Figure 1.Complete DNA sequence for the Prostate Cancer Therapeutic. This gene includes sequences for Uracil DNA Glycosylase and five Zinc-Finger Protein (ZFP) domains, as well as the two types of linkers that connect them.

The gene sequence for our proposed prostate cancer therapeutic (PCT) was synthesized and cloned into the pET-24a(+) vector system using the BamHI and NdeI cloning sites by Twist Bioscience. This system is an overexpressing construct containing the kanamycin resistance gene. Codon optimization was carried out for E. coli codon usage using the Twist Bioscience sequence optimization software. Verification of the gene insert was carried out using Sanger sequencing (Twist Bioscience). Sequence data showed that the gene had the correct restriction sites and the gene sequence was correct. The full DNA sequence for the PCT can be seen in Figure 1; Figure 2 shows the protein sequence of the entire designed therapeutic (54.54 kDa calculated molecular weight; bioinformatics.org).

Figure 2.Protein sequence for the Prostate Cancer Therapeutic (PCT). Amino acids are color-coded as follows: Uracil DNA Glycosylase (UDG)—black letters; Zinc-Finger (ZF) domains—green letters; SR linkers—red letters; Bridge-linkers/basic tail sequences—orange letters.

BL21 (DE3) cells (Invitrogen) were transformed by adding 1.0 ng of the pET 24a(+) vector containing the gene sequence for PCT. Cells were then heat-shocked as follows: Cells were incubated on ice for 1 hour, placed in a 40˚C hot water bath for 30 s, then immediately placed on ice for 5 min. 500 μL of 37˚C SOC media (National Diagnostics) was added and incubated at 37˚C for 1 h. The cells were spun down and spread out on a LB-agar plate containing the kanamycin antibiotic. The plate was incubated overnight at 37˚C.

A single colony was picked and inoculated in 5 mL of LB media containing 50 μg/mL kanamycin in a 15-mL conical tube. The resulting solution was incubated overnight at 37˚C in a shaking incubator. Cells were diluted 1:100 in 200 mL LB media containing 50 μg/mL kanamycin and grown at 37˚C in a shaking incubator until the optical density at 600 nm reached 0.6 OD units. At this point, 1 mM IPTG was added and the cells were left to shake for 3 h before being removed. The cell suspension was centrifuged at 3000×g for 30 min at 4˚C. The cell pellet was then suspended in liquid nitrogen for 1 min, then placed in a −80˚C freezer for storage.

PCT proteins were isolated from crude cellular extracts by resuspending 10 g of wet cell pellets in 40 mL metal chelate (MC) binding buffer solution (5 mM imidazole, 250 mM NaCl, and 20 mM Tris-HCl (pH 7.9). A lysis buffer was made by adding 0.1% NP-40 (to promote solubility), 0.5 mM PMSF, and 1.0 mM benzamidine (to inhibit any endogenous proteases) to the 40 mL MC buffer containing the resuspended cells. Cells were lysed by subjecting them to a single round of a freeze/thaw cycle (−20˚C for 24 h), followed by 10 repetitions of a 10-s sonication with a 1-min cooling cycle on ice. Sonicated cell extracts were then centrifuged at 3000×g for 40 min at 4˚C. The supernatant containing the PCT proteins was collected and applied to a 7-mL Nickel-NTA HisPur^TM Econo column (Bio Rad/Thermo Scientific). Prior to loading the cell extract, the affinity column was equilibrated in the MC metal chelate-binding buffer. The supernatant was loaded onto the column at a flow rate of 1 mL/min. The column was washed with 4 column volumes of MC buffer before starting the elution step. The bound protein was eluted over 3 column volumes of metal chelate elution buffer (1 M imidazole, 500 mM NaCl, and 20 mM Tris-HCl (pH 7.9)). Fractions were analyzed on 12% SDS-PAGE gels. Each protein eluted as a single peak within 3 - 5 fractions at an imidazole concentration of approximately 350 mM.

LC-MS/MS was performed using the same conditions as previously described [23]. Analysis was performed by the Proteomics Department at the University of Vermont (Burlington, VT).

3.1.1. Building a Specific Zinc Finger Array for Binding to the TMPRSS2:ERG4 Gene

The Zinc Finger Consortium (http://www.zincfingers.org/default2.htm) has studied binding of individual ZF modules to their target DNA sequences [2][24]. The ZiFiT (Zinc Finger Targeter) program [25][26], which has been developed by the Consortium, facilitates the design of custom zinc finger proteins (ZFPs) by identifying suitable DNA target sites and matching them with validated zinc finger modules using a modular assembly approach. We utilized ZiFiT’s “Modular Assembly” method, which aims to construct ZFPs by linking three individual pre-characterized ZF modules, each recognizing three consecutive and specific 3-bp DNA subsites, together. The ZiFiT program accepts as input uploaded DNA sequences, which it then divides into 9-bp regions composed of three consecutive 3-bp triplets. It then compares each triplet of triplets against a library of experimentally-validated ZF modules. Because three ZF proteins are required to bind to 9-bp regions, the selectivity of the designed ZFP system is increased and the chance of off-target binding is decreased. Only compatible triplets (especially GNNs) are retained for reliable design. Each potential 9-bp target site is evaluated based on GNN Score (preference is given to sites with more GNN triplets, as they yield higher functional success) and Affinity Score (a predictive metric indicating DNA-binding strength, where lower values suggest stronger binding and better functionality). Individual ZF–to–triplet affinities (e.g., a numeric K_d for a single ZF, such as ZF83 binding to AGC) have not been published or centrally curated, and these data are not reported in ZiFDB or Consortium notes. However, it has been reported that assembled multi-finger proteins typically bind with high affinity in the low nanomolar range (K_d ≈ 1 − 20 nM) and with high specificity [27].

