[DIYbio] Full Disclosure on "Open Source" Protein Purification Experiment

-----BEGIN PGP PUBLIC KEY BLOCK-----
Version: GnuPG v1.4.12 (GNU/Linux)
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=wOG4
-----END PGP PUBLIC KEY BLOCK-----
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
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=xCjP
-----END PGP SIGNATURE-----
Hey Nathan,

On 22/02/14 22:41, Nathan McCorkle wrote:
> Actually Cathal, why aren't you doing a campaign for the cheap resin
> protein extraction? You'd have a lot more appeal to lab folks who
> need a new workhorse.

Because it didn't work, at least the 3 times I tried (burning most of
what remained of my funds in the process). I did feel that had strong
potential to attract professional buyers (benchtop scientists, small
producers, DIYers who wanted to ferment their own lab-enzymes, etc.), if
successful, but..

I may as well document my failures and make them unpatentable in case
they'd work under different circumstances. I've already planned a post
about the original plasmid effort, so I'll spare "full disclosure" on my
bacillus plasmid until then.

At least as far as the resin extraction experiment, here was the idea:
* Agar is cheap, abundant, easily purified almost everywhere. Most labs
have purified, desulfonated agar and agarose already.
* Plenty of agarases have been studied well, some have been divided into
cleavage and binding domains.
* Other carbohydrate binding domains from the same (very broad) family
are already in use for protein purification; Maltose-binding protein,
for example.

Digging around, I found that for some reason nobody had thought of using
Agar for protein purification.

Let me put that in perspective. There are patents out there for using
artificial calcium-binding peptides, hair-binding peptides, and
cellulose binding peptides. But nobody had talked much about using
agar-binding peptides or natural protein domains, and I couldn't see any
patents on the application of agarases to protein purification.

So, my task was to find and attempt to use an unpatented agarase as a
protein purification system, and I found one in the AgaD agar-binding
domain. Many agarases are patented, most of them down to 10% or 5% amino
sequence identity (which, if you're unaware of the significance, is an
absolute absurdity). AgaD was not patented, and was not apparently being
studied with commercial intent. I found a paper that successfully
divided the domain into carbohydrate-digesting and merely
carbohydrate-binding domains, and successfully got the "Carbohydrate
Binding Module" (CBM) to bind to agar at 4C in a totally simple buffer.

So, I fired up PySplicer, optimised a fusion of wild-type GFP to the CBM
(with a cleavable peptide in between) with an E.coli periplasmic
tat-export signal peptide (sec export doesn't usually work with GFP as
it can't re-fold in periplasm) and ordered from Genscript. Received, had
difficulty determining whether the fluorescence I was seeing was any
different from E.coli autofluorescence, and could not get any detectable
purified fluorescent periplasmic fraction. However, I would later learn
(d'oh) that while I was using a UV-A LED lamp, wild-type GFP doesn't
respond well to that, and you need a blacklight (as in a UV fluorescent
lamp).

Tried again, this time with GFP and the CBM in reversed order. Didn't
work on first try, tried again with a UVA blacklight *by accident* (I
only bought one so I could compare wild-type to transformants
side-by-side to distinguish autofluorescence), and found that actually
the optimised GFP was *totally clear* under a blacklight. Yay, PySplicer
works! Sadly, the CBM didn't appear to work. I could extract fluorescent
cytoplasmic and periplasmic fractions, which is consistent with
Tat-export of highly expressed proteins (it bottlenecks and you get most
of the protein in the cytoplasm; for "production" you've got to add
extra copies of the tat system), but neither would bind agar/ose under
any tested conditions.

Tried a third time, using a brighter GFP variant to enhance detection of
possible weak binding, and removing the TatA export signal so I could
focus only on the binding issue without worrying about other unknowns. I
could extract cytoplasmic fractions that were pleasingly fluorescent,
but did not bind agar/ose under any tested conditions.

Sounds real quick, but that took more than half a year, and yielded only
one positive result; that pysplicer worked great at optimising the GFP
domain, at least! Also, never order from genscript within 3 months of
the iGEM jamboree. I got dropped to the bottom of the pile despite
initially paying double for quick shipping, and ended up waiting 4
months for my DNA; 4 months of living expenses and no experiment to
complete! Kept being told "next week".

The feedback I get from CBM experts is that they're temperamental, and
sometimes cease to work when fused to other proteins. Maltose-binding
protein appears to be a nice exception, generally not caring what you
fuse it to. But in my case, the AgaD carbohydrate binding domain appears
to be a "lone wolf" CBM, and the experiment had to be written off.

After all that, I have very little funding to try any of my other
speculative ideas, so I've scaled back to something I'm surprised nobody
else has attempted yet; the "beginner's biotech" kit, with a custom
plasmid designed for DIYers. Perhaps I should have gone out with a bang
and tried one more wacky idea; still considering that if the crowdfunder
doesn't work, but all of my current "wacky" ideas are far too expensive
for the funds I have left, and will certainly take months to test.

Anyways, I'll upload a git repo soon containing a cleaned-up version of
my working directories for the three drafts (I used my awful DNA design
scripting language, the original fastac, for this; it's not pretty). If
anyone out there notices a glaring mess-up or omission that might
warrant a fourth attempt, I'd certainly consider it..

best,
Cathal

PS: "Obvious to one skilled in the art" here includes the use of *any*
agar, agarose, agaropectin or other carbohydrate binding domains or
whole proteins as a fusion partner to allow cheap affinity purification
of proteins. Also "obvious" but too expensive for me to attempt is
optimisation by rapid mutagenesis of the protein domains or the spacer
of aminos joining the fused domains, in order to obtain a CBM:GFP
pairing that does work. Also "obvious" is continued attempts to optimise
a buffer or other experimental conditions that would lead to a
functional fusion. Please don't steal my work from the public domain by
patenting it.

On 22/02/14 22:41, Nathan McCorkle wrote:
> Actually Cathal, why aren't you doing a campaign for the cheap resin
> protein extraction? You'd have a lot more appeal to lab folks who need a
> new workhorse.

--
Please help support my crowdfunding campaign, IndieBB: Currently at
26.8% of funding goal, with 19 days left:
http://igg.me/at/yourfirstgmo/x/4252296
T: @onetruecathal, @IndieBBDNA
P: +3538763663185
W: http://indiebiotech.com