A protein array containing antigens from different diseases is one of the main parts of the DiaCHIP. Manufacturing of this array can be done by "traditional" protein expression and spotting of the proteins on a specific surface or via cellfree expression. Our project focused on establishing a cellfree expression system for expression of the antigens but nevertheless we used different E. coli strains for expression of proteins as backup. Working with this two systems in parallel, cellfree expression on the one hand and expression with E. coli on the other hand, involved a lot of cloning.
That is why we designed our own cloning strategy and multiple cloning site which are adapted to our project. After some initial setbacks we finally found the system best suited for our needs, a vector we called pET_igem.
We soon realised that protein expression is probably a problem many iGEM teams around the world are facing during their projects. Therefore we decided to improve the original pSB1C3 in a similar way, thus providing an iGEM conform backbone (pOP)that can be used by other teams for easy expression of proteins.
During our cloning work we mainly used two methods: Classical cloning using restriction enzymes and Gibson cloning. We really got to appreciate the differences and advantages of the two different methods, which inspired us to write a short review comparing the two systems.

1. Our own cloning site

2. Vector design

3. Classical cloning vs. Gibson cloning

4. Detailed cloning strategy

One very basic task of our project is the cloning of all constructs that enable the expression of every antigen with several different tags. Additionally, those constructs have to be adapted for standard overexpression in E. coli as well as for cell-free expression. A first calculation, based on the assumption of using about 15 different proteins and 4 different tags, resulted in a total of 120 constructs to be cloned. Therefore, we decided to create our own multiple cloning site which could be used to combine every antigen with the needed tag(s) in a standardized matter, making it easy for many different people to work with it.

As a new cloning site, we did not use one of the available iGEM standards for several reasons. First, we need at least three distinct sites which can be used for insertion of a part, one for an N-terminal tag, one for an antigen and one for a C-terminal tag. Using, for example, SpeI and XbaI sites to insert a part in the middle would result in a 50 % chance to have the insert cloned in the wrong orientation, because SpeI and XbaI restriction botH produce the compatible overhang GATC. Another reason is that we originally planned to insert the whole expression cassette into pSB1C3. This iGEM standard plasmid already contains RFC [10], so using another iGEM standard between pre- and suffix is not possible.

The basic idea was to have distinct restriction sites which can be used for the insertion of an antigen, while others can be used for the exchange of tags.
Finally, this resulted in a cloning site containing Acc65I-, BamHI-, HindIII- and AlfII restriction sites (Fig. 1). Using those sites for the assembly of different antigen-tag combinations will never result in a frameshift, if the coding sequence is in frame with the restriction sites flanking it.

Figure 1: Self-designed cloning site: Our own cloning site containing the restriction sites Acc65I, BamHI, HindIII and AflII, that are assembled in a way that facilitates exchange of tags without getting a frameshift.

In practice, this means that the restriction sites BamHI and HindIII can be used to insert an antigen by classical cloning. Afterwards, either Acc65I and BamHI can be used to add an N-terminal tag or HindIII and AflII can be used to fuse a C-terminal tag to the antigen. Linker regions, creating some space between the tag and the antigen to avoid interactions between them, should be attached to the tag sequence. Other combinations of restriction sites can be used to insert antigen-tag combinations which might already be available. In addition to classical ligation, the respective restriction sites can be used to open the backbone for other cloning methods, e.g. Gibson Assembly.

The restriction sites have been chosen according to the amino acids which result by translation of their sequence. Acc65I (G/GTACC) and BamHI (G/GATCC) are translated into glycine and either serine or threonine. Those are hydrophilic, soluble amino acids which are commonly used in flexible linkers. The translation of HindIII (A/AGCTT) results in lysine and leucin. Especially lysine is highly hydrophobic, which is why some soluble amino acids should be used as a linker in front of a C-terminal tag. The last restriction site, AflII (C/TTAAG), is also translated to leucin and lysine. This will not much affect protein function or folding since these are the last translated codons. But in order to prevent rapid protein degradation according to the N-end-rule in bacteria, few additional amino acids have been inserted between the AflII site and the stop codon (TGA).

