[BBF Standards] Tom Knight RFC (plain text)

[Drew Endy]

Just a quick resend of Tom Knight's RFC as plain text.

******

Request for comments

Biobrick assembly standard modifications

8 July 2008

Tom Knight


Background:

Over the past several years the original Biobrick assembly standard  
has proven to be a useful DNA assembly technique.  Despite this, a  
significant flaw has been the composition of the mixed base scar, T  
ACTAGA G.  Since this scar is 8 bp long, it makes protein fusions,  
aligned on three base codon boundaries, quite difficult.  Ira Phillips  
at the Silver Lab worked around this problem by ignoring the flanking  
T and G sites (inserted for protection against methylation issues) and  
using the mixed site ACTAGA.  This resulted in the amino acid sequence  
Thr-Arg inserted into fusion protein designs, two amino acids with  
significant chemical difficulties in many contexts.

Chris Anderson at Berkeley worked around this problem in a different  
way, by adopting a  new restriction enzyme set, BglII (prefix, AGATCT  
site) and BamHI (suffix, GGATCC site).  These enzymes are insensitive  
to methylation, and produce a scar GGATCT (Gly-Ser).  The Gly-Ser  
amino acids are near ideal for most protein fusion work, and the  
enzymes are cheap and effective.  Unfortunately, neither of these  
enzymes can be heat inactivated, making automated assembly with them  
substantially more difficult.

Another difficulty with these enzymes is the frequency of the BamHI  
and BglII sites in many natural DNA sequences.  For example, in the E.  
coli genome the BamHI average fragment length is 9,000, while the  
average fragment length of XbaI fragment is 120,000.  This reflects  
the relative rarity of the CTAG sequence in E. coli genomic DNA (for  
reasons poorly understood).  The high frequency of sites causes two  
problems.  First, making new Biobricks from existing genomic DNA  
becomes substantially more difficult.  Second, the frequent occurrence  
of these sites in contaminating genomic DNA in minipreps results in  
short fragments which can replace desirable parts in assembly  
reactions, yielding incorrect products.

Proposal:

Two additional restriction enzymes exist with a CTAG overhang: AvrII  
(CCTAG site) and NheI (GCTAGC site).  AvrII cannot be heat killed, and  
produces poorer codon choices than NheI.  I propose that we  
restructure the cloning site and flanking sites of Biobrick parts with  
the following structure:

……<EcoRI>……<SpeI> Part <NheI>…..<PstI>…..

The part would be flanked by bare SpeI and NheI sites.  The mixed site  
formed by assembly of these fragments, using standard approaches,  
would be GCTAGT, coding for Ala-Ser.  The Ala-Ser amino acids are  
almost as fusion-friendly as the Gly-Ser of the Anderson fusion  
technique.

The NheI enzyme can be heat killed, and thus is more amenable to  
automated assembly processes.

The rarity of the NheI site in E. coli genomic DNA means that many  
fewer fragments accidentally cut from genomic DNA contamination of  
minipreps will clone in place of the desired part.

Transition issues:

We would need to construct new cloning vectors with the new cloning  
site.  Parts would need to be recloned into the new vectors, probably  
using PCR with new primers.  Manual assembly of parts mixed between  
old and new formats would likely be possible in many cases as an  
interim solution, since the parts retain a common CTAG overhang.

We should rethink the use of the EcoRI enzyme for the prefix outside  
cutter.  There are likely more robust enzymes usable.

We should rethink the need/desirability of the NotI sites between the  
outside and inside restriction enzyme sites.  Some DNA fragment is  
necessary there, but it need not be that sequence, and the two  
sequences need not be identical.

Plan:

1)    Circulate this document for comments and blunder stopping

2)    Analyze the frequency of NheI sites in existing registry parts

3)    Test the efficiency of NheI and any other recommended enzymes

4)    Test for the ability to heat kill the enzymes

5)    Design automated programs to assist in the primer design for  
transition

6)    Choose which parts are worth transitioning

7)    Design desirable part collections for protein fusion work