Effects of Ionizing Radiation on Biological Molecules—Mechanisms of Damage and Emerging Methods of Detection

Abstract

Significance: The detrimental effects of ionizing radiation (IR) involve a highly orchestrated series of events that are amplified by endogenous signaling and culminating in oxidative damage to DNA, lipids, proteins, and many metabolites. Despite the global impact of IR, the molecular mechanisms underlying tissue damage reveal that many biomolecules are chemoselectively modified by IR. Recent Advances: The development of high-throughput “omics” technologies for mapping DNA and protein modifications have revolutionized the study of IR effects on biological systems. Studies in cells, tissues, and biological fluids are used to identify molecular features or biomarkers of IR exposure and response and the molecular mechanisms that regulate their expression or synthesis. Critical Issues: In this review, chemical mechanisms are described for IR-induced modifications of biomolecules along with methods for their detection. Included with the detection methods are crucial experimental considerations and caveats for their use. Additional factors critical to the cellular response to radiation, including alterations in protein expression, metabolomics, and epigenetic factors, are also discussed. Future Directions: Throughout the review, the synergy of combined “omics” technologies such as genomics and epigenomics, proteomics, and metabolomics is highlighted. These are anticipated to lead to new hypotheses to understand IR effects on biological systems and improve IR-based therapies. Antioxid. Redox Signal. 21: 260–292.

Introduction

Radiation is a phenomenon present in our daily lives, originating from natural and manmade sources. Living organisms are profoundly affected by radiation-induced cellular damage, threatening healthy and diseased tissues alike. In humans, there is a wide range of response to radiation, which is determined by parameters including the radiation source, radiation dosage (amount of radiation energy received), length of exposure, and, importantly, the genetic and epigenetic makeup of the exposed individual. These parameters can range widely, and humans may be exposed to low-dose radiation from commonly used diagnostic tools in medicine such as computed tomography (CT) scanning or high doses of radiation such as those used for radiotherapy and generated by nuclear disasters. The genetic and epigenetic aspects are significant across many conditions and may determine, for example, the likelihood of an individual to develop cancer or to respond to a cancer treatment (e.g., some tumors have either an intrinsic resistance to ionizing radiation (IR) or have acquired this property through cycled accumulation of genetic mutations and selection for increased survival and proliferation).

Though much progress has been made in understanding the basic principles of IR-induced effects on individual components of biological systems, less is known about how localized IR effects on target molecules regulate the cellular networks that contain these modified species, the interactions between networks (e.g., signaling and bioenergy metabolism), and the overall state of a biological system (cell, tissue/tumor, and individual/patient). The availability of high-end technologies and detection methods to monitor single or global radiation-induced modification of biomolecules has increased sharply in recent years, uncovering a much wider context for how radiation impacts the cellular life cycle. In this review, we highlight many recently developed techniques for uncovering radiation targets, though to date, not all have been applied toward further elucidating the unique biological response of radiation. The current review will discuss (i) the impact of IR on biological macromolecules (nucleic acids/DNA, lipids, and proteins); (ii) classical and modern methods of detection of IR-modified species; (iii) cellular processes affected by the interaction of IR with DNA, lipids, and proteins; and (iv) how state-of-the-art “omics” methods and technologies could be applied to decipher the complex interaction networks that exist between the products of IR-induced modifications (e.g., DNA damage, lipid peroxidation, and protein oxidation).

Fundamentals of IR Biochemistry

This chapter discusses the fundamentals of IR in the context of other types of radiation along with the key effector molecules that are responsible for the transforming effects of IR on biomolecules.

Fundamentals

Radiation is classified in two major forms: ionizing and non-ionizing. Environmental radiation is largely the non-ionizing type, including ultraviolet (UV) rays from the sun and electromagnetic radiation associated with radio waves and microwaves. The ability of sources of non-IR, such as UV rays (from the sun or tanning beds), to harm biological tissues is now well-established (168, 382). The interaction of IR with biomolecules is, however, much more aggressive than non-IR due to the ability of IR to induce atom ionization. The source of IR is a class of unstable radionuclides (radioisotopes) that emit high-energy particles which are capable of displacing atomic electrons, facilitating a chain reaction of electron ejection. The major types of IR are alpha particles, beta particles, X-rays, and gamma rays. Since alpha and beta particles can be stopped by physical barriers, such as a sheet of paper or an aluminum plate, while X- and gamma rays are more penetrating, environmental exposure to gamma rays induces a greater degree of biological damage than exposure to alpha or beta particles. However, all four types of radiation are successfully utilized for therapeutic purposes and are capable of causing significant cellular damage (189).

