[BBF Standards] Fwd: new assembly standard proposal

[Raik Gruenberg]

Hi all,

sorry for the late response to Chris' last mail about the assembly format. If 
you haven't done it yet, please have a look at my table comparing the different 
formats:
http://tinyurl.com/5ch4qa

I agree with Chris that all the new format proposals are solving the issue of 
protein scars pretty well. There are some differences when it comes to 
preserving the native N-terminal of a protein in a way that it can be combined 
with independent RBS parts. There also appear to be some differences in the 
efficiency and background frequency of the chosen restriction sites and I cannot 
judge how important these differences are.

I would also agree with Chris that backwards compatibility is not the most 
important feature *IF* we assume that the new standard is good enough for 
wide-spread adoption and will serve us well for several years to come. Here, 
however, Chris is raising a very important issue -- the current assembly 
chemistry may not fit this shoe:
(1) Even after several attempts, it seems to stubbornly defy reliable 
automation: perhaps not surprising, since we have to shuttle DNA in and out of 
E. coli several times.
(2) It is slow: assembly of 5 parts requires 3 cloning rounds taking between 
(ambitious) 7 and (more likely) 10 working days.
(3) It leaves a scar...

There are of course also many advantages to the assembly method but the 
shortcomings are serious enough to scare off potential users and motivate the 
new institutes and even some iGem teams to explore altogether different 
strategies. In order to justify a full registry conversion, a new format would, 
in my opinion, need to fulfill either of the two following criteria:
(A) support a new fast and automation-friendly assembly method that (almost) 
everyone is happy with, or...
(B) or support the "upgrade" of a part into new prefix/suffix flanks in one 
simple and efficient cloning step.

Since we have neither A or B, we could simply declare defeat and leave things in 
the chaotic state that we now face. However, we could also look for a mid-term 
solution that, at least, solves the pressing protein fusion issue and buys us 
some time for tackling the larger problem. The Freiburg format is such a 
solution and has the unmatched advantage of being a backwards compatible BBa 
extension rather than a full-fledged new format.

My suggestion would thus be to:
* recommend the Expression part (aka Freiburg) format as BBa extension (pBBa?, 
BBx?) for those teams that are working with protein parts
* still recommend the classic BBa for everyone else
* encourage everyone to explore and document alternative assembly schemes
* call for a collective brain storming to solve problems A or B

BTW, a solution for criterion B would not only open an upgrade route but also 
offer some very interesting new assembly shortcuts (like the method from Wendell 
Lim's 2007 UCSF team).

On a practical note, I should soon be able to submit the three pSB1A*3 
construction plasmids with Freiburg flanks to the registry (pSB1AC3 and AK3 are 
ready, AT3 under construction). The modification is not actually needed for 3A 
assembly but helpful for the construction of new BioBricks. Some dedicated 
documentation of the Expression part protocols on OOW would also be needed, though.

