Fwd: telomerase as a biological typewriter

[Bryan Bishop]

---------- Forwarded message ---------
From: fenn <fenn...@gmail.com>
Date: Sat, Nov 19, 2022 at 9:47 PM
Subject: Fwd: telomerase as a biological typewriter
To: Bryan Bishop <kan...@gmail.com>, Max Berry <maxb...@gmail.com>

sent after biohacktheplanet but it never made it to the enzymatic synthesis list for unknown reasons.

maybe you'll find the ideas interesting, maybe not. it sounds like the telomerase 6-mer sequence (GGTTAG) can't be changed on the fly. her scheme could be used for data storage at least, assuming you had a sequencer that can handle the repetitive sequences. perhaps the telomerase template string could be changed to one that is compatible with a scar-free deletion process, leaving just the variable nucleotides.

Forwarded Conversation
Subject: telomerase as a biological typewriter
------------------------

From: fenn lipkowitz <fenn...@gmail.com>
Date: Sun, Aug 27, 2017 at 3:10 AM
To: <cpa...@ucsc.edu>
Cc: <enzymatic...@googlegroups.com>, <fenn...@gmail.com>

hello christina,

i saw your presentation about telomeres at biohack the planet, and
i've been dreaming about how to synthesize new DNA in a cellular system
since high school.

the original in-vivo DNA synthesis idea was to have a light driven
polymerase with 2 or 4 different light wavelengths to grab hold of a
specific nucleotide, and another wavelength to allow the attachment of the
nucleotide to the growing strand and drive the enzyme forward. 3 or 5
"antenna" chemicals in the optical range would tug on the protein so that
it assumes a conformation that would perform the desired function, whether
that's grabbing a base or attaching it. you control the output DNA
sequence by flashing lights in that sequence. very simple.

but it's become clear that we don't have the computing power to simulate
protein folding quickly enough to do rational protein engineering, and
not many people seem to be working in this field. so instead we have to
find an enzyme in the wild that does what we want, and use that.

one way to do it* with existing enzymes would be to ligate short fragments
of DNA together with ligase one molecule at a time, but in order for
ligase to work you need at least 5 base pairs of overlap, which means the
short fragment is at least 10 bases long. there are 4^10 = 1048576 or
about a million different permutations of ten base pairs. you can create
each permutation with a traditional bulk chemistry DNA synthesizer, then
put them in a microfluidic linear storage array - basically water drops
full of DNA in a narrow tube, separated by oil. you drive the water drops
back and forth while electronically counting the number that have passed
until you find the right droplet. then break off part of that droplet with
suction and add it to the microfluidic reaction chamber with the growing
strand. the oil and water never mix, so the droplets will stay separated,
as long as everything is working as planned and you never get any bubbles
or broken droplets.

only a single short fragment is added to the growing strand by ligase
because it is extremely unlikely that the other half of the 10-mer that is
hanging off the end is exactly the complementary sequence to the first
half of the 10-mer.

AGCTATCGAT   <- extremely unlikely 10-mer
TCGAT        <- growing strand

AGCTATCGATAGCTATCGATAGCTATCGATAGCTATCGAT  <- ligase repeats
TCGATAGCTATCGATAGCTATCGATAGCTATCGAT

you can predict in software when this will happen. just start adding bases
to the start, to shift the 10-mer frame until it doesn't happen anymore.

TTTAGCTATC   <- no more palindromey thing
AAATC

this whole scheme is a lot of work, you have to synthesize ONE MILLION
sequences, and then corral them around without messing up, ever. but it
could work.

--------
new idea

but you said human telomerase copies 6-mers from an RNA strand, that's a
whole lot less than one million permutations, only 4^6 = 4096
permutations, or 4096*6 = 24576 bases long.

what if you had a synthetic circular RNA strand, 24.5kbase long and
containing every possible 6-mer permutation, that shuffled back and forth
through the template spot in the telomerase molecule, instead of the oil
droplets? the strand is dragged forward a base at a time by a light-driven
inchworm enzyme. (or even better, six bases at a time. i'm sure we could
come up with an enzyme that already does this inchworm motion; it's so
simple.) then the telomerase would copy the current 6 base sequence in the
template spot to its growing strand. by driving the circular strand around
and around, and somehow inhibiting telomerase until the right spot, you'd
end up copying any sequence you wanted onto the end of a DNA strand.

some problems: how do you get the 24.5kbase strand to the right spot in
the telomerase? i don't know how the telomerase knows what particular RNA
to use as a template in the first place.

since it has every possible 6-mer permutation, maybe it gets transported
somewhere unfavorable, or degraded by site recognition enzymes, or covered
in protein, or tangled up in itself. but there's 4096 factorial different
ways to lay out the circular strand, so these scenarios are avoidable with
careful ordering of the permutations. permutations of permutations.

maybe one of those old fashioned embossing label maker wheels is a better
metaphor than a typewriter.

