Ribonucleotide reductase, a possible agent in deoxyribonucleotide pool asymmetries induced by hypoxia

Still no Rest for the Reductases: Ribonucleotide Reductase (RNR) Structure and Function: An Update

Herein we present a multidisciplinary discussion of ribonucleotide reductase (RNR), the essential enzyme uniquely responsible for conversion of ribonucleotides to deoxyribonucleotides. This chapter primarily presents an overview of this multifaceted and complex enzyme, covering RNR's role in enzymology, biochemistry, medicinal chemistry, and cell biology. It further focuses on RNR from mammals, whose interesting and often conflicting roles in health and disease are coming more into focus. We present pitfalls that we think have not always been dealt with by researchers in each area and further seek to unite some of the field-specific observations surrounding this enzyme. Our work is thus not intended to cover any one topic in extreme detail, but rather give what we consider to be the necessary broad grounding to understand this critical enzyme holistically. Although this is an approach we have advocated in many different areas of scientific research, there is arguably no other single enzyme that embodies the need for such broad study than RNR. Thus, we submit that RNR itself is a paradigm of interdisciplinary research that is of interest from the perspective of the generalist and the specialist alike. We hope that the discussions herein will thus be helpful to not only those wanting to tackle RNR-specific problems, but also those working on similar interdisciplinary projects centering around other enzymes.

Tipping cancer cells over the edge: the context-dependent cost of high ploidy

Tetraploidy is an aneuploidy-permissive condition that can fuel tumorgenesis. The tip-over hypothesis of cytotoxic therapy-sensitivity proposes that therapy is effective if it pushes a cell's aneuploidy above a viable tipping point. But elevated aneuploidy alone may not account for this tipping point. Tissue micro-environments (TMEs) that lack sufficient resources to support tetraploid cells can explain the fitness cost of aneuploidy. Raw materials needed to generate deoxynucleotides, the building blocks of DNA, are candidate rate-limiting factors for the evolution of high-ploidy cancer cells. Understanding the resource cost of high ploidy is key to uncover its therapeutic vulnerabilities across tissue sites with versatile energy supplies.

Ribonucleotide reductase inhibitors suppress SAMHD1 ara-CTPase activity enhancing cytarabine efficacy

The deoxycytidine analogue cytarabine (ara-C) remains the backbone treatment of acute myeloid leukaemia (AML) as well as other haematological and lymphoid malignancies, but must be combined with other chemotherapeutics to achieve cure. Yet, the underlying mechanism dictating synergistic efficacy of combination chemotherapy remains largely unknown. The dNTPase SAMHD1, which regulates dNTP homoeostasis antagonistically to ribonucleotide reductase (RNR), limits ara-C efficacy by hydrolysing the active triphosphate metabolite ara-CTP. Here, we report that clinically used inhibitors of RNR, such as gemcitabine and hydroxyurea, overcome the SAMHD1-mediated barrier to ara-C efficacy in primary blasts and mouse models of AML, displaying SAMHD1-dependent synergy with ara-C. We present evidence that this is mediated by dNTP pool imbalances leading to allosteric reduction of SAMHD1 ara-CTPase activity. Thus, SAMHD1 constitutes a novel biomarker for combination therapies of ara-C and RNR inhibitors with immediate consequences for clinical practice to improve treatment of AML.

Association of genetic polymorphisms in genes involved in Ara-C and dNTP metabolism pathway with chemosensitivity and prognosis of adult acute myeloid leukemia (AML)

Background

Cytarabine arabinoside (Ara-C) has been the core of chemotherapy for adult acute myeloid leukemia (AML). Ara-C undergoes a three-step phosphorylation into the active metabolite Ara-C triphosphosphate (ara-CTP). Several enzymes are involved directly or indirectly in either the formation or detoxification of ara-CTP.

Methods

A total of 12 eQTL (expression Quantitative Trait Loci) single nucleotide polymorphisms (SNPs) or tag SNPs in 7 genes including CMPK1, NME1, NME2, RRM1, RRM2, SAMHD1 and E2F1 were genotyped in 361 Chinese non-M3 AML patients by using the Sequenom Massarray system. Association of the SNPs with complete remission (CR) rate after Ara-C based induction therapy, relapse-free survival (RFS) and overall survival (OS) were analyzed.