For our study, we uploaded the TMPRSS2/ERG DNA sequence of interest into ZiFiT and selected target sites with high GNN content and favorable Affinity Scores. We selected five promising ZF proteins that should, when expressed as a covalently-linked system, bind with high specificity to DNA triplets found around the breakpoint region of the prostate cancer gene TMPRSS2:ERG4 [3][4], specifically AGC (for the first zinc finger) and GCG (for the second) in the TMPRSS2 exon 1 portion of the breakpoint region, and AGG (for the third zinc finger), AAT (for the fourth) and CTT (for the fifth) on the ERG4 portion of the breakpoint region. Based on these results, we selected ZF83, ZF23, ZF84, ZF77 and ZF103 as modules for the ZF array using the ZF numbering introduced by the Zinc Finger Consortium (see Figure 3). Figure 4 shows a schematic for the overall concept for the design of the PCT protein.

Figure 3.TMPRSS2/ERG DNA sequence. The TMPRSS2 sequence is shown in orange, and the ERG sequence is shown in blue. The red triangle indicates the break-region. Zinc Finger Proteins are shown at their respective triple base-pair binding sites (ZF numbering is from the Zinc-finger Consortium) and include ZF83 (AGC); ZF23 (GCG); ZF84 (AGG); ZF77 (AAT); and ZF103 (CTT).

Figure 4. Schematic showing the PCT protein overlapping the TMPRSS2/ERG4 break region. The image shows the ribbon structure of UDG linked to five Zinc-Finger Proteins (ZFP). Each ZFP is aligned along the DNA in their approximate positions with respect to the break region.

3.1.2. Molecular Motor Design

UDG2 has an NLS, which should be useful in vivo to locate the system to the chromosomal DNA in an organized fashion as opposed to the ZF array randomly encountering its target sequence through diffusion. UDG2 can also act as a molecular motor, pulling the system along the DNA in search of its target sequence. One of UDG’s functions is to unpackage, unwind and clear chromosomal DNA, which may assist in locating the system to the nucleus and increasing the accessibility of the DNA target. In addition, the presence of the UDG may serve to overcome nonspecific interactions of the zinc finger system to random DNA sequences, as the force of UDG hopping and short-range scanning may overcome any nonspecific, weaker interactions the ZF array may experience.

3.1.3. Linkers/Anchor Design

Wild-type GATA proteins have ZF motifs that are primarily responsible for binding to specific DNA sequences. They also contain a lysine-rich C-terminal tail (or anchor) that has been shown to be necessary for maintaining high DNA binding affinities [15][16]. We predict that inclusion of a C-terminal tail found in the GATA1 protein (protein sequence DVIKKRNR) will improve the binding affinity of the UDG-containing fusion system to the TMPRSS2:ERG4 gene sequence. We have also used the DVIKKRNR sequence to bridge between sets of two zinc finger modules; each individual module is joined by short SR linkers (protein sequence SR). The fifth ZF motif is preceded by the longer DVIKKRNR sequence to introduce necessary flexibility for binding to the DNA target.

3.2.1. Protein Expression

Protein expression was carried out according to the conditions outlined in Section 2.4. Colony growth on antibiotic-inclusive plates were excellent, with large, well-distributed colonies present after 12 - 15 h incubation. Figure 5 shows the expression profile for the induction of the protein therapeutic PCT. When comparing the pre- and post-induction lanes, we see that they share a number of similar-sized protein bands. However, the post-induction has an additional heavy band at approximately 40 kD that is not present in the pre-induction lane. These results are not consistent with the expected molecular weight of 54 kD for the PCT protein.

Figure 5.Gel of expression for the PCT/pET24a(+) construct. PL = protein ladder (MW values are labeled to the left of the gel); Pre is pre-induction; Postis post-IPTG induction after 4 hrs. The black arrow points to the heavy band at approximately 40 KD in the post-induction lane that indicates the overexpression of the E.coli OmpF protein, which is verified by LC-MS/MS as shown in Table 1.

3.2.2. Affinity Chromatograph

Figure 6. Gel of the eluted protein from Nickel-NTA purification. PL indicates protein ladder (MW values are labeled to the left of the gel); EP is the eluted protein lane. The black arrow points towards the band at approximately 40 KD in the eluted protein sample lane (EP) that indicates overexpression and elution of the E. coli OmpF protein as verified by LC-MS/MS (see Table 1).

Protein extracts were purified using a Nickel-NTA column under the conditions outlined in Section 2.4. The eluted protein sample was visualized on a 12% SDS-PAGE gel (Figure 6). The only band present is at the 40 kD mark, which corresponds to the overexpressed protein in Figure 3. In order to determine if this was a truncated PCT or a different protein entirely, the gel piece containing the band was removed and analyzed by LC-MS/MS.

3.2.3. Protein Identification by Liquid Chromatography-Tandem Mass Spectroscopy (LC-MS/MS)

From the LC-MS/MS data, it can be seen that the predominant protein found is an E. coli protein, Outer Membrane Porin F (Table 1). This protein was identified with high confidence and a high coverage (>10 peptides with a high peptide identification score). From the SDS-PAGE analysis, these results have been consistent throughout the overexpression optimization process for PCT. No PCT fragments were found in the LC-MS/MS results, indicating that protein expression of the target therapeutic was not successful.

Table 1. Sample Identification by Liquid Chromatography-Tandem Mass Spectroscopy (LC-MS/MS). The table shows the identity and information for the 25 most significant proteins detected from an excised protein sample as shown in Figure 6. The proteins are listed from the most to least abundant proteins in the sample analyzed.