   - fenn

cc'd to enzymaticsynthesis because it's the right place

*i don't know who originated this microfluidic permutation library method

----------
From: Mail Delivery Subsystem <mailer...@google.com>
Date: Sun, Aug 27, 2017 at 3:10 AM
To: <fenn...@gmail.com>

Hello fenn...@gmail.com,

We're writing to let you know that the group you tried to contact (enzymaticsynthesis) may not exist, or you may not have permission to post messages to the group. A few more details on why you weren't able to post:

 * You might have spelled or formatted the group name incorrectly.
 * The owner of the group may have removed this group.
 * You may need to join the group before receiving permission to post.
 * This group may not be open to posting.

If you have questions related to this or any other Google Group, visit the Help Center at https://groups.google.com/support/.

Thanks,

Google Groups

----- Message truncated -----

----------
From: Christina Palka <cpa...@ucsc.edu>
Date: Sun, Aug 27, 2017 at 10:38 PM
To: fenn lipkowitz <fenn...@gmail.com>
Cc: <enzymatic...@googlegroups.com>

Fenn!

I'm really really excited about your idea. I've been thinking about it all day. Let me do a bit of reading and get back to you with a detailed response.

Let's build ourselves a biological keyboard/typewriter/label maker!

Christina

----------
From: Christina Palka <cpa...@ucsc.edu>
Date: Mon, Sep 4, 2017 at 2:42 PM
To: fenn lipkowitz <fenn...@gmail.com>

Hi again Fenn,

Sorry about the delay. It's be a hectic week...

So I'm not super familiar with the microfluidic permutation library method but I feel like it could be the solution to a problem that I was never quite able to figure out with regards to my telomerase typewriter idea. I'll give you an overview of what I was thinking, the conceptual problems I've encountered, and we'll see where we can go from there :)

I'm not sure why but my emails are bouncing from your enzymatic synthesis group so feel free to forward this info onto them if you want.

So telomerase is a reverse transcriptase that uses an RNA template to make DNA. Telomerase is unique among reverse transcriptases in that it re-uses the RNA template to make iterative DNA repeats (telomeres). The process of re-using the template is called Repeat Addition Processivity (RAP) RAP consists of a few steps: 1) Recognizing the end of the template 2) Breaking the DNA:RNA hybrid that was just synthesized 3) Maintaining contact with the DNA stand while the RNA template and protein machinery resets itself at the beginning of the template 4) Continue adding nucleotides till the end of the template (Nucleotide addition processivity) 5) Repeat.

A lot of the details with regards to how RAP is achieved aren't known. It's actually what I'm researching for my PhD thesis. However, enough is known that makes me think telomerase could actually be used effectively to create de novo DNA sequences. First is how the enzyme binds with different affinities to different lengths of the DNA it's making.

The human telomere sequence is TTAGGG. However, in the actual enzyme the end of the template and DNA repeat sequence is GGTTAG and the complementary RNA template sequence is CAAUCCCAAUC. You can see how in this way the final five nucleotides are complementary with the beginning of the sequence (GTTAG on the DNA matches with CAAUC at the beginning of the RNA template). Given the template is 11 nucleotides and 10 of them are involved in complementary pairing of DNA synthesis processivity, you only have 1 nucleotide to play with. My idea was to create 4 templates, altering that middle nucleotide with either a C, G, A, or U. 

*Sorry they don't exactly line up in register*

Four templates: 3'-CAAUCCCAAUC-5', 3'-CAAUCGCAAUC-5', 3'-CAAUCUCAAUC-5', 3'-CAAUCACAAUC-5'

DNA created:    5'-GTTAGGGTTAG-3',  5'-GTTAGCGTTAG-3',  5'-GTTAGAGTTAG-3',  5'-GTTAGTGTTAG-3',

These nucleotides would be the ones to be analyzed and use in the alphabet created. Simple scripts could get rid of the repetitive sequence leaving you only with the middle nucleotide and using a three letter code similar to the amino acid codons could give you A - Z. I'd always imaged the RNA template being moved by some sort of light activated protein or something, but honestly I haven't thought through that aspect of it in great detail which is why I'm interested in your microfluidic idea. 

In my mind, telomerase would be introduced to a template, and create 1 register of this sequence, stop (there are a lot of mutations that have been characterized that are attempting to dissect different amino acids involved in the repositioning of the DNA and RNA template during the RAP process), a new template would be introduced, and the new end of the DNA would reanneal to the new template, and the DNA strand would be extended again.

I'd kind of imagined the sequencing of the DNA that was synthesized by the Nanopore or something as it has the potential to get very long reads and I think it does well with repetitive DNA sequences. But again, this aspect of it isn't terribly well thought out.

Making telomerase is challanging but it can be made in small quantities. For this work, I was imagining DNA being created at a single molecule (or at least very small quantities) so that should be ok. Our lab does single molecule work on the system so I'm familiar with the in's and out's of the system :)

Problems to sort out:

Sequencing repetitive DNA sequences is hard

Not super clear how to introduce the template to the telomerase molecule

Engineering a set up in which a telomerase molecule would be 

Telomerase has an exonuclease activity that isn't well characterized

Exactly how RAP is achieved remains not well characterized 

Interesting sequencing ideas to consider:

Zero-mode wave guide : http://science.sciencemag.org/content/299/5607/682

Looking forward to your thoughts!

Christina