Results

Three SNPs were observed to be associated increased risk of chemoresistance indicated by CR rate (NME2 rs3744660, E2F1 rs3213150, and RRM2 rs1130609), among which two (rs3744660 and rs1130609) were eQTL. Combined genotypes based on E2F1 rs3213150 and RRM2 rs1130609 polymorphisms further increased the risk of non-CR. The SAMHD1 eQTL polymorphism rs6102991 showed decreased risk of non-CR marginally (P = 0.055). Three SNPs (NME1 rs3760468 and rs2302254, and NME2 rs3744660) were associated with worse RFS, and the RRM2 rs1130609 polymorphism was marginally associated with worse RFS (P = 0.085) and OS (P = 0.080). Three SNPs (NME1 rs3760468, NME2 rs3744660, and RRM1 rs183484) were associated with worse OS in AML patients.

Conclusion

Data from our study demonstrated that SNPs in Ara-C and dNTP metabolic pathway predict chemosensitivity and prognosis of AML patients in China.

Electronic supplementary material

The online version of this article (10.1186/s12967-018-1463-1) contains supplementary material, which is available to authorized users.

EGFR Mutations Compromise Hypoxia-Associated Radiation Resistance through Impaired Replication Fork-Associated DNA Damage Repair

EGFR signaling has been implicated in hypoxia-associated resistance to radiation or chemotherapy. Non-small cell lung carcinomas (NSCLC) with activating L858R or ΔE746-E750 EGFR mutations exhibit elevated EGFR activity and downstream signaling. Here, relative to wild-type (WT) EGFR, mutant (MT) EGFR expression significantly increases radiosensitivity in hypoxic cells. Gene expression profiling in human bronchial epithelial cells (HBEC) revealed that MT-EGFR expression elevated transcripts related to cell cycle and replication in aerobic and hypoxic conditions and downregulated RAD50, a critical component of nonhomologous end joining and homologous recombination DNA repair pathways. NSCLCs and HBEC with MT-EGFR revealed elevated basal and hypoxia-induced γ-H2AX-associated DNA lesions that were coincident with replication protein A in the S-phase nuclei. DNA fiber analysis showed that, relative to WT-EGFR, MT-EGFR NSCLCs harbored significantly higher levels of stalled replication forks and decreased fork velocities in aerobic and hypoxic conditions. EGFR blockade by cetuximab significantly increased radiosensitivity in hypoxic cells, recapitulating MT-EGFR expression and closely resembling synthetic lethality of PARP inhibition.Implications: This study demonstrates that within an altered DNA damage response of hypoxic NSCLC cells, mutant EGFR expression, or EGFR blockade by cetuximab exerts a synthetic lethality effect and significantly compromises radiation resistance in hypoxic tumor cells. Mol Cancer Res; 15(11); 1503-16. ©2017 AACR.

Ribonucleotide Reductase Requires Subunit Switching in Hypoxia to Maintain DNA Replication

Cells exposed to hypoxia experience replication stress but do not accumulate DNA damage, suggesting sustained DNA replication. Ribonucleotide reductase (RNR) is the only enzyme capable of de novo synthesis of deoxyribonucleotide triphosphates (dNTPs). However, oxygen is an essential cofactor for mammalian RNR (RRM1/RRM2 and RRM1/RRM2B), leading us to question the source of dNTPs in hypoxia. Here, we show that the RRM1/RRM2B enzyme is capable of retaining activity in hypoxia and therefore is favored over RRM1/RRM2 in order to preserve ongoing replication and avoid the accumulation of DNA damage. We found two distinct mechanisms by which RRM2B maintains hypoxic activity and identified responsible residues in RRM2B. The importance of RRM2B in the response to tumor hypoxia is further illustrated by correlation of its expression with a hypoxic signature in patient samples and its roles in tumor growth and radioresistance. Our data provide mechanistic insight into RNR biology, highlighting RRM2B as a hypoxic-specific, anti-cancer therapeutic target.