Figure 2: pIG15_001. pSB1C3 backbone containing an altered cloning site in the middle of an expression cassette for improved assembly of protein coding sequences with different tags (N- or C-terminal), which avoids the formation of frameshifts. Important parts for overexpression of proteins in E. coli were added inside the RFC[10].

Figure 3: pIG15_002. For cell-free expression the pSB1C3 backbone was altered similarly, but without a pelB secretion signal or lacI coding sequence and with a CMV promotor instead of a T7 promoter.

As all our constructs are supposed to be overexpressed and purified either in E. coli or via cell-free expression, we wanted to combine the advantages of the commonly used expression vector pET22b(+) with the iGEM standard vector pSB1C3 which allows us to hand in all our constructs to the iGEM registry.
Therefore, the main features of pET22b(+) were integrated between the BioBrick prefix and suffix of pSB1C3. Those features are namely the T7 promoter and terminator, a ribosomal binding site, the lacI gene and a signal sequence for periplasmic translocation of the protein (pelB). In between the pelB sequence and the terminator, a cloning site is arranged which we designed to simplify the cloning intents (see next section for more details). The resulting plasmid pIG15_001 is shown in Figure 2.
The backbone for E. coli-based cell-free expression purposes was designed analogous, but without a pelB secretion signal and the lacI gene, because those are not needed for cell-free expression (Fig. 3).

Unfortunately, after the first construct was ready for expression in E. coli, we realized that it did not work as expected. Instead of wasting time trying to optimize our own expression vector, we went on using the original vector and modified the cloning site to fit our requirements. The original vector pET22b+ and the modified version pET_iGEM are compared in Figure 4.
Using this expression vector exhibits a lot of advantages compared to working with a pSB1C3-based vector. One advantage is that the vector is already established for protein expression, so we could access the experience of more practised people. Additionally, this vector carries an ampicillin resistance as a selection marker instead of the chloramphenicol resistance in pSB1C3. This is advantageous in the case of protein expression because many E. coli strains, which are adapted for this purpose, are chloramphenicol-resistant. Thus, a double selection for successful transformation would not be possible.
However, the cell-free expression backbone was not changed because the template for expression is a PCR product amplified from promoter to terminator. This means the backbone, which is used for cloning the construct, is irrelevant for protein expression. So, we decided to clone the constructs in a cloning-optimized, high copy plasmid, pSB1C3.

Figure 4: Comparison of pET_iGEM and pET22b+. The two vectors differ in the the cloning site, which contains only four restriction sites in pET_igem (Acc65I, BamHI, HindIII, AflII) compared to the more complex multiple cloning site of the original pET22b+ (BamHI, EcoRI, SacI, SalI, HindIII, NotI, AvaI, XhoI). Also the 6x His-Tag from pET22b+ was removed, to enable easy tag-exchanges with our own system.

As mentioned before, we established a cloning strategy which is easy to handle for a lot of people. All our cloning efforts begin with either pET_iGEM, the modified version of pET22b+, for E. coli-based protein expression or pIG15_104, where the expression site derived from pET22b+ is inserted into pSB1C3.

The basic constructs (purification)

pET_iGEM: To modify the multiple cloning site of pET22b+ for our purposes, we performed a Gibson Assembly with only one fragment. Therefore we designed primers for the amplification of the whole plasmid, except the MCS. The cloning site we designed ourselves was part of the primer extention.
pET_803: To insert the very small but commonly used His-tag, we again used Gibson Assembly, because classical cloning of such small parts is hard. The respective primers were designed for amplification of antigen 8 (Herpes simplex TypeI, glycoprotein G1) to combine the insertion of the tag and an antigen. While the forward primer was a common Gibson primer, the reverse primer additionally contained the sequence of a 10xHis-tag. After amplification, the part was inserted into the BamHI and AflII digested backbone pET_iGEM, resulting in the first antigen with N-terminal His-tag.
pET_804: Another tag we use for immobilization of the antigens on the surface is the Spy-tag. This part of 39 bp length was inserted into pET_iGEM analogous to the 10xHis-tag. The only difference was, that the forward primer for amplification of antigen 8 included a standard 6xHis-tag, which is needed for protein purification. Consistent with our cloning strategy, the amplified part was inserted between Acc65I and AflII, because it contains an N-terminal as well a a C-terminal tag. BamHI and HindIII restriction sites were retained flanking the antigen making it exchangable.
pET_805: The only tag we used which was not inserted by Gibson Assembly was the Halo-tag. With a length of about 900 bp, classical cloning of this part was no trouble. The Halo-tag exists in two variants, one being optimized for C-terminal tagging of the protein, the other for N-terminal tagging (Ref. Promega?). As the tag we have access to is the C-terminal one, the Spy-tag in pET_804 was exchanged using the restriction sites HindIII and AflII resulting in pET_805.