The international unit of measure for absorbed radiation (radiation dose) is the gray (Gy), defined as J/kg of mass. Since equal doses of IR elicit differential effects depending on source and properties of the biological target, the unit of sievert (Sv) is used to express the equivalent dose. Individuals receive an average of 2.4 mSv per year of IR from natural sources, though this figure is increased in more developed nations (173). While natural sources of gamma rays (K-40) exist, gamma rays most widely used in research and therapies are from manmade sources such as Co-60 and Cs-137. The focus of this review will be primarily on the interactions of gamma rays with biological macromolecules, though references to other types of radiation are included where relevant.

Radiation from X-rays, including from CT scans, is a form of IR similar to gamma radiation but of lower energy. The energy level of X-rays enables visualization of dense areas (e.g., bones) in the human body, which cannot be penetrated as efficiently as soft tissue due to the photoelectric effect. CT scanning provides more thorough diagnostic details than X-ray but at the cost of patient exposure to a higher radiation dose, estimated at 15–30 mGy (compared with 0.01–0.15 mGy from X-ray imaging) (47). Cumulative CT-related low-dose IR has been correlated with detrimental biological effects such as DNA damage, bystander effects, tissue injury, and, in some cases, carcinogenesis (218). The biological consequences of low-dose IR have been recently reviewed (218), and illustrate that although CT is a valuable diagnostic tool, excessive exposure, especially in children, may lead to cancer and should be limited (241, 263). On the other hand, exposure to low-dose IR has been reported to increase immunity and induce an adaptive response, defined as a priming that enables an improved protective response to subsequent high-dose IR (261). Since the adaptive response is relevant to both the prevention and treatment of cancer, this intriguing phenomenon is under investigation by a number of groups, with an emphasis on identifying the underlying mechanisms of the adaptive response and the conditions that enable its onset. For example, the exposure of human colon carcinoma cells to a dose similar to that generated in image-guided radiotherapy (<100 mGy) followed by two doses of 2 Gy (24 h apart) resulted in an adaptive response which was further linked to the anti-apoptotic protein survivin (131). In vitro, the adaptive response has also been observed in primary human fibroblasts exposed to 100–500 mGy of X-ray 24 h before a 2 Gy dose (also X-ray) (92). This low-dose IR priming altered the response of the DNA repair protein phosphorylated histone H2AX (γH2AX) and increased secreted cytokine levels relative to non-IR-primed cells. Interestingly, these effects were not replicated by priming the cells with cytokines IL6 or transforming growth factor β (TGF-β) alone, thus ruling out a bystander effect through these signaling cues.

As this exciting area of research continues to develop and the mechanisms involved are further elucidated, new avenues will emerge for the manipulation of low-dose radiation in the prevention and treatment of cancer.

Effectors of IR

Gamma radiation of cellular water rapidly generates the reactive oxygen species (ROS) hydroxyl radical (^•OH) and ionized water (H₂O⁺), as well as the less investigated reductants hydrogen radical (H^•) and hydrated electrons (e_aq⁻). Within one ps (10⁻¹² s), superoxide (O₂^•−) and hydrogen peroxide (H₂O₂) are formed as secondary ROS products of IR (308). Subsequent chemical cascades affect the intracellular stoichiometry of these reactive species and generate additional cell-damaging molecules. For example, metal catalysis by intracellular ferrous and/or cuprous ions converts O₂^•− and H₂O₂ to form additional amounts of ^•OH (64). In a separate, but critically significant process, O₂^•− couples with endogenous nitric oxide (^•NO), forming peroxynitrite anion (ONOO⁻) (85). Cumulatively, these species, along with peroxynitrous acid (ONOOH), nitrogen dioxide (NO₂^•), dinitrogen trioxide (N₂O₃), and others, are referred to as reactive nitrogen species (RNS). The increased formation of RNS and the generation of additional ROS equivalents are particularly harmful to the cell, as the reaction products are in many cases more reactive with biomolecules than their precursors.

