Cell lines were purchased from ATCC (Manassas, VA). PC3, DU145 (both prostate carcinoma) and HEK293 (human embryonic kidney) cells were grown in RPMI-1640 medium (Gibco, 11875, NY, USA), DMEM, and DMEM (Gibco, 11885), respectively; both media were supplemented with 10% FBS, penicillin‒streptomycin (Gibco 15140), and 1 mM L-glutamine (A2916801).
wtHEK293T and ADAR1 KO HEK293T cells13 were a generous gift from Dr. Charles M. Rice at Rockefeller University, NY, USA. These cells were grown in DMEM supplemented with 10% FBS, 1X nonessential amino acids and penicillin‒streptomycin.
Cells were fixed with 4% formaldehyde and analyzed with an LSM 880 META confocal laser scanning microscope (Zeiss, Oberkochen, Germany) equipped with 63x and 40x A-Plan oil immersion objectives using ZEN imaging software in multitrack mode.
Generation of the uneditable AZIN1 construct
The codon corresponding to amino acid residue 367 in AZIN1 was mutated from AGC (Ser) to TCC (Ser). The mutations were introduced by site-directed mutagenesis using a QuikChange Lightning Site-Directed Mutagenesis Kit according to the manufacturer’s protocol (Agilent Technologies). The forward primer GATCAAATTGTGGAATCCTGTCTTCTTCCTGAGCTGAATGTGGG and the reverse primer CCCACATTCAGCTCAGGAAGAAGACAGGATTCCACAATTTGATC were used.
Protein expression and purification
The bacterial expression vectors Clover-pBAD (Addgene, #54575), mRuby2-pBAD (Addgene, #54771), and pBAD-mTAG-BFP2 (Addgene, #54572) encoding the 6X-HIS-tagged fluorescent protein with a TEV protease cleavage site were purchased from Addgene. The human OAZ1 gene was codon optimized for expression in E. coli (https://www.idtdna.com/CodonOpt) and synthesized de novo (ThermoFisher). The DNA sequences encoding the N-terminal fluorescent fusion protein of hAZIN1 and the E. coli expression-optimized hOAZI were synthesized by overlap extension PCR using previously described methods14. Expression vectors of C-terminal fluorescent fusion proteins were synthesized using NEBuilder HiFi DNA Assembly (New England Biolabs). Point mutations in the hAZIN1 gene were introduced using PCR-based site-directed mutagenesis with Pfu Ultra High-Fidelity DNA Polymerase (Agilent) using the described protocols. The fusion proteins were expressed and purified according to our previously described methods15.
To generate mammalian expression vectors, we amplified the full-length cDNA sequences encoding Clover-Antizyme and Clover-edited-antizyme from the pBAD expression vectors described above by PCR with the primers
F: 5’- cgcGCTAGCcGccATGgTGAAATCCTCCCTGCAGcg-3’ and
R: 5’-cgcAAGCTTCTTACTTGTACAGCTCGTCCATCC-3’, and used the primers
F: 5’- cgcGCTAGCcGccATGgTGAGCAAGGGCGAG-3’ and
R: 5’-cgcAAGCTTTTATGCTTCAGCGGAAAAGCTGTC-3’ to similarly amplify the Antizyme-mRuby2 sequence. Subsequently, the purified products were ligated into pcDNA3.1 + (Addgene, MA, USA) according to the manufacturer’s instructions. The other plasmids used, pcDNA3.1-Clover-mRuby2 (plasmid #49089) and pcDNA3 mRuby2 LIC cloning vector (6H), were commercially available (Addgene, MA, USA).
All sequences were verified by multiple Sanger sequencing runs using forward and reverse primers (Eton Biosciences) and analyzed with 4peaks software. Gene, primer, and vector sequences can be found in the supplementary information (Supplementary Table 3).
The AZIN1-antizyme binding affinity was measured by a FRET assay as previously described16. In brief, “purified recombinant protein stocks in 50% glycerol were diluted with TBS (pH 7.4) + 0.15% Tween-20 to 2X final working concentration (100 nM for Clover containing donor proteins and 2 µM mRuby2 acceptor proteins) and transferred to a 96-well plate (Corning 3821). Specifically, acceptor protein or diluent control was transferred into a low-volume 384 well plate (CoStar 3356) in triplicate and serially diluted. Plates were equilibrated at room temp for 1 h before reading.
Fluorescence was measured using an EnVision plate reader (Perkin Elmer) with a 470/40 excitation filter and either a 515/30 or 600/8 emission filter to measure Clover or mRuby-2 fluorescence, respectively. Curve fitting on donor fluorescence (quenching) was performed in GraphPad Prism v. 7, using global fitting for maximum and minimum fluorescence intensity, constraining the Hill coefficient to 1, and fitting the IC50. Similar IC50s were obtained when fitting was performed on acceptor sensitization after excluding acceptor concentrations above 100 nM. In our in vitro FRET determination for AZIN and antizyme, we observed 17% donor quenching and the quantum yield for mRuby2 is 0.38, predicting a 6.4% acceptor sensitization”.