•

RRM2B is induced in response to hypoxia in both cell models and patient datasets

•

RRM2B retains activity in hypoxic conditions and is the favored RNR subunit in hypoxia

•

Loss of RRM2B has detrimental consequences for cell fate, specifically in hypoxia

•

RRM2B depletion enhanced hypoxic-specific apoptosis and increased radiosensitivity

Visualizing cell-cycle kinetics after hypoxia/reoxygenation in HeLa cells expressing fluorescent ubiquitination-based cell cycle indicator (Fucci)

Hypoxia induces G1 arrest in many cancer cell types. Tumor cells are often exposed to hypoxia/reoxygenation, especially under acute hypoxic conditions in vivo. In this study, we investigated cell-cycle kinetics and clonogenic survival after hypoxia/reoxygenation in HeLa cells expressing fluorescent ubiquitination-based cell cycle indicator (Fucci). Hypoxic treatment halted cell-cycle progression during mid-S to G2 phase, as determined by the cell cycle-regulated E3 ligase activities of SCF(Skp2) and APC/C(Cdh1), which are regulators of the Fucci probes; however, the DNA content of the arrested cells was equivalent to that in G1 phase. After reoxygenation, time-lapse imaging and DNA content analysis revealed that all cells reached G2 phase, and that Fucci fluorescence was distinctly separated into two fractions 24h after reoxygenation: red cells that released from G2 arrest after repairing DNA double-strand breaks (DSBs) exhibited higher clonogenic survival, whereas most cells that stayed green contained many DSBs and exhibited lower survival. We conclude that hypoxia disrupts coordination of DNA synthesis and E3 ligase activities associated with cell-cycle progression, and that DSB repair could greatly influence cell-cycle kinetics and clonogenic survival after hypoxia/reoxygenation.

Quantitative Targeted Proteomics of Pancreatic Cancer: Deoxycytidine Kinase Protein Level Correlates to Progression-Free Survival of Patients Receiving Gemcitabine Treatment

The purpose of the present study is to identify the determinant(s) of gemcitabine (dFdC)-sensitivity in pancreatic cancer tissues of patients treated with dFdC alone and in pancreatic cancer cell lines exposed to dFdC in vitro. Protein expression levels of 12 enzymes and 13 transporters potentially involved in transport and metabolism of dFdC in pancreatic cancer cell lines and tissues were quantified by means of our LC-MS/MS-based quantitative targeted proteomics technology. Protein expression levels of deoxycytidine kinase (dCK), uridine monophosphate-cytidine monophosphate (UMP-CMP) kinase, cytosolic nucleotidase III (cN-III), and equilibrative nucleoside transporter 1 (ENT1) were significantly correlated with IC50 or 1/IC50 in five cell lines with different sensitivities to dFdC (p < 0.05). Expression levels of the selected proteins in pancreatic cancer tissues of 10 patients with different progression-free survival (PFS) (49-955 days) were quantified, and their relationship with PFS was examined. Only the protein expression level of dCK was significantly correlated with PFS (p < 0.05). Multiple regression analysis was also performed, and combinations of ENT1, UMP-CMP kinase, CTPS1, and dCK were highly correlated with PFS. Our results indicate that the protein expression level of dCK in pancreatic cancer tissue is a good predictor of PFS, and thus dCK may be the best biomarker of dFdC sensitivity in pancreatic cancer patients treated with dFdC, although other proteins would also contribute to dFdC-sensitivity at the cellular level in vivo and in vitro.

Studies of ribonucleotide reductase in crucian carp-an oxygen dependent enzyme in an anoxia tolerant vertebrate

The enzyme ribonucleotide reductase (RNR) catalyzes the conversion of ribonucleotides to deoxyribonucleotides, the precursors for DNA. RNR requires a thiyl radical to activate the substrate. In RNR of eukaryotes (class Ia RNR), this radical originates from a tyrosyl radical formed in reaction with oxygen (O(2)) and a ferrous di-iron center in RNR. The crucian carp (Carassius carassius) is one of very few vertebrates that can tolerate several months completely without oxygen (anoxia), a trait that enables this fish to survive under the ice in small ponds that become anoxic during the winter. Previous studies have found indications of cell division in this fish after 7 days of anoxia. This appears nearly impossible, as DNA synthesis requires the production of new deoxyribonucleotides and therefore active RNR. We have here characterized RNR in crucian carp, to search for adaptations to anoxia. We report the full-length sequences of two paralogs of each of the RNR subunits (R1i, R1ii, R2i, R2ii, p53R2i and p53R2ii), obtained by cloning and sequencing. The mRNA levels of these subunits were measured with quantitative PCR and were generally well maintained in hypoxia and anoxia in heart and brain. We also report maintained or increased mRNA levels of the cell division markers proliferating cell nuclear antigen (PCNA), brain derived neurotrophic factor (BDNF) and Ki67 in anoxic hearts and brains. Electron paramagnetic resonance (EPR) measurements on in vitro expressed crucian carp R2 and p53R2 proteins gave spectra similar to mammalian RNRs, including previously unpublished human and mouse p53R2 EPR spectra. However, the radicals in crucian carp RNR small subunits, especially in the p53R2ii subunit, were very stable at 0°C. A long half-life of the tyrosyl radical during wintertime anoxia could allow for continued cell division in crucian carp.