Cellular macromolecules are modified by direct ionization and via the reactivity of the high-energy species originating from water radiolysis (indirect effects of ionization), affecting an estimated 2000 primary ionization events (351). The timing attributes of cellular damage inflicted by IR range from chemical reactions occurring as rapidly as 0.01 ps after IR to major cellular effects that occur in the range of minutes to hours (308). Direct radiation damage is initiated in the range of 10⁻¹⁴–10⁻¹² s with the breaking of S–H, O–H, N–H, and C–H bonds. Widespread biomolecular damage induced by radiolytic products of water begins within 1 ps (10⁻¹² s), along with thiol depletion and further bond breaking (e.g., C–C and C–N). By 1 ms after IR exposure, the reactions of nascent ^•OH, H^•, and e_aq⁻ are mostly completed and DNA repair processes are initiated. Though the activity of some reactive IR products has diminished, an important event occurring through ∼10 s post-irradiation is the increased intracellular formation of ROS and RNS species through mechanisms described next (section “Endogenous propagation of IR-induced ROS”). The cumulative effects of the early, rapid biochemical processes are manifested in later stages of cellular damage, including the slowing of mitosis, damage to protein signaling networks, and membrane rupture, estimated to occur over the course of minutes to 10 h (308).

Endogenous propagation of IR-induced ROS

The overall amount of ROS generated from primary ionization events is further propagated via the intracellular activation of endogenous ROS-producing systems such as nicotinamide adenine dinucleotide phosphate, reduced form (NADPH) oxidase, and the mitochondrial electron transport chain (ETC) (12, 191, 235, 247, 351). IR exposure has been definitively linked to mitochondria-dependent ROS/RNS generation in tumor cells (95). Increased ROS generation in mitochondria after low-dose IR significantly contributed to radiosensitivity and cell survival (10). Whole body irradiation of rats resulted in the increased activity of cytochrome oxidase and NADH-cytochrome c reductase, decreased antioxidant activity, and increased lipid peroxidation in live mitochondrial fractions (170). Irradiation of A549 cells induced mitochondrial ROS production, increased mitochondrial membrane potential, and promoted respiration and ATP production (367). Similarly, an increased expression of NADPH oxidase was reported after irradiation with 10 Gy in rat brain microvascular endothelial cells, and the inhibition of NADPH oxidase led to a decrease in IR-generated ROS (79). IR-induced chromosomal instability in hematopoietic stem cells (6.5 Gy) was reversed by NADPH oxidase inhibition using diphenylene iodonium (262). The mechanisms of NADPH oxidase activation by IR may involve ceramide signaling, which is discussed later in this review. In addition to NADPH oxidase activation, a 2.5 Gy dosage of IR was shown to induce mitochondrial ROS production that can be blocked by inhibitors of mitochondrial respiration (66).

The temporal propagation of IR effects is also achieved through nitrosative stress mechanisms. A study of murine bone marrow stromal cells showed that irradiation with 2–50 Gy stimulated the expression of nitric oxide synthase (inducible nitric oxide synthase [iNOS]), leading to a dose-dependent increase in ^•NO levels in vitro along with the increased occurrence of nitrated tyrosine residues in vivo (128). Significant increases in the expression of iNOS and elevated levels of nitrate and nitrite have been associated with radiation-induced epithelial dysfunction in the colon (112). In addition to directly modifying tyrosines, cysteines, and hemes, ^•NO is the endogenous precursor to ONOO⁻ and other RNS (23). The activation of ROS- and RNS-producing pathways by IR is particularly important, as it illustrates a targeted localized increase in these reactive species as a consequence of global IR-induced ROS production, selectively altering cellular signaling and a host of metabolic pathways.

The complex chemical interplay of ROS/RNS generated directly by IR and through derivative systems such as NADPH oxidase, iNOS, and mitochondrial ETC is summarized in Figure 1. O₂^•−, H₂O₂, ^•OH, and ONOO⁻ are especially reactive, damage a wide range of cellular biomolecules, and react with each other to generate additional ROS/RNS. For example, the powerful oxidant peroxynitrite decays rapidly in acidic conditions (pK_a 6.8) and forms the highly potent secondary oxidant NO₂^• (22, 40). In general, the relatively milder oxidants (e.g., H₂O₂) target biomolecules in a more selective manner, while ROS with higher reactivity, such as ^•OH, are promiscuous; selected chemical examples appear throughout the next few sections.

FIG. 1.

IR generates the potent intracellular oxidants H₂O₂, O₂^•−, and ^•OH, along with reductants H^• and e_aq⁻. Endogenous ROS propagation occurs through the mitochondrial ETC and the increased expression of signaling enzymes such as NOX and iNOS. Reductive stress induced by IR leads to loss of sulfur in protein methionine and cysteine residues. ^•OH, hydroxyl radical; O₂^•−, superoxide; e_aq⁻, hydrated electron; H^•, hydrogen radical; H₂O₂, hydrogen peroxide; iNOS, inducible nitric oxide synthase; IR, ionizing radiation; NOX, NADPH oxidase; ROS, reactive oxygen species. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

