View source for The BioBricks Foundation:Standards/Technical/Formats

'''Biobrick Formats''': 
This working group aims to specify a new BioBrick DNA format.


= Aim / Application scenarios for this standard =

Formulate a successor to the current BBa physical assembly format of standard biological parts.
[add further]

= Overview over existing and proposed Biobrick formats =

All biobrick formats proposed so far follow the same basic scheme where restriction and ligation of two biobricks forms a new biobrick: 

[[Image:Bbformats_general.png|250px]]


== classic Biobrick format (BBa)==

This is the format used by most iGem teams and most BioBricks in the MIT registry.

<table>

<tr>
  <td><center>prefix</center></td><td></td><td><center>suffix</center></td>
</tr>

<tr>
<td><pre>
5' GAATTC GCGGCCGC T TCTAGA G
   EcoRI    NotI      XbaI
</pre></td>

<td>...part...</td>

<td><pre>
T ACTAGT A GCGGCCG CTGCAG 3'
   SpeI     NotI    PstI </pre></td>

</tr>

<tr>
  <td>Protein coding prefix</td><td></td><td></td>
</tr>

<tr>
<td><pre>
5' GAATTC GCGGCCGC T TCTAG
   EcoRI    NotI     XbaI 
</pre></td>

<td>ATG.part</td>
<td></td>
<tr>

</table>

Fusing two parts leaves the following scar:

<table>
<tr>
<td>5' ...part A...</td>
<td><pre>
TACTAGAG
 *  *
</pre></td>
<td>...part B... 3'</td>
</tr>
</table>

<!-- [http://openwetware.org/wiki/Image:Bbformats_classic.png See also Image] -->

[http://parts.mit.edu/registry/index.php/Help:BioBrick_Prefix_and_Suffix  description at parts.mit.edu]

====Advantages====
* standard
* well tested and documented
* native protein start codon can be preserved while using RBS parts.
* large and still growing set of parts

====Disadvantages====
*  no protein fusions (frame shift, stop codon)
*  a single mutation (at the fused region) can upset the setup?

== Biofusion (Silver lab) ==

The Silver lab modified the classic 1.0 format to allow for protein fusions:

<table>

<tr>
  <td><center>Prefix</center></td><td></td><td><center>Suffix</center></td>
</tr>

<tr>
<td><pre>
5' GAATTC GCGGCCGC T TCTAGA
   EcoRI    NotI      XbaI
</pre></td>

<td>...part...</td>

<td><pre>
ACTAGT A GCGGCCG CTGCAG 3'
 SpeI     NotI    PstI </pre></td>

</tr>

</table>

Fusing two parts now leaves the following scar:

<table>
<tr>
<td>5' ...part A...</td>
<td><pre>
ACTAGA
 T  R
</pre></td>
<td>...part B... 3'</td>
</tr>
</table>
<!-- [[Image:Bbformats_biofusion.png]] -->

[http://www.openwetware.org/wiki/Silver:_BB_Strategy  description by Silver lab]

====Advantages====
* in-frame fusion of protein parts
* restriction-compatible to 1.0 parts -- no new enzymes
* also protein parts can, theoretically, be fused N-terminally to to BBa protein parts, as long as the frameshift is corrected by an adapter part
* used by several iGem teams

====Disadvantages====
* Arg in scar can be problematic
* N-terminal Thr-Arg = destabilization signal (N-end rule)
* Dam methylation blocks cloning when prefix is followed by "TC"
* unexpected side-effects for users not aware of the shortened prefix/suffix
* non-coding parts may be not functionally compatible due to the changed bp distance
* frameshift with respect to what is expected from protein coding 1.0 parts
* not possible to preserve native protein start (but equivalent to BBa coding part could be constructed)

== Fusion parts (Freiburg iGem2007 team) ==

The Freiburg iGem2007 team proposed an extension of BBa, which would enable protein fusions but alleviate the disadvantages of the Biofusion format. Two restriction sites are added within the standard BBA sites. These additional sites provide compatible ends, can be employed using the same cloning strategy as for the standard restriction sites, and code for amino acids suited for linkers:

<table>

<tr>
  <td><center>Prefix</center></td><td></td><td><center>Suffix</center></td>
</tr>

<tr>
<td><pre>
5' GAATTC GCGGCCGC T TCTAGA GT GCCGGC
   EcoRI    NotI     XbaI  Met NgoMIV
</pre></td>

<td>...part...</td>

<td><pre>
ACCGGT TAAT ACTAGT A GCGGCCG CTGCAG 3'
 AgeI   *    SpeI     NotI    PstI </pre></td>