Significance of the Cgl1427 gene encoding cytidylate kinase in microaerobic growth of Corynebacterium glutamicum

The Cgl1427 gene was previously found to be relevant to the microaerobic growth of Corynebacterium glutamicum (Ikeda et al. Biosci Biotechnol Biochem 73:2806-2808, 2009). In the present work, Cgl1427 was identified as a cytidylate kinase gene (cmk) by homology analysis of its deduced amino acid sequence with that of other bacterial cytidylate kinases (CMP kinases) and on the basis of findings that deletion of Cgl1427 results in loss of CMP kinase activity. Deletion of the cmk gene significantly impaired the growth of C. glutamicum in oxygen-limiting static culture, and the impaired growth was restored by introducing a plasmid containing the cmk gene, suggesting that this gene plays an important role in the microaerobic growth of C. glutamicum. On the other hand, in the main culture with aerobic shaking, a prolonged lag phase was observed in the cmk disruptant, despite an unchanged growth rate, compared to the behavior of the wild-type strain. The prolongation was observed when using seed culture grown to later growth stages in which oxygen limitation occurred, but it was not observed when using seed culture grown to an earlier growth stage in which oxygen remained relatively plentiful. Since nucleotide biosynthesis in C. glutamicum requires oxygen, we hypothesized that the ability of the cmk disruptant to synthesize nucleotides was influenced by oxygen limitation in the later growth stages of the seed culture, which caused the prolongation of the lag phase in the following shaken culture. To verify this hypothesis, a plasmid containing genes encoding all components of a homologous ribonucleotide reductase, a key enzyme for nucleotide synthesis that requires oxygen for its reaction, was introduced into the cmk disruptant, which significantly ameliorated the lag phase prolongation. Furthermore, this experimental setup almost completely restored the growth of the cmk disruptant in the oxygen-limiting static culture. These results indicate that CMP kinase plays an important role in normal nucleotide biosynthesis under an oxygen-limiting environment.

E2F4 and ribonucleotide reductase mediate S-phase arrest in colon cancer cells treated with chlorophyllin

Chlorophyllin (CHL) is a water-soluble derivative of chlorophyll that exhibits cancer chemopreventive properties, but which also has been studied for its possible cancer therapeutic effects. We report here that human colon cancer cells treated with CHL accumulate in S-phase of the cell cycle, and this is associated with reduced expression levels of p53, p21, and other G(1)/S checkpoint controls. At the same time, E2F1 and E2F4 transcription factors become elevated and exhibit increased DNA binding activity. In CHL-treated colon cancer cells, bromodeoxyuridine pulse-chase experiments provided evidence for the inhibition of DNA synthesis. Ribonucleotide reductase (RR), a pivotal enzyme for DNA synthesis and repair, was reduced at the mRNA and protein level after CHL treatment, and the enzymatic activity was inhibited in a concentration-dependent manner both in vitro and in vivo. Immunoblotting revealed that expression levels of RR subunits R1, R2, and p53R2 were reduced by CHL treatment in HCT116 (p53(+/+)) and HCT116 (p53(-/-)) cells, supporting a p53-independent mechanism. Prior studies have shown that reduced levels of RR small subunits can increase the sensitivity of colon cancer cells to clinically used DNA-damaging agents and RR inhibitors. We conclude that by inhibiting R1, R2, and p53R2, CHL has the potential to be effective in the clinical setting, when used alone or in combination with currently available cancer therapeutic agents.

Influence of vector design and host cell on the mechanism of recombination and emergence of mutant subpopulations of replicating retroviral vectors

Background

The recent advent of murine leukaemia virus (MLV)-based replication-competent retroviral (RCR) vector technology has provided exciting new tools for gene delivery, albeit the advances in vector efficiency which have been realized are also accompanied by a set of fresh challenges. The expression of additional transgene sequences, for example, increases the length of the viral genome, which can lead to reductions in replication efficiency and in turn to vector genome instability. This necessitates efforts to analyse the rate and mechanism of recombinant emergence during the replication of such vectors to provide data which should contribute to improvements in RCR vector design.