Detection of intracellular ROS and RNS is routinely achieved using targeted probes, including fluorescent redox dyes such as dichlorodihydrofluorescein (DCF assay) and chemiluminescent methods (e.g., for ^•NO detection) (138, 352). Numerous caveats are associated with these methods, particularly with regard to cross-reactivity, and, thus, the development of more selective methods, such as the H₂O₂-specific peroxylfluor-1, is critically needed (56). Nevertheless, it has been reported that the amount of H₂O₂ generated by 2 Gy of IR is 10⁻⁷ M (351), and the extracellular H₂O₂ produced immediately after 5 Gy exposure has been measured at 12 μM (3). While seemingly low, these amounts are, at minimum, in the same range as physiological regulatory levels of H₂O₂, if not 10-fold greater [estimated physiological H₂O₂ concentration is 10⁻⁸–10⁻⁷ M (55)]. These concentrations have been shown to affect physiological processes such as proliferation, cell cycle arrest, senescence, and apoptosis (8, 87). We should also point out that a direct correlation between the steady-state levels of physiological or IR-induced ROS and the modification of biomolecules is not straightforward, and one should consider differences in the sensitivity and accuracy of detection methods, ROS compartmentalization, and evolving hypotheses for localized accumulation of ROS such as the “floodgate” hypothesis (360).

Cellular defense against IR-generated ROS

The capability of cells to survive an IR-based insult is dependent on a complex network of ROS-metabolizing molecules and detoxifying enzymes that comprise the cellular oxidative stress defense. O₂^•−, such as that generated through water radiolysis, is dismutated to H₂O₂ and O₂ by the activity of superoxide dismutases (SODs). Enzymatic detoxification of H₂O₂ is accomplished through the activity of catalase, peroxiredoxins (Prx), and glutathione peroxidases (GPxs). The roles of these antioxidant enzymes are paramount even under normal cellular conditions to keep ROS levels and pro-oxidant mechanisms in check. Aiding in this cause are low-molecular-weight (MW) endogenous antioxidants such as glutathione (GSH), ascorbate (Vitamin C), melatonin, lipoic acid, ubiquinone (Coenzyme Q₁₀), and Vitamin E, which target water radiolysis products, partially oxidized biomolecules, and peroxynitrite (153, 282, 313, 325, 332).

Almost immediately after cellular exposure to IR, the low MW antioxidant supply becomes compromised, leading, for example, to a rapid decrease in reduced GSH levels [(193), 10 Gy]. In an effort to combat the oxidative burst imparted by IR, cellular transcription factors, including nuclear factor (erythroid-derived 2)-like 2 and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB), are activated, resulting in the increased expression of ROS detoxifying enzymes, including catalase and SOD along with GPx, glutathione S-transferase (GST), heme oxygenase-1, and several others (224, 257). In addition, increased levels of many mammalian Prx isoforms have been noted after 10 Gy IR exposure, further enhancing the cellular defense mechanisms (4, 120, 193, 361). Both cumulative and acute IR readily disrupt the cellular redox balance, eventually overwhelming the cellular antioxidant system, an event marked by enzyme inactivation, a low GSH/glutathione disulfide ratio, and a decreased pool of low MW antioxidants. The consequence of such redox imbalance is manifested in the efficient modifications of nucleic acids, lipids, proteins, and other biomolecules (detailed in the next few sections and summarized in Table 1).

Table 1.

Oxidative Modifications of Biomolecules Associated with Exposure to Ionizing Radiation and Their Corresponding Methods of Detection