Tagged constructs (purification)

We wanted to establish a tag-system which minimizes unspecific binding of proteins to the chip surface, therefore we tried out different tag systems.
His-tag: Every antigen was genetically fused to a C-terminal 10xHis-tag, which could be used for its immobilization on the surface as well es for its purification. The basic construct for this purpose was pET_803. The Herpes simplex derived antigen which was initially completed had to be excised by restriction with BamHI and HindIII. Thus, all the other antigens digested in the same way could be ligated into this backbone.
Spy-tag: Another tag for surface immobilization of the antigen is the Spy-tag. To enable protein purification via Ni-NTA constructs with Spy-tag were additionally fused to a C-terminal 6xHis-tag. The basic vector for cloning of those constructs was pET_804. Excision of the Herpes simplex derived antigen and insertion of other antigens was done analogous to antigen exchange in His-tag constructs.
Halo-tag:As a second covalent tag-system for immobilization of the antigens to the surface we used the Halo-tag, which specifically binds to a chloroalkane surface. Using our own cloning strategy, the Halo-tag can be easily inserted between HindIII and AflII, using pET_803 as basis. Cloning of the Halo-tag into pET_803 leads to a new vector, pET_805. This vector is then used to substitute the antigens. As in the Spy-tag-construct, this new vector contains a N-terminal 6x His-tag for purification of the antigens.
HA-tag:?

Cellfree backbone

As described above we decided to use the pSB1C3 standard vector as backbone for our cellfree expression constructs. Inserted between the RFC 10 prefix and suffix is our own cloning site, containing the restriction sites: Acc65I, BamHI, HindIII and AflII. We included a T7 promotor and a ribosomal binding site between the RFC 10 prefix and the Acc65I restriction site. A Stopcodon-cassette and a T7 terminator were inserted between the AflII restriction site and the RFC 10 suffix. These similarities to the expression vector facilitate the easy exchange of fragments between the expression vector and the cell free vector. Using our own cloning site it is rather simple to exchange different tag-combinations. The advantage of the pSB1C3 backbone is the high copy origin, allowing a rapid multiplication and easy cloning. For the cellfree expression itself the vector backbone is dispensable as only the PCR amplified part between the prefix and the suffix is mandatory for cellfree expression.

The basic constructs (cell free expression)

pIG15_104: For the first basic construct we ordered a gBlock from IDT containing T7 promotor, ribosomal binding site, our cloning site and the T7 terminator with adjoining Stopcodon-cassette. Between BamHI and HindIII the coding sequence for tYFP was inserted and between HindIII and AflII the Spy-tag sequence. This gBlock was then inserted into the pJET vector. Afterwards it was digested out of pJET and classically cloned into the pSB1C3 backbone resulting in the first basic construct for cell free expression. We inserted tYFP as default fragment into the cellfree expression backbone because we already had some anti-tYFP available. Hence first measurements for analysis of the binding of anti-TYFP to tYFP could be made as soon as the basic construct was finished. Besides tYFP could be used as reporter. Using BamHI and HindIII restriction sites tYFP can be excised and replaced by antigen sequences analogous to the cloning system used for the expression vector.
pIG15_105: Using the HindIII and AflII restriction sites the Spy-tag sequence was excised. For our second covalent tag system we inserted the sequence for the Halo-tag via classical cloning thereby producing our next basic construct for cellfree expression. Using identical cloning strategies as with the first basic construct (pIG15_104) all antigen sequences can be inserted into this vector using BamHI and HindIII.