Results

In this study, we have performed detailed molecular analyses on packaged vector genomes and proviral DNA following propagation of MLV-based RCR vectors both in cell culture and in pre-formed subcutaneous tumours in vivo. The effects of strain of MLV, transgene position and host cell type on the rate of emergence of vector recombinants were quantitatively analysed by applying real-time PCR and real-time RT-PCR assays. Individual mutants were further characterized by PCR, and nucleotide sequence and structural motifs associated with these mutants were determined by sequencing. Our data indicate that virus strain, vector design and host cell influence the rate of emergence of predominating vector mutants, but not the underlying recombination mechanisms in vitro. In contrast, however, differences in the RNA secondary structural motifs associated with sequenced mutants emerging in cell culture and in solid tumours in vivo were observed.

Conclusion

Our data provide further evidence that MLV-based RCR vectors based on the Moloney strain of MLV and containing the transgene cassette in the 3' UTR region are superior to those based on Akv-MLV and/or containing the transgene cassette in the U3 region of the LTR. The observed discrepancies between the data obtained in solid tumours in vivo and our own and previously published data from infected cells in vitro demonstrates the importance of evaluating vectors designed for use in cancer gene therapy in vivo as well as in vitro.

A mutant allele of MRE11 found in mismatch repair-deficient tumor cells suppresses the cellular response to DNA replication fork stress in a dominant negative manner

The interaction of ataxia-telangiectasia mutated (ATM) and the Mre11/Rad50/Nbs1 (MRN) complex is critical for the response of cells to DNA double-strand breaks; however, little is known of the role of these proteins in response to DNA replication stress. Here, we report a mutant allele of MRE11 found in a colon cancer cell line that sensitizes cells to agents causing replication fork stress. The mutant Mre11 weakly interacts with Rad50 relative to wild type and shows little affinity for Nbs1. The mutant protein lacks 3'-5' exonuclease activity as a result of loss of part of the conserved nuclease domain; however, it retains binding affinity for single-stranded DNA (ssDNA), double-stranded DNA with a 3' single-strand overhang, and fork-like structures containing ssDNA regions. In cells, the mutant protein shows a time- and dose-dependent accumulation in chromatin after thymidine treatment that corresponds with increased recruitment and hyperphosphorylation of replication protein A. ATM autophosphorylation, Mre11 foci, and thymidine-induced homologous recombination are suppressed in cells expressing the mutant allele. Together, our results suggest that the mutant Mre11 suppresses the cellular response to replication stress by binding to ssDNA regions at disrupted forks and impeding replication restart in a dominant negative manner.

Apoptosis induced by replication inhibitors in Chk1-depleted cells is dependent upon the helicase cofactor Cdc45

Checkpoint kinase 1 (Chk1) responds to disruption of DNA replication to maintain the integrity of stalled forks, promote homologous recombination-mediated repair of replication fork lesions, and control inappropriate firing of replication origins. This response is essential for viability as replication inhibitors trigger apoptosis in S-phase cells depleted of Chk1. Given the complex network of cellular responses controlled by Chk1, our aim was to determine which of these protect cells from apoptosis following replication stress. Work with cell-free systems has shown that RPA-ssDNA complex forms following replication inhibition through the uncoupling of replication and helicase complexes. Here we show that replication protein A (RPA) foci form in cells treated with replication inhibitors and that the number of foci dramatically increases together with hyperphosphorylation of RPA34 in Chk1-depleted cells in advance of the induction of apoptosis. RPA foci, RPA34 hyperphosphorylation, and apoptosis were suppressed by siRNA-mediated knockdown of Cdc45, an essential replication helicase cofactor required for both the initiation and elongation steps of DNA replication. In contrast, loss of p21, a negative effector of origin firing, stimulates both the accumulation of RPA foci and apoptosis. Taken together, these results suggest that the loss of control of replication origin firing following Chk1 depletion triggers the accumulation of the RPA-ssDNA complex and apoptosis when replication is blocked.