Modification	Detection method	References
Intracellular ROS	DCF assay	Hafer et al. (138)
Intracellular RNS	Chemiluminescence	Wardman (352)
DNA SSBs and DSBs	TUNEL, Comet, FISH, MS	Weimann et al. (354)
8-oxodG	ELISA, IHC	Rossner and Sram (289)
Lipid hydroperoxides (LOOH)	HPLC	Miyazawa et al. (233); Yamamoto (368)
	Phosphine (DPPP)	Okimoto et al. (252)
	Hydrazine with MS	Milic et al. (229, 230)
Hydroxynonenal (HNE)	HPLC/fluorescence	Tanaka et al. (326)
HNE, MDA, Acrolein	GC-MS	Kawai et al. (169)
Acrolein-DNA adducts	UHPLC-MS/MS	Yin et al. (374)
Protein radicals	EPR	Gordy and Miyagawa (129); Symons and Taiwo (320)
Protein carbonylation	DNPH antibody	Yan and Forster (369)
	DNPH with radiolabeling	Lenz et al. (199)
	DNPH with MS detection	Guo and Prokai (136); Bernevic et al. (30); Bollineni et al. (43)
	GPR with MS detection	Mirzaei and Regnier (232)
	Label free MS detection	Rauniyar et al. (276)
	Gel fluorescence	Madian and Regnier (219); Tamarit et al. (324)
	Biotin-conjugated probes	Chavez et al. (59); Chung et al. (70)
Methionine sulfoxide (MetO)	Infrared spectroscopy	Ravi et al. (277)
	MetO antibody	Nakaso et al. (246)
	Label-free MS detection	Guan et al. (135); Xiang et al. (363)
Reversibly oxidized Cys (SOH, SS, SNO, SN) and tailored for nitrosated Cys	Switch-tag (OxICAT)	Leichert et al. (197); Sethuraman et al. (301)
	NOxICAT	Lindemann and Leichert (207)
S-nitrosocysteine (CySNO)	Switch-tag	Lu et al. (213); Greco et al. (132); Camerini et al. (52)
	On-resin (SNO-RAC)	Forrester et al. (111)
	Phosphine probes	Wang and Xian (347); Bechtold et al. (21); Zhang et al. (381)
Cysteine sulfenic acid (CySOH)	Switch-tag	Saurin et al. (295)
	CySOH antibody	Seo and Carroll (300)
	1,3-Dicarbonyls: fluorescent	Poole et al. (267)
	1,3-Dicarbonyls: biotinylated	Charles et al. (57); Poole et al. (267); Klomsiri et al. (178); Nelson et al. (248); Qian et al. (271)
	1,3-Dicarbonyls: click chem.	Reddie et al. (278); Leonard et al. (200); Seo and Carroll (299); Qian et al. (272)

8-oxodG, 8-oxo-2′-deoxyguanosine; DCF, dichlorofluorescein; DNPH, 2,4-dinitrophenylhydrazine; DPPP, diphenyl-1-pyrenylphosphine; DSB, double-strand break; ELISA, enzyme-linked immunosorbent assay; EPR, electron paramagnetic resonance spectroscopy; FISH, fluorescence in situ hybridization; GC-MS, gas chromatography coupled to mass spectrometry; GPR, Girard's P reagent; HPLC, high-performance liquid chromatography; IHC, immunohistochemistry; MDA, malondialdehyde; MS, mass spectrometry; NOxICAT, isotope-coded affinity tag for detecting nitrosated and oxidized cysteine; OxICAT, isotope-coded affinity tag for detecting oxidized cysteine; RNS, reactive nitrogen species; ROS, reactive oxygen species; SNO-RAC, resin-assisted capture for S-nitrosothiols; SSB, single-strand break; TUNEL, terminal transferase; UHPLC-MS/MS, ultra high-performance LC coupled to tandem MS.

Chronic IR-mediated oxidative effects—evaluation based on in vivo evidence

Without doubt, the fundamental reaction chemistry underlying acute radiation damage involves rapid and widespread oxidation events that play critical functions in tumor cell killing. In addition, indirect contributions from inflammation (240), changes in vasculature (1), growth factor signaling (130), cytokine expression (385), mitochondrial dysfunction (380), and other cellular responses have also been demonstrated to influence the overall outcome of radiation treatment in vivo. Many of these processes have an underlying mechanism of oxidative regulation and/or action.

A remaining question is whether the late effects of radiation on normal tissue (e.g., chronic oxidative stress occurring over weeks to months after treatment) are regulated by oxidative mechanisms (81). For example, when using radiation nephropathy as an indicator of normal tissue damage [see ref. (78) for translational relevance and descriptions of mechanisms involved], there is evidence that both supports and refutes a function of late IR-induced oxidative injury. Supporting evidence includes the identification of chronic oxidative DNA damage (8-oxo-2′-deoxyguanosine [8-oxodG]) in glomerulus and tubules for approximately 24 weeks after IR treatment (288). On the other hand, antioxidant treatment in a rat model of radiation nephropathy using three different reagents (deferiprone, an iron chelator; genistein, an anti-inflammatory isoflavone; and apocynin, an NADPH oxidase inhibitor) did not significantly protect against kidney damage (73). Other studies using microarray analyses for gene expression have found either no or only limited evidence of up-regulation of known antioxidant proteins (72, 185).

In other tissues susceptible to late radiation effects (e.g., skin, lung, and brain), treatment with antioxidants, antioxidant enzymes, or enzyme mimetics has been shown to reduce the late effects of radiation (384). Clinical studies have shown that the administration of Lipsod, a lipid-encapsulated form of antioxidant enzyme SOD, led to a reduction in radiation-induced skin fibrosis (89). Similarly, others have shown that administration of the SOD mimetic EUK-207 along with genistein decreased the extent of radiation damage in normal lung but not skin tissue in a rat model of radiation (147).