Droplet Digital PCR (ddPCR)
The ddPCR primers Forward (GAGCCTCTGTTTACAAGCAG) and Reverse (CATGGAAAGAATCTGCTCCC) and probes wtAZIN (5’-/5HEX/GCTCAGGAAGAAGACAGCTTTCCAC/3IABkFQ/-‘3) and edAZIN (5’-/56-FAM/GCTCAGGAAGAAGACAGCCTTCCA/3IABkFQ/-‘3) used in this study were designed using Primer3 software to target AZIN. Droplet digital PCR was performed as follows. The PCR mixture contained ddPCR Super Mix (Bio-Rad, Hercules, CA; final concentration: 1X), the wtAZIN and edAZIN probes (0.25 μM each), forward and reverse primers (1 μM each), and up to 30 ng of template DNA in a 25 μL total volume. The reaction mixture was emulsified into approximately 16–17,000 droplets using a QX100 Droplet Generator (Bio-Rad) according to the manufacturer’s instructions. PCR was performed (10 min. at 98 °C; 40 cycles of 30 s at 94 °C, 60 s at 58 °C, and 20 s at 72 °C; 10 min at 72 °C; and holding at 12 °C); the samples were analyzed within 24 h using a QX100 Droplet Reader (Bio-Rad), and the data were analyzed with QuantaSoft software (Bio-Rad).
RNA extraction and Illumina mRNA library preparation
HEK293 cells were transfected for 24 h with plasmids expressing fluorescent protein (Empty); wild-type AZIN1 (wt), which can be edited endogenously; pseudoedited AZIN1 (ed); or AZIN1 with an uneditable codon 367 (uneditable-wt). RNA was extracted using a Qiagen kit and was then sequenced and analyzed by Macrogen as described17.
Protein concentrations in lysates were quantified, and proteins were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS‒PAGE). After separation, proteins were transferred onto a polyvinylidene difluoride (PVDF) membrane (Bio-Rad Laboratories, Inc. and incubated with a primary antibody followed by a secondary antibody. Luminescence was visualized with Enhanced Chemiluminescence Substrate (ECL) (Amersham Biosciences, Sweden). The results of western blotting were analyzed with NIH ImageJ 1.62 software. The following antibodies were used: anti-ACTG1 (actin gamma 1) and anti-ACTA2 (alpha-smooth muscle actin) from Sigma–Aldrich, anti-AZIN1 from Abbexa, and anti-ADAR1, anti-Myosin-9, anti-GAPDH, anti-α-Tubulin, anti-snail, anti-slug, anti-MMP2, anti-MMP9 and anti-Rpb1 CTD from Cell Signaling.
Mysoin-9 CRISPR knockout cells
We digested lentiCRISPRv2 blast (Addgene, cat# 98293) and/or lentiCRISPRv2 hygro (Addgene, cat# 98291) with the BsmB1 restriction enzyme (NEB cat# R0739S) at 55 °C for 1 h and performed gel purification (0.7 g agarose in 100 ml of TAE buffer) using a gel extraction kit (NEB cat# T1020S). sgRNAs were selected using the ChopChop website (https://chopchop.cbu.uib.no/). Oligos/sgRNAs were purchased from Eton Bioscience (Supplementary Table 4) and were phosphorylated by the addition of T4 PNK in T4 DNA ligase buffer (37 °C for 50 min). Thereafter, the oligos were annealed by holding at 95 °C for 5 min followed by ramping down to 25 °C at 5 degrees/minute. The oligos and plasmids were ligated by incubation with Quick Ligase for 10 min at RT and were then transformed into DH5alpha (NEB cat# C2987I) following the manufacturer’s instructions. Following sequence confirmation and purification with a maxiprep kit (Qiagen), the plasmids together with envelope pVSV-G plasmid (Addgene #138479) and packaging psPAX plasmid (Addgene #12260) were transfected into HEK295T cells for 24 h, and the medium containing lentivirus was then changed and extracted after an additional 24 and 48 h, respectively. DU145 and PC3 cells were exposed to the lentivirus-containing medium for 4 h, and fresh medium was then added to the cells for an additional 24 h. Antibiotic selection was performed by treating the cells with 5 ng/µl blasticidin and/or 250 µg/µl hygromycin for 4 days. Myosin 9 knockout was confirmed by western blotting.
ADAR1 CRISPR knockout cells
The ADAR1 genomic locus was modified, which resulted in abolishing ADAR1 expression in the HEK293T cell line, by using CRISPR-Cas9 genome editing as described earlier13.
