Creation of zebularine-resistant human cytidine deaminase mutants to enhance the chemoprotection of hematopoietic stem cells

Abstract

Human cytidine deaminase (hCDA) is a biomedically important enzyme able to inactivate cytidine nucleoside analogs such as the antileukemic agent cytosine arabinoside (AraC) and thereby limit antineoplastic efficacy. Potent inhibitors of hCDA have been developed, e.g. zebularine, that when administered in combination with AraC enhance antineoplastic activity. Tandem hematopoietic stem cell (HSC) transplantation and combination chemotherapy (zebularine and AraC) could exhibit robust antineoplastic potency, but AraC-based chemotherapy regimens lead to pronounced myelosuppression due to relatively low hCDA activity in HSCs, and this approach could exacerbate this effect. To circumvent the pronounced myelosuppression of zebularine and AraC combination therapy while maintaining antineoplastic potency, zebularine-resistant hCDA variants could be used to gene-modify HSCs prior to transplantation. To achieve this, our approach was to isolate hCDA variants through random mutagenesis in conjunction with selection for hCDA activity and resistance to zebularine in an Escherichia coli genetic complementation system. Here, we report the identification of nine novel variants from a pool of 1.6 × 10⁶ transformants that conferred significant zebularine resistance relative to wild-type hCDA2. Several variants revealed significantly higher K_i values toward zebularine when compared with wild-type hCDA values and, as such, are candidates for further exploration for gene-modified HSC transplantation approaches.

Keywords: cytidine deaminase; cytosine arabinoside; hematopoietic stem cells; random mutagenesis; zebularine.

Fig. 1

Overview of AraC metabolism. AraC is phosphorylated by deoxycytidine kinase (dCK) and other enzymes into its active, cytotoxic form AraCTP, which can incorporate into DNA and induce cell death. AraC can also be deaminated into its inactive metabolite AraU by hCDA. When AraC combined with zebularine (Zeb) is used to treat leukemia, HSCs gene-modified with a zebularine-resistant hCDA mutant (mt) will be protected from AraC cytotoxicity and by association minimize myelosuppression, while leukemic cells will be eradicated.

Fig. 2

Outline of hCDA2 random mutagenesis procedure and agarose gel analysis of EP-PCR and DNA shuffling products. A single round of mutagenesis consists of introducing random mutations (filled circle) into the hCDA2 gene by EP-PCR, digestion of the EP-PCR products into small fragments with DNase I, the reassembly of small fragments by primerless PCR and reamplification. The resultant recombinant fragments were then ligated into pETHT, transformed into E. coli strain SØ5201(DE3) and selected for growth on zebularine-containing CDA plates. This process was repeated for multiple rounds with zebularine concentrations being increased in each round. Mutants that grew on plates were used as EP-PCR templates for subsequent rounds of mutagenesis. The products of EP-PCR and DNA shuffling were analyzed on 1.2% agarose gels after each round and one representative gel is shown. Lane M, low molecular weight DNA ladder; Lane 1, EP-PCR product (∼720 bp); Lane 2, DNase I treatment of EP-PCR product into fragments of about 50–200 bp; Lane 3, the smear product of primerless PCR by using small fragments as template; Lane 4, the product of reamplification (∼720 bp) using primerless PCR DNA as template.

Fig. 3

Functional complementation assays of CDA-deficient E. coli strain SØ5201(DE3) by wild-type and mutant hCDA2 enzymes. E. coli SØ5201(DE3) harboring the expression vector (pETHT), pETHT:hCDA2 (hCDA2), or one of the nine mutants (in clock-wise position from pETHT:hCDA2, pETHT:hCDA2-8102, pETHT:hCDA2-8173, pETHT:hCDA2-8202, pETHT:hCDA2-8224, pETHT:hCDA2-8362, pETHT:hCDA2-8474, pETHT:hCDA2-8727, pETHT:hCDA2-9019 and pETHT:hCDA2-9028) were grown at 37°C for 24–36 h on CDA selection plates (A) in the absence of zebularine (−zebularine) or (B) in the presence of 30 µg/ml zebularine (+zebularine).

Fig. 4

Growth curves of CDA-deficient E. coli strain SØ5201(DE3) harboring pETHT:hCDA2 (empty circle) or one of the nine mutants [pETHT:hCDA2-8102 (empty square), pETHT:hCDA2-8173 (empty diamond), pETHT:hCDA2-8202 (empty triangle), pETHT:hCDA2-8224 (filled circle), pETHT:hCDA2-8362 (filled square), pETHT:hCDA2-8474 (filled diamond), pETHT:hCDA2-8727 (filled triangle), pETHT:hCDA2-9019 (cross) and pETHT:hCDA2-9028 (asterisk)] at 37°C in CDA selection medium (A) without zebularine or (B) with 30 µg/ml of zebularine was monitored for 8 h by OD₆₀₀ readings. The experiments were repeated three times, and similar results were obtained. Representative plots are shown.

Fig. 5

Alignment of the amino acid sequences of wild-type hCDA2 and nine selected mutants. The sequences were aligned using BioEdit sequence alignment editor (version 7.0.9.0). Mutant names (numbers) are noted to the left of the sequences and an asterisk denotes a stop codon. The bottom line is the amino acid sequence corresponding to the B. subtilis CDA (P19079; Mizuno et al., 1989) with residues conserved between hCDA2 and the Bacillus CDA in gray. Underlined residues denote identities between Bacillus CDA residues and those found in at least one of the mutant hCDA2 sequences.

Fig. 6

SDS–PAGE of the purified wild-type hCDA2 and nine mutant hCDA2 enzymes from E coli BL21(DE3). Protein samples (∼600 ng) were analyzed in a 15% SDS–polyacrylamide gel. Wild-type hCDA2 is shown in the first lane with molecular weight markers (M) in the far right lane.

Fig. 7

Enzyme assays of purified hCDA2 and nine hCDA2 mutants. Purified enzymes (1 µg each) were subjected to spectrophotometric enzyme assays by using (A) 140 µM of CdR or (B) 140 µM of AraC as substrate. The enzymes assayed were wild-type hCDA2 (empty circle) or one of the nine mutants (8102 (empty square), 8173 (empty diamond), 8202 (empty triangle), 8224 (filled circle), 8362 (filled square), 8474 (filled diamond), 8727 (filled triangle), 9019 (cross) and 9028 (asterisk)). The enzyme assays were repeated three times with similar results observed. Representative plots are shown.