A similar prevention or mitigation of radiation late effects was obtained with blockers of the renin–angiotensin system (RAS) and peroxisome proliferator-activated receptor gamma agonists (77). A redox mechanism for their mode of action was proposed for protection against loss of cognitive functions in radiation brain injury (287, 296). Angiotensin II (Ang II), an active peptide of the RAS, binds to angiotensin II type 1 receptor (AT1R) and generates ROS via NADPH oxidase. There is evidence of both increased expression of AT1R in a lung fibrosis model (255) and increased production of Ang II in rat lung at 2 and 6 months post-irradiation with 20 Gy (54). Preclinical studies indicate a role of Ang II blockers in ameliorating radiation-induced tissue damage. Angiotensin-converting-enzyme inhibitors (ACEI) and AT1R antagonists are effective in the treatment and prevention of radiation nephropathy (74–76, 238). Other studies reported that Ang II blockers reduced radiation-induced chronic injury, including lung fibrosis, vascular injury, and inflammation (119, 179, 225, 234). ACEI ramipril, given after stereotactical irradiation with 30 Gy using a single collimated beam, reduced the severity of optic nerve damage and retained nerve function (176, 290). In addition, chronic administration of AT1R antagonist, L-158,809, or ACEI ramipril prevented radiation-induced cognitive impairment in rats assessed at 6 months post-irradiation (196, 286). AT1R antagonist L158,809 blocked radiation-induced expression of heme oxygenase-1, a sensitive indicator of oxidative stress, suggesting a role of Ang II in radiation-mediated chronic oxidative stress (86). In vitro studies further confirmed a role for both Ang II and NADPH oxidase in the generation of ROS after radiation (79). Furthermore, blockade of Ang II with ACEIs or ATR1 antagonists reduced the expression of inflammatory proteins (79, 319) and prevented AP-1 and NF-κB activation (42, 384).

Cumulatively, these results demonstrate tissue- and antioxidant-dependent effects of late radiation injury and its mitigation as a potential explanation for the sometimes opposing reports in this area of research. In the subsequent sections of this review, we reference a number of studies that have identified markers of IR-induced damage (products of IR interaction with nucleic acids, metabolites, and proteins) shown to persist for hours to months after IR exposure. While in vivo evidence linking radiation exposure to both acute and chronic oxidation of biomolecules continues to emerge (11), there is a clear need for more investigations to determine the functional consequences of long-lived products of acute IR on tissue-specific radiation injury.

Interaction of IR and IR Effectors with Nucleic Acids

The interaction of IR with the cell nucleus has for many years been considered the primary mechanism responsible for the genotoxic effects of radiation fueling the central dogma of radiation biology. Among others, investigations by Munro in the early 1970s showed that a significantly higher dosage of radiation is needed to kill cells when the radiation is targeted selectively to cytosol compared with the nucleus (242). This view has evolved in more recent years, and while the interaction of IR with nucleic acids is certainly critical, studies are deciphering new IR-mediated pathways that lead to mutations, carcinogenesis, or cell death. Such mechanisms are discussed in the context of IR interaction with lipids (section “Interaction of IR and IR effectors with lipids”) and proteins (section “Interaction of IR and IR effectors with proteins”).

Chemistry

Nucleic acid damage in response to cellular IR is widely documented and is the classical paradigm for understanding the harmful effects of IR. Given the preponderance of combined therapies, it is important to note that the occurrence of DNA damage is increased when chemotherapeutics such as cisplatin are used in conjunction with radiation (284). DNA damaging events inflicted by IR alone include the deleterious alteration of bases and sugars, cross-link formation, single- and double-strand breaks (SSBs/DSBs), and DNA clustering (96, 331). Of the water radiolysis products, ^•OH is the most abundant and particularly destructive to nucleic acid molecules.

Radiation damage to deoxyriboses is the primary event underlying strand breakages, which occur in a high frequency and randomly along the DNA backbone in response to both direct ^•OH attack and the activity of nucleic acid-binding enzymes (46). DSBs, in particular, originate through the coordinated reactivity of two ^•OH radicals at nearby ribose sites, ultimately leading to strand breaks through subsequent radical pathways (14, 24, 350). Both the nucleobases and deoxyribose are targets for ^•OH -mediated damage. For purine nucleobases, ^•OH adds at C4, C5, and C8, generating reactive adduct radicals that lead to a variety of products, with the most common being the 8-hydroxypurines, specifically 8-oxodG, which serve as well-known hallmarks for oxidative DNA damage (Fig. 2) (94). The persistence of oxidative DNA damage in vivo has been illustrated by elevated levels of 8-oxodG in the mouse kidney in response to IR (20 Gy maximum) even at 24 weeks after treatment (288). Pyrimidine olefins are also susceptible to ^•OH addition, particularly at C5 and C6, generating pyrimidine glycols in the presence of O₂ (115).

