Scientists and companies use it as their genetic tool of choice? Have the
> majority of Scientists in the world for the past 50? years been wrong and
> you and Cathal are correct?
> If I had to guess why people don't use *B. subtilis* I would say:
> 1. *B. subtilis* might be autocompetent but homologous recombination from
> what I read requires overlapping regions of > 500bp(
) (if you could point me
> to a paper that shows less is required without a complicated strain of
> complex manipulations that would be cool). This adds complicated work to
> any design and multiple cloning steps. *E. coli* homologous recombination
> can be done with ~40bp which means it can be added to the end of a cheap
> primer and the gene created in a single PCR step. Further, needing to
> express comK is similar to needing to express a protein in *E. coli* for
> homologous recombination making it no more useful.
> 2. Under optimal growth conditions the doubling time of *B. subtilis* is
> always slower than *E. coli*. *B. subtilis* is measured to be somewhere
> around 50-120min(http://jb.asm.org/content/167/1/219.full.pdf
). While *E. coli* is consistently
> measured to be 20 minutes or less(http://jb.asm.org/content/180/3/547.full
> This is very significant considering it would take triple the amount of
> time for the ~same cell mass.
> 3. Transformation efficiency as you have stated and as I have read in other
> papers is staggeringly low in *B. subtilis*, on the order of 10^5 /ug of
> usually <=10^3. *E. coli* on the other hand is around 10^10 or greater /ug
> of DNA generically. When creating genetic libraries or performing simple
> transformations this is highly significant and greatly restricts the use of *B.
> subtilis*. A typical ligation reaction I perform has around <=100ng of
> final product that would yield <=1 colony. I know regularly when cloning I
> often only receive a few colonies of *E. coli* DH5a or DH10B cells when
> transforming a quickchange or a ligation, this would likely yield zero
> colonies in *B. subtilis *and cause much more work.
> 4. According to the NCBI database there are being worked on or have been
> completed 237 projects to sequence *E. coli* genomes(
> subtilis* currently has 23 projects total (
> genetics/metabolism &c. This makes the longterm viability of *E. coli* much
> more inviting. It is like using an open source package that has 23 versus
> 237 people working on it. Which one do you think will have all the snazzy
> new features and still be around in 1, 2, 5, 10 years? Further, according
> to Google Scholar in 2013 ~190,000 papers contain the term "e coli" and
> ~26,000 contain the term "b subtilis"(no quotes used in actual queries).
> 5. Many proteins don't unfold/refold well, as occurs during secretion. Many
> protein have cofactors or require chaperones not found outside cells. This
> makes secretion only a slightly useful tool. Searching for optimal protein
> expression yield I couldn't find data for *E. coli* and *B. subtilis* in
> which they used the same protein and conditions so I don't know which is
> better for protein expression.
> I agree *B. subtilis* has its positives. This discussion has even made me
> consider using it for one of my projects that requires protein secretion
> but I don't think you are going to convince me or most Scientists that it
> is better than *E. coli *as a genetic or synthetic biology tool.
> I am not saying to stop your work I am just suggesting that if you want to
> use *B. subtilis* you should focus your work on the advantages it has over *E.
> coli* instead of attempting to compete with an already established
> genetic/synthetic biology tool.
> Josiah Zayner, Ph.D.
> Synthetic Biology Group
> NASA Ames Research Center
> On Friday, January 3, 2014 9:26:11 PM UTC-6, Koeng wrote:
>> Bacillus subtilis is not a "new species"... It is the second most studied
>> bacterium in the world and has more information on its genetic structure
>> then E coli. It has the most complete minimal gene set of any living
>> First of all, remember that *plasmids are NOT being used*. So yes, they
>>> *Bacillus*. I assume the strains of* Bacillus* are not endA- and
>>> recA-(meaning your DNA and the DNA of the host have a chance of being
>>> screwed up) and you probably cannot do blue white screening with them, a
>>> useful genetic tool.
>>> Say you want to express a protein with a T7 or T5 promoter are you just
>>> going ask someone for the prophage and lysogenize these strains? Genetic
>>> editing is probably much more difficult without Warner strains and such. Do
>>> you know if *Bacillus* can produce comparable amount of protein compared
>>> to* E. coli* using T7/lac/IPTG? Does *Bacillus* restrict non-methylated
>>> DNA? Would make working from PCR/synthesized products difficult. Seems like
>>> an interesting hobby to test these things but Scientists probably will not
>>> use *Bacillus* just because it is slightly better than *E. coli* as *E.
>>> coli* has been used and tested and strains have been perfected for
>>> 10s/100s of years? Maybe *Bacillus* would have been the correct choice
>>> to start with cloning 100 years ago but it is kind of too late. Remember
>>> you are fighting against years and years of research and
>>> optimization(probably more than a single lifetime) with the
>>> thousands/millions of strains of *E. coli* it doesn't seem like the