summaryrefslogtreecommitdiff
path: root/transcripts/hgp-write/2017-05-09/human-artificial-chromosomes.mdwn
blob: 4c7ed7240b306d74ab1d1ddae738c48290af105a (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Human artificial chromosomes that shuttle the three kingdoms

Alina Chan, Harvard Medical School

<https://twitter.com/ayjchan>

<https://www.youtube.com/watch?v=qTUUkYzRVSI&t=0s&list=PLHpV_30XFQ8RN0v_PIiPKnf8c_QHVztFM&index=26>

I believe these are foundational technologies for synthesizing megabases of DNA and delivering the DNA into cells. Most significantly, in my process of building them, it's a brutal test of our knowledge of human biology. We can't build what we don't understand. We want HACs to shuttle between yeast, bacteria and humans. We want to do the cloning in bacteria and yeast where we are really good at doing that already.

We want to leverage existing technologies to use cloning methods that are known-- see Brown DB, Chan YA et al, NAR 2016. Frankenstein cells. A case study that we did is fuse yeast that carries a herpes genome and a yeast that carries the ebola protein, to human cells, to show that the human cell produces the infectious virus in combination with the ebola part and the other part. So this was just one example, but you could imagine a combinatorial toolkit.

We can build and deliver megabase -sized construct. We still need a HAC for shuttling from yeast and bacteria, and efficiently establish itself in human cells. Our HACs cannot hsuttle between yeast and bacteria yet.

Our HAC prototype is easy to manipulate in bacteria or yeast. 31 kb human artificial chromosome. Human mammalian cell selection markets, like a blastocyte resistant gene, GFP, etc. The most important part is the centromere which helps your newly replicated chromosome go into daughter cells during each cell division, otherwise you don't have a functioning chromosome. There are also epigenetic protein interactions with your DNA. And based on the knowledge that we have about human centromeres, we tried two different proteins, one is a histone-like protein, and the other is the chaperone, CENPA. HUURP chaperones, CENPA to centromeres. HUURP-lacI. We also had to insulate it to prevent epigenetic modifications from spreading.

Moving on to the centromere in more detail, it's an eentirely de novo centromere. It did not reply on repetitive replication, and it did not rely on excising from natural occurring centromeres. We used the 171 bp consensus sequence, any place whe could vary the nucleotides, we tried to remove repetition. There are two sites where there are lacO sequences which is what we use to target our public .. which we use to target our centormere... pUC-based vector, pCC1BAC-basedvector. It took many months to move this.

How do we characterize our HAC prototype? Right now it's only 30 kb, it's pretty small. I lipofectamine transfected into my cells. Used lacI-helpers. One of the first observations I had were that insulators are extremely important, because even after selection you have low rates of GFP cells compared to other variants. Controlling epigenetics is really important to HACs. Here's the stunning result, which is that compared to other variants, when you have the variants alongside HUURP-lacI. This goes on for months after selection. We think these are due to recombinants rather than independent HAC. We could take those HACs been in human cells for month, and shove them back into ecoli bacteria and check the sequence and see what has happened to the HAC. So right now we're analyzing that.

There are many more things we could do-- we're looking for collaborators to help us build new HAC variants. We're excited to be part of the international consortium. Also, we're distributing the HAC through Addgene as well. I have a lot of crazy undergrads working for me, and Clovis Basier, Sinem Saka Kirli, Thomas Westerling, David M. Brown, Elena Schafer, Michael Tschemer, Isaac Plant, Leah Bury, Natalia Koustseva, Matthew Solomonson, Mihael A. P. Karrebelt, Steen Thomas, Daniel G. Gibson, John I Glass, Jeffrey C Way, Ian Cheeseman, Pamela Silver.

# Q&A

Q: .. not enriched.. DNA binding properties.. maybe you give it some advantage because now you have a binding site..

A: Those are good points. Trying different variants of the centromere, we didn't have the bandwidth for that, because it was a super high risk project. Now that we have shown that it works, we can try other things.

Q: Did you check size of ... in human cells?

A: One thing I know is that when I shuttle it into bacteria, the size is still the same, and some of them have the entire sequence unaltered. Some of them seem rearranged, but the size is still 30 kb, so I'm unable to detect any changes through FISH or fluorescence methods.

Q: Epigenetic state changes?

A: I don't know that right now. Those modifications might be lost when moving to bacteria. I'm doing some experiments right now.