FIG. 2.

Effects of IR on cellular DNA: oxidative damage to nucleobases and riboses mediated by ^•OH.

Radiation-induced nucleobase lesions include oxidatively modified bases as well as abasic sites, but do not immediately result in strand breakage (344). Both ^•OH and e_aq⁻ react with the nucleobases at diffusion-controlled rates, adding to unsaturated bonds and abstracting H^• from methyl and amino substituents (94). These radical products are structurally diverse and are involved in many secondary reactions as oxidants or reductants, depending on the structure and the reactive species in proximity. A common fate for such species is the diffusion-controlled reaction with O₂, producing peroxyl radicals, hydroperoxides, ring-opening events, and ring-contraction products. Specific kinetic and thermodynamic details of radiation-induced nucleobase damage, as well as the characterization of downstream oxidation products, have been compiled in a recent review (94).

Without unsaturations, deoxyriboses are modified via ^.OH-mediated hydrogen abstraction (Fig. 2). Though hydrogen abstraction has been noted at all ribose carbons, radical formation at C4′ dominates (227). Hydrogen abstraction at ribose carbons is the initiation event for reparable damage (e.g., of nucleobases) as well as irreparable DNA strand breaks. The extent of radiation-associated DNA damage is dramatically increased in the presence of bivalent metal ions (Cu²⁺, Fe²⁺) along with cellular reductants (GSH) via the Haber–Weiss generation of ^.OH and H₂O₂ (9, 270, 279). The capacities of hydrated (e_aq⁻) and prehydrated electrons to facilitate DNA strand breaks have garnered attention only recently but appear to play an important role in the cumulative effects of IR on DNA (44, 307, 346). O₂^•−, in contrast, has been shown to be significantly less reactive with DNA than with other water radiolysis products (266).

Biological consequences

The immediate response to IR-induced ROS/RNS-mediated DNA damage is the activation of the cell cycle checkpoint response, an intricately controlled network involving sensor, transducer, and effector proteins that respond to the DNA damage signal by initiating a cytoprotective response—the DNA damage response (DDR). The cellular mechanisms for DNA repair are extensive and have been recently reviewed in detail (331). A brief discussion is included here to highlight the interplay between the IR-induced DNA damage, IR-induced protein modifications, and chromatin remodeling that, ultimately, converges to repair nucleic acid damage or signal for the initiation of cell death pathways. Sensor and transducer proteins recognize DNA damage sites, initiate, and amplify a biochemical cascade. The principal proteins implicated as sensors and transducers of DNA damage include the Mre11–Rad50–Nbs1 complex, BRCA1/2 proteins, ATM/ATR, DNA-dependent protein kinase (DNA-PK), checkpoint kinases 1/2 (Chk1/2), and poly(ADP-ribose) polymerase (143, 265, 297, 342, 364, 365). The activities of these proteins are tightly regulated by multiple post-translational modifications (PTMs). For example, ATM, one of the initial enzymes activated in DDR and a key regulator of all three cell cycle checkpoints, is activated by IR in a mechanism involving phosphorylation at Ser1981 (13). ATM is also directly activated by IR-induced ROS, facilitating dimerization through the formation of an intermolecular disulfide (93, 137). In addition to modifying other crucial cell cycle checkpoint proteins (e.g., Chk1/2 and p53), ATM mediates the phosphorylation of Kap1, promoting heterochromatin relaxation and increasing the efficiency of DNA repair (127, 250). The ATM polymorphism was also found to play a function in determining the sensitivity to radiation, as detailed in the section “Genomics and epigenomics.” Another observation pointing to the significance of oxidative PTMs to the DDR comes from studies using freshwater invertebrates, Adineta vaga and Caenorhabditis elegans, revealing that diminished carbonylation maintains the activity of DDR enzymes, enhancing post-IR survival in A. vaga compared with the genomically similar C. elegans (183, 184). Protein carbonylation, an oxidative modification of many amino acids, is discussed in the section “Carbonylation.” Thus, cellular DNA damage after IR exposure is largely dependent on the capacity of the cellular proteins to prevent and repair the DNA modification induced by IR.

Methods of detection

Historically, many techniques and methods have been used to detect the DNA damage. Direct methods include those designed to detect specific SSBs and DSBs such as the well-known terminal transferase, comet, and fluorescence in situ hybridization assays, and those targeted for the analysis of oxidized nucleic acid metabolites [e.g., liquid or gas chromatography (GC) coupled with electrochemical or mass spectrometry (MS) detection] (354). Other methods rely on enzyme-linked immunosorbent assay and immunohistochemistry as commonly applied for detection of 8-oxodG, one of the most abundant oxidized nucleosides (289). Indirect methods for the detection of DNA damage include Western blotting or imaging analysis of proteins involved in the DDR that undergo PTMs after IR-induced DNA damage. Examples include the detection of γH2AX histone that undergoes phosphorylation and binds to sites of DNA DSBs (159, 291). Only the most frequently used methods of detection were highlighted here; for a more in-depth discussion, including the pros and cons of various techniques, the reader is directed to a recent book on this topic (91).

Interaction of IR and IR Effectors with Lipids

Another biomolecule target of radiation-generated ROS is the lipid layer within cell membranes. The lipid component of cell membranes is generally estimated to be ∼5 nm in thickness with significant exposure to the aqueous cellular environment (315). Though radiation is capable of directly damaging lipids, lipid bilayer mimetics have indicated that indirect damage induced by water radiolysis products is a larger contributor toward overall lipid modification by IR (25). Radiation induces lipid peroxidation, particularly the peroxidation of polyunsaturated fatty acids (PUFAs), leading to an increase in membrane permeability, disruption of ion gradients and other transmembrane processes, and altered activity of membrane-associated proteins (80, 359). Studies of IR targeted to the cell membrane revealed the induction of apoptosis at 5–10 Gy via increased ceramide levels (139). This outcome was observed even in cells without a nucleus, revealing another pathway for cellular IR damage independent of nucleic acid damage.

Chemistry

Unsaturated fatty acids are unreactive toward molecular oxygen, but readily oxidized in a number of radical-mediated processes (269). Peroxidation of PUFA, such as linoleic and arachidonic acids, leads to the generation of diene hydroperoxides, isomerization of cis-alkenes, and/or degradation to form small molecule reactive carbonyls such as malondialdehyde (MDA), acrolein, and 4-hydroxy-2-nonenal (HNE) (113). Alkene isomerization is initiated by the abstraction of hydrogen from a PUFA bisallylic site by ^•OH, thiyl radical (RS^•), or other radicals, generating a carbon-centered PUFA radical. Subsequent capture by oxygen results in a cis-trans isomerization and forms a peroxyl radical in a diffusion-controlled reaction (Fig. 3A, top path) (269, 323). Alternatively, cis-trans isomerization may also occur via the direct addition of RS^• to a PUFA alkene (Fig. 3A, bottom path). Ejection of the R′′S^• radical accompanies isomerization, a process documented even in the presence of oxygen (228). Peroxyl radicals are terminated by the formation of lipid hydroperoxides (LOOH) and PUFA fragmentation to MDA, HNE, and acrolein (Fig. 3B). The reactive aldehyde products of PUFA peroxidation persist intracellularly for approximately 2 min, providing ample time for the modification of proteins, nucleic acids, and other biomolecules (69, 154, 264). In vivo evidence for increased levels of reactive aldehyde MDA in response to IR has been shown in the kidney, lung, and liver of rats exposed to 8 Gy of total body IR (298).

FIG. 3.

Effects of IR on cellular lipids. (A) Oxidative cis-trans isomerization of polyunsaturated fatty acids. (B) Reactive aldehydes generated by lipid peroxidation. (C) Chemical probes for the detection of intracellular ROS (DPPP) and reactive aldehydes (DNPH; CHH). CHH, 7-(diethylamino)coumarin-3-carbohydrazide; DNPH, 2,4-dinitrophenylhydrazine; DPPP, diphenyl-1-pyrenylphosphine.

Biological consequences

A major factor affecting the fate of LOOH and the magnitude of its effects in cells is the extent of lipid peroxidation, correlated directly to cellular redox status. Moderate levels of LOOH activate the oxidative stress response, which, when exceeded, leads to apoptosis; high levels, and widespread LOOH, resulting in global damage to the cell membrane and intracellular content, triggering membrane lysis and necrosis (123). Chemically, the degradation of LOOH proceeds through one-electron and/or two-electron reduction pathways. One-electron reduction of LOOH occurs via aerobic reduction mediated by Fe²⁺ ions and providing epoxyallylic peroxy radicals (OLOO^•) with subsequent radical reactions leading to additional lipid damage. The detoxification of LOOH by two-