Transcriptional repression of human cad gene by hypoxia inducible factor-1alpha

De novo biosynthesis of pyrimidine nucleotides provides essential precursors for DNA synthesis and cell proliferation. The first three steps of de novo pyrimidine biosynthesis are catalyzed by a multifunctional enzyme known as CAD (carbamoyl phosphate synthetase-aspartate carbamoyltransferase-dihydroorotase). In this work, a decrease in CAD expression is detected in numerous cell lines and primary culture human stromal cells incubated under hypoxia or desferrioxamine (DFO)-induced HIF-1α accumulation. A putative hypoxia response element (HRE) binding matrix is identified by analyzing human cad-gene promoter using a bioinformatic approach. Promoter activity assays, using constructs harboring the cad promoter (−710/+122) and the −67/HRE fragment (25-bases), respectively, demonstrate the suppression of reporter-gene expression under hypoxia. Suppression of cad-promoter activity is substantiated by forced expression of wild-type HIF-1α but abolished by overexpression of dominant-negative HIF-1α. A chromatin immunoprecipitation assay provides further evidence that HIF-1α binds to the cad promoter in vivo. These data demonstrate that the cad-gene expression is repressed by HIF-1α, which represents a functional link between hypoxia and cell-cycle arrest.

A macroscopic kinetic model for DNA polymerase elongation and high-fidelity nucleotide selection

The enzymatically catalyzed template-directed extension of ssDNA/primer complex is an important reaction of extraordinary complexity. The DNA polymerase does not merely facilitate the insertion of dNMP, but it also performs rapid screening of substrates to ensure a high degree of fidelity. Several kinetic studies have determined rate constants and equilibrium constants for the elementary steps that make up the overall pathway. The information is used to develop a macroscopic kinetic model, using an approach described by Ninio [Ninio J., 1987. Alternative to the steady-state method: derivation of reaction rates from first-passage times and pathway probabilities. Proc. Natl. Acad. Sci. U.S.A. 84, 663-667]. The principle idea of the Ninio approach is to track a single template/primer complex over time and to identify the expected behavior. The average time to insert a single nucleotide is a weighted sum of several terms, including the actual time to insert a nucleotide plus delays due to polymerase detachment from either the ternary (template-primer-polymerase) or quaternary (+nucleotide) complexes and time delays associated with the identification and ultimate rejection of an incorrect nucleotide from the binding site. The passage times of all events and their probability of occurrence are expressed in terms of the rate constants of the elementary steps of the reaction pathway. The model accounts for variations in the average insertion time with different nucleotides as well as the influence of G + C content of the sequence in the vicinity of the insertion site. Furthermore the model provides estimates of error frequencies. If nucleotide extension is recognized as a competition between successful insertions and time delaying events, it can be described as a binomial process with a probability distribution. The distribution gives the probability to extend a primer/template complex with a certain number of base pairs and in general it maps annealed complexes into extension products.

Cell proliferation and gill morphology in anoxic crucian carp

Is DNA replication/cell proliferation in vertebrates possible during anoxia? The oxygen dependence of ribonucleotide reductase (RNR) could lead to a stop in DNA synthesis, thereby making anoxic DNA replication impossible. We have studied this question in an anoxia-tolerant vertebrate, the crucian carp (Carassius carassius), by examining 5'-bromo-2'-deoxyuridine incorporation and proliferating cell nuclear antigen levels in the gills, intestinal crypts, and liver. We exposed crucian carp to 1 and 7 days of anoxia followed by 7 days of reoxygenation. There was a reduced incidence of S-phase cells (from 12.2 to 5.0%) in gills during anoxia, which coincided with a concomitant increase of G(0) cells. Anoxia also decreased the number of S-phase cells in intestine (from 8.1 to 1.8%). No change in the fraction of S-phase cells ( approximately 1%) in liver was found. Thus new S-phase cells after 7 days of anoxia were present in all tissues, revealing a considerable rate of DNA synthesis. Subsequently, the oxygen-dependent subunit of crucian carp RNR (RNRR2) was cloned. We found no differences in amino acids involved in radical generation and availability of the iron center compared with mouse, which could have explained reduced oxygen dependence. Furthermore, the amount of RNRR2 mRNA in gills did not decrease throughout anoxia exposure. These results indicate that crucian carp is able to sustain some cell proliferation in anoxia, possibly because RNRR2 retains its tyrosyl radical in anoxia, and that the replication machinery is still maintained. Although hypoxia triggers a 7.5-fold increase of respiratory surface area in crucian carp, this response was not triggered in anoxia.