Greetings,
Raik

Drew Endy wrote:
> (Forwarding a great reply from Chris Anderson at Berkeley to the full 
> standards list; for everybody's consideration and additional discussion).
> 
>> Hey Tom,
>>  
>>
>> That's an interesting proposal, and I do think we are still in an era 
>> in which exploring different assembly schemes is prudent, and possibly 
>> necessary.
>>
>>  
>>
>> First of all, there is the question of whether hybrid enzyme standards 
>> like Raik's make more or less sense than single pair standards.  My 
>> impression is that the hybrid strategy only exists to avoid re-making 
>> the collection of parts but serves no specific function beyond that.  
>> There aren't yet enough basic parts in existence that moving them to a 
>> new standard would be a substantial effort.  Therefore this strikes me 
>> as a weak motivator, and your proposal of trying a new standard that 
>> would be fundamentally incompatible with BBa suggests to me you see it 
>> the same way.  Nevertheless, the EcoRI-AgeI-part-NgoMIV-PstI could be 
>> a standard, and I'll pretend for the sake of argument that this is 
>> what Raik proposed (and the existing parts would of course remain 
>> compatible with such a standard).
>>
>>  
>>
>> In evaluating these things, it seems there are two distinct criteria 
>> categories to consider.  First is the acceptability of the scar 
>> sequence, and the second is the subtle effects of the enzymes on 
>> assembly chemistry.
>>
>>  
>>
>> In terms of broad acceptance outside of iGEM, I think there will 
>> always be holdouts on the idempotent standard assembly concept as long 
>> as the scar is not scarless.  I have never seen a chemistry providing 
>> a scarless solution to this that would be robust enough to serve as 
>> the backbone of the long-term assembly solution (not for a lack of 
>> looking, and kudos to Austin Che for trying!)  I think ultimately we 
>> can get there, but it will require some pretty difficult protein 
>> engineering.  In the meantime, we have a variety of 6 bp scar options 
>> that give reasonable scars for protein fusions--the Silver lab 
>> standard, BglII/BamHI, AgeI/NgoMIV, and SpeI/AvrII (and there may be 
>> other reasonable ones).  The peptides they encode all seem arguably 
>> reasonable to me, so this does not seem to be a viable criterion for 
>> distinguishing the standards.
>>
>>  
>>
>> The second criterion is the ease of assembly (including side reactions 
>> and frequent complications).  In practice this is very hard to 
>> evaluate as you must take into account the particular reaction scheme 
>> (prefix/suffix insertions, 3ab method, 1-2-3 method, one-pot method, 
>> dbbs (an ipcr scheme), the Goler PCR scheme, etc.) and the enzymes as 
>> they perform in that scheme.  The criteria you fault BglII/BamHI 
>> for--heat insensitivity and proliferation of genome fragments--both 
>> strike me as criteria predicated on a 3ab assembly scheme.  I would 
>> agree with you that BBb enzymes perform poorly in 3ab reactions.  
>> However, the 1-2-3, one-pot, and pcr-based schemes of BBb do not 
>> involve heat killing or 3-part ligations.  In the case of the 1-2-3 
>> and one-pot methods, the reactions involve site-specific methylation 
>> and background subtraction with the assembly enzymes.  So, the 
>> dominant issues are the performance of the cognate methyltransferases 
>> expressed from the coli genome and the overall efficiency of 
>> cutting-to-completion by the enzymes.  It is on these criteria that 
>> BglII/BamHI stands out.  I've never used NgoMIV, but I have worked 
>> with AvrII, AgeI, and of course SpeI and XbaI.  The reactivity of 
>> AvrII and SpeI are definitely better than AgeI and XbaI.
>>
>>  
>>
>> One also has to evaluate these standards for the ease of assembly for 
>> both the short term in the pre-automation era, their perfomance in the 
>> soon-to-come automation-with-reagents era, the reagent-free assembly 
>> era, the scarless era, and potentially a later phage-based or 
>> cell-free era.  How well, for example, does NgoMIV express and perform 
>> in an in vitro transcription/translation mixture?  It may seem 
>> premature to consider such things now, but this is the inevitable path 
>> of development for idempotent assembly schemes and these issues will 
>> ultimately determine what assembly chemistries remain in use 5 years 
>> from now.
>>
>>  
>>
>> One also must consider the course of development of 
>> oligonucleotide-based total synthesis that evolves in parallel to 
>> assembly schemes (and I won't even get into the implications of SLIC 
>> derivative methods).  If something dramatic happens in the market in 
>> the next 3 years bringing the price down two orders of magnitude, then 
>> this is all a big waste of brain energy.  If the total synthesis cost 
>> comes down at least to the point where pcr-based cloning of basic 
>> parts is impractical relative to total synthesis, then complications 
>> such as the frequency of internal restriction sites are irrelevant.  
>> Of course, that day may also never happen and restriction site 
>> frequency may still be a concern 10 years from now.
>>
>>  
>>
>> ...so, I think this is all pretty complicated.  I think about it a 
>> great deal, and my guestimations of the future lead me to 
>> BglII/BamHI.  However, it is all sufficiently complicated that the 
>> criteria that would lead away from BBb may be more relevant than my 
>> guestimations suggest, and I therefore think that having "felt out" a 
>> variety of alternative assembly chemistries is of great value.  They 
>> help guarantee that we avoid the worst case scenario--that we lock 
>> ourselves into the status quo and find ourselves at the same place we 
>> are now 5 years out.
>>
>>  -Chris
>>
>>
>> On Tue, Jul 8, 2008 at 12:34 PM, Tom Knight <tk at csail.mit.edu 
>> <mailto:tk at csail.mit.edu>> wrote:
>>
>>
>>
>>
>>
>>
>> -- 
>> J. Christopher Anderson, Ph.D.
>> Assistant Professor
>> Department of Bioengineering
>> http://andersonlab.qb3.berkeley.edu/
>>
>> Office: 308A Stanley Hall
>> Lab: 327 Stanley Hall
>>
>> Mailing Address:
>> J. Christopher Anderson
>> University of California, Berkeley
>> 327 Stanley Hall, Mailcode #1762
>> Berkeley, CA 94720
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Standards mailing list
> Standards at biobricks.org
> http://biobricks.org/mailman/listinfo/standards_biobricks.org

-- 
________________________________

Dr. Raik Gruenberg
http://www.raiks.de/contact.html
________________________________