Soft agar colony formation assays
A total of 5 × 103 cells in 0.4% Bacto agar were seeded on top of a solidified layer of 0.6% Bacto agar in 6-well plates. Colonies consisting of more than 50 cells were counted after 19 days, and the data are expressed as the mean ± s.e.m. of triplicate wells in the same experiment.
Matrigel invasion assay
We performed the invasion assay using 24-well BioCoat Matrigel Invasion Chambers (BD Biosciences) according to the manufacturer’s instructions. Briefly, 2 × 105 cells were seeded in the top compartment, and DMEM containing 10% FBS was added to the bottom compartment as a chemoattractant. After 24 h of incubation, cells that invaded the Matrigel were fixed and stained with crystal violet (Sigma‒Aldrich). Cells were counted in 10 fields of view under a 20× objective and imaged using SPOT imaging software (Nikon, Japan).
HEK293 cells were transfected with 2.0 μg of the Clover-tagged wild-type or edited AZIN1 plasmid (GFP-wtAZIN1 or GFP-edAZIN1) and analyzed via Co-IP using a V5-specific antibody. Approximately 10 mg of the total cell lysate was immunoprecipitated with 5 μg of an anti-FLAG antibody at 4 °C overnight. Immunocomplexes were then precipitated using 100 μl of protein G-agarose, which was provided in the immunoprecipitation kit (Roche Diagnostics Co., Indianapolis, IN). After extensive washing in washing buffer, the beads were boiled in 50 μl of loading buffer and analyzed by western blotting using antibodies against myosin-9 (cell signaling), ACTG1 (actin gamma 1-), and ACTA2 (alpha-smooth muscle actin) from Sigma–Aldrich. We utilized 5% of the total lysate (5% input) as a positive control. Mouse immunoglobulin G (Santa Cruz Biotechnology) was used as a negative control.
Patients and tissue microarray (TMA)
Prostate cancer patients (n = 202) who underwent radical prostatectomy at Massachusetts General Hospital (Boston, MA) between September 1993 and March 1995 were included. Patients who received neoadjuvant hormonal treatment or adjuvant hormonal and/or radiation treatment before recurrence were excluded. The Gleason score was reassigned based on the current International Society of Urological Pathology (ISUP) recommendation18. TMAs were constructed as previously described19. Index tumor foci from each case were selected for inclusion in the TMA along with 26 adjacent benign tissues. The study protocol was approved by the human study committees at MGH (IRB# 2005P000774). For the high Gleason score cohort, patients who underwent radical prostatectomy for localized PC between 1993 and 2007 were reviewed, and patients with Gleason scores of 7 or higher were included in this study.
TMAs were deparaffinized with xylene, rehydrated, and subjected to brief proteolytic digestion and peroxidase blocking. Slides were incubated overnight at 4 °C with a 1:400 dilution of a polyclonal anti-antizyme inhibitor 1 antibody (#11548–1-AP; Proteintech Group, Inc. Chicago, IL). The antibody was validated using cells overexpressing AZIN1 plasmids. After washing, a peroxidase-labeled polymer and the substrate chromogen TMB were used to visualize the staining of the proteins of interest (DAKO EnVision System, Dako Diagnostics, Zug, Switzerland). Normal prostate tissues were used as controls. Slides were scored independently by two board-certified pathologists blinded to the patient data, and discrepancies were resolved through concurrent re-examination of the slide to arrive at a consensus. The percentage of cells showing nuclear or cytoplasmic AZIN staining was tabulated; specimens were scored as positive if >3% tumor cells exhibited immunoreactivity.
In the tissue microarray cohort, PSA failure-free survival was defined as the time between the initial surgery and the appearance of detectable PSA in patients with two consecutive increases in PSA, and metastasis-free survival was defined as the time to clinical or radiographic detection of metastasis.
Data from The Cancer Genome Atlas (TCGA) Prostate Adenocarcinoma (PRAD) cohort of patients with primary prostate cancer20 were obtained using previously described methods21. Analyses were restricted to the set of n = 333 tumors with high-quality RNA sequencing data. The edAZIN1 level was quantified by Han et al.6 in n = 292 tumors. All tumors with missing edAZIN1 calls had missing PSA values but otherwise similar characteristics to tumors with edAZIN1 calls. Grading of cribriform morphology was performed as described in22.
To quantify associations between edAZIN1 and mRNA expression levels, Pearson correlation analysis was used after log-transforming mRNA levels. Predictors of high edAZIN1 levels (ADAR mRNA expression, fraction of the genome altered, age at diagnosis, and AZIN1 mRNA level) were assessed using multivariable linear regression, and all predictors were standardized to interquartile ranges (25-percentile increase) for comparability between predictors.
For survival analyses in the tissue microarray cohort and the TCGA cohort, Cox proportional hazards regression analysis was used to estimate hazard ratios and 95% CIs. Multiplicative effect measure modification was evaluated using the Wald test for the product term (pinteraction).