</tr>

<tr>
  <td>"N-part" prefix</td><td></td><td></td>
</tr>

<tr>
<td><pre>
5' GAATTC GCGGCCGC T TCTAG
   EcoRI    NotI     XbaI 
</pre></td>

<td>ATG.part</td>
<td></td>
<tr>
</table>

Fusing two parts now leaves the following scar:

<table>
<tr>
<td>5' ...part A...</td>
<td><pre>
ACCGCC
 T  G
</pre></td>
<td>...part B... 3'</td>
</tr>
</table>

<!--[[Image:Bbformats_freiburg.png]]-->

[http://parts.mit.edu/igem07/index.php/Freiburg07/report_fusion_parts  description by Freiburg iGem team]

For cases where the native ATG needs to be conserved, the Freiburg team introduces an "N-part" which has the classic BBa coding part prefix and the Expression part suffix. N-parts would need to be cut with XbaI in place of NgoMIV.

====Advantages====
* in-frame fusion of protein parts
* benign protein scar
* N-end rule safe (long protein half-life)
* provision for preserving native N-terminal while using RBS parts
* both new enzymes can be heat-inactivated
* stand-alone protein expression (start + stop in prefix / suffix)
* full BBa compatibility -- functionally & compositionally equivalent to BBa protein coding part
* blunt-cutting isochizomer of NgoMIV (NaeI) -- possibility of directional cloning with two inner restriction sites enables part transfer between different formats and other potentially interesting transfer reactions.

====Disadvantages====
* N-parts are assembled with a different enzyme combination.
* not compatible to BioFusion protein parts (frame shift + stop codon), but see below!

<table frame=box>
<tr><td>
====Expression / BioFusion format conversion strategies====

Until recently, the Silver lab BioFusion format was the only well-described way to compose fusion proteins from (unofficial) BioBricks. Due to the frameshift, the Expression Part format cannot be made compatible to both BBa and BioFusion at the same time. Since I (Raik) am collaborating with BioFusion users, I am constructing a conversion vector that introduces a prefix and suffix that are compatible to both BioFusion and Expression Parts. I would not consider this a good candidate for a real format though.

<!--
=====Expression -> BioFusion adapter parts=====

* C-terminal adapter: a part ending in an incomplete 2-bp codon 
<sub>...would correct the frameshift to Biofusion and override the STOP codon. Biofusion parts could then be appended to the adapter using 1.0 restriction. The resulting scar would be ugly though: P V N (T R)
</sub>
* N-terminal adapter: a part beginning with a single stray nucleotide
<sub>...would allow to allow to couple Biofusion parts in front of the adapter. The scar would be again ugly: (T R) W P R (regardless of the stray nucleotide).
</sub>
-->
=====Expression -> BioFusion conversion vectors=====

Construction vectors with a modified Freiburg prefix / suffix could bring a protein part in frame with BioFusion parts and remove the STOP between the AgeI and SpeI site. Restriction / ligation with AgeI + NaeI can (theoretically) transfer Expression parts into this conversion vector which can then be used for normal BioFusion cloning (at the cost of adding a T G before and after the part). 
NaeI is an isoschizomer to NgoMIV but generates blunt ends which should allow for a directional transfer.
<table>
<tr>
<td><pre>
5' GAATTC GCGGCCGC T TCTAGA GCCGGC
   EcoRI    NotI     XbaI   NgoMIV
</pre></td>

<td>...part...</td>

<td><pre>
ACCGGT ACTAGT A GCGGCCG CTGCAG 3'
 AgeI   SpeI     NotI    PstI </pre></td>

</tr>
</table>

=====BioFusion --> Expression part conversion=====
None really. Introducing the modified flanks by PCR seems the only way.
</td></tr>
</table>

== The Berkeley (BBb) Format ==
BBb is used by several researchers at UC Berkeley and is based on idempotent assembly with BamHI and BglII restriction enzymes.  In a nutshell, most plasmids look like this:

<table>

<tr>
  <td><center>Prefix</center></td><td></td><td><center>Suffix</center></td>
</tr>

<tr>
<td><pre>
5' GAATTC atg AGATCT
   EcoRI      BglII
</pre></td>

<td>...part...</td>

<td><pre>
GGATCC taa CTCGAG 3'
BamHI   *   PstI </pre></td>
</tr>

</table>

Fusing two parts leaves the following scar:

<table>
<tr>
<td>5' ...part A...</td>
<td><pre>
GGATCT
 G  S
</pre></td>
<td>...part B... 3'</td>
</tr>
</table>

Note, however, that BBb is intended as a minimal physical assembly standard, and only those features needed for interconversion of BBb plasmids are formally defined.  Therefore, "atg" and "taa" spacers are not core definitions of the standard.

'''Formal Definition:'''
*A BBb part is a DNA sequence flanked on the 5' end by "GATCT" and on the 3' end by "G" lacking BglII, BamHI, EcoRI, and XhoI restriction sites
*A BBb vector is a DNA sequence flanked on its 5' end by "GATCC" and on its 3' end by "A"
*A BBb entry vector has a unique EcoRI site, no BamHI or BglII restriction sites, and at most one XhoI site 5' to the EcoRI site
*A BBb plasmid is represented as <vector_name>-<part_name> and has the sequence obtained by concatenating the vector and part sequences
*Further definition constraints are "sub-standards" of the BBb format

===Advantages===
* in-frame fusion of protein parts
* benign protein scar
* enzymes selected for efficient cutting

===Disadvantages===
* BglII cannot be heat-inactivated -- current 3A standard assembly won't work
* incompatible to BBa format
* incompatible to BioFusion format

== Tom Knight's BBb proposal ==

Tom Knight has recently proposed to replace BBa by the following format: 

<table>

<tr>
  <td><center>Prefix</center></td><td></td><td><center>Suffix</center></td>
</tr>

<tr>
<td><pre>
5' GAATTC...ACTAGT
   EcoRI     SpeI
</pre></td>

<td>...part...</td>

<td><pre>
GCTAGC...CTCGAG 3'
 NheI     PstI </pre></td>
</tr>

</table>

Fusing two parts would then leave the following scar:

<table>
<tr>
<td>5' ...part A...</td>
<td><pre>
GCTAGT
 A  S
</pre></td>
<td>...part B... 3'</td>
</tr>
</table>

===Advantages===
* in-frame fusion of protein parts
* benign protein scar
* N-end rule save
* introduction of only one new enzyme
* NheI is comparatively rare in the E. coli genome -- less background fragments from genomic DNA
* NheI can be heat-inactivated
* assembly scar ends in ATG
** RBS parts could thus be redesigned to furnish (start-less) protein parts with the ATG
** the native N-terminal of a protein part would then be automatically preserved

===Disadvantages===
* incompatible to BBa format
* incompatible to BioFusion format

=Comparison of different format proposals=

<table frame=box>
<tr align=center>
<td width=300></td>             <th width=85>BBa</th> <th width=85>Silver</th> 
                                           <th width=85>Freiburg</th> <th width=85>Berkeley</th> <th width=85>Knight</th></tr>

<tr><td><i>protein issues</i></td></tr>
<tr align=center>
  <td>(1) protein fusion?</td>      <td>no</td>     <td>yes</td>     <td>yes</td>      <td>yes</td>      <td>yes</td></tr>
<tr align=center>
  <td>(2) good protein scar?</td>   <td>n/a</td>    <td>no</td>      <td>yes</td>      <td>yes</td>      <td>yes</td></tr>
<tr align=center>
  <td>(3) N-end rule save?</td>     <td>n/a</td>    <td>no</td>      <td>yes</td>      <td>yes</td>      <td>yes</td></tr>
<tr align=center>
  <td>(4a) native N' after RBS part?</td>  
                                    <td>coding part</td> <td>no</td> <td>N-part</td>   <td>no</td>       <td>p-RBS</td></tr>
<tr align=center>
  <td>(4b) native N' after Kozak part?</td>  
                                    <td>no</td>     <td>no</td>      <td>no</td>       <td>no</td>       <td>no</td></tr>

<tr><td><i>BBa compatibility</i></td></tr>
<tr align=center>
  <td>(5) same ass. enzymes?</td>   <td>yes</td>    <td>yes</td>     <td>no</td>       <td>no</td>       <td>no</td></tr>
<tr align=center>
  <td>(6) restriction-compatible?</td>
                                    <td>yes</td>    <td>yes</td>     <td>yes</td>      <td>no</td>       <td>no</td></tr>
<tr align=center>
  <td>(7) ..side effects?</td>      <td>none</td>   <td>some</td>    <td>none</td>     <td>n/a</td>      <td>n/a</td></tr>
<tr align=center>
  <td>(8) directional transfer with inner restr. sites?</td>   
                                    <td>no</td>     <td>no</td>  <td>sticky/blunt ligation</td>  <td>no</td>     <td>no</td></tr>

<tr><td><i>standard assembly issues</i></td></tr>
<tr align=center>
  <td>(9) heat inactivation?</td>   <td>yes</td>    <td>yes</td>     <td>yes</td>      <td>no</td>       <td>yes</td></tr>
<tr align=center>
  <td>(10)inner restr. site occurrence in E. coli</td>
                                    <td>240</td>    <td>240</td>     <td>2100</td>     <td>1200</td>     <td>500</td></tr>
<tr align=center>
  <td>(11) enzyme efficiency?</td>  <td>ok</td>     <td>ok</td>      <td>?</td>        <td>good</td>     <td>?</td></tr>

<tr><td><i>other issues</i></td></tr>
<tr align=center>
  <td>(12) current adoption?</td>   <td>wide</td>   <td>some</td>    <td>some</td>     <td>some</td>     <td>none</td></tr>
</table>

Explanations:

<sub>
(2) good meaning small and biochemically nonperturbing

(3) Some N-terminal amino acids are strong destabilization signals in both pro- and eucaryotes. The assembly scar turns into the protein N-terminal if a RBS+start part is coupled with the protein part -- the scar should hence not code for destabilizing residues.

(4a) Some protein parts, in particular signaling peptides, critically depend on preserving their sequence at the N-terminal. That means any additional scar residues between a "start" part and the old N-terminal would disrupt function. Procaryotic RBS tolerate a variable 6bp spacer between RBS and AUG and special parts-layout can shift the assembly scar into this spacer. BBa and Freiburg define a specialized sub-format where the ATG is shifted into the part and partly overlaps with the end of the scar. p-RBS: Tom's assembly scar ends in ATG and RBS parts could thus be redesigned to furnish (start-less) protein parts with the ATG.
see: [http://www.ambion.com/techlib/append/rbs_requirements.html RBS, Kozak explanation]

(4b) All bets are, apparently, off for eucaryotic proteins where the Kozak reaches 4 bp (ATG G) into the reading frame. Only a scar-less assembly method would allow us to freely combine special Kozak parts (lacking ATGG) with special N-term protein parts (starting with ATG). 
Kozak and N-terminal signaling peptides are thus not decomposable with any of the current schemes.

(5) Uses same restriction enzyme combination as the BBa assembly.

(6) The format allows to couple parts with "old" BBa parts using the BBa assembly.

(7) Side effects mean that parts in the new format do not behave exactly like BBa parts after assembly with "old" BBa parts.

(8) Can the part be directionally transferred into other vectors that introduce different flanking sequences (more or less) directly right and left of it? This may be critical for offsite-cutter (IIS) based or other future assembly schemes.

(9) Can all restriction enzymes used be heat-inactivated?

(10) Approximate combined occurrence of the two inner restriction sites in E. coli. 

</sub>

=entirely different strategies=

=== IIS restriction and multi-fragment ligation ===

The IIS restriction strategy from the UCSF iGem2007 team could probably be extended into a more general multi-ligation Biobrick system:
[http://parts.mit.edu/igem07/index.php/UCSF/Cloning_Strategy1 UCSF 2007 cloning strategy]

=== BioBrick ++ ===

... was an early (2004) proposal for a more versatile BioBrick format, which somehow didn't catch on. BioBrick ++ is based on a sophisticated combination of IIS (offsite cutters) and nicking restriction enzymes, and was intended to allow both seamless and normal BioBrick assembly, flipping of BioBricks and other operations. There are some disadvantages though [Raik's opinion, add your own view]: 

(1) the large combination of restriction sites makes the system not quite easy to understand. (2) ++ was designed without keeping protein fusions in mind -- the proposed standard assembly would again introduce a frameshift and a stop codon, although the more sophisticated blunt assembly would of course work for protein fragments. (3) the different assembly methods produce different frames. (4) Some of the proposed enzymes or ligation schemes may not behave as ideally as assumed (?) (5) Several of the proposed "operations" involve two sequential restriction/ligation/transformation cycles which, in practice, may amount to more work then a normal single step conversion by PCR. 

Nevertheless, BioBrick ++ describes, at least, two core innovations that may be very helpful for a second (or third?) generation BioBrick format:
# IIS-restriction (offsite cutting) in prefix and suffix uncouples the cohesive ends from the enzyme recognition sites -- overhangs can therefore end directly at the part boundary (allowing for blunt ligation strategies and parts "upgrade")
# Construction plasmids can be created with any overhang by inside-out IIS restriction or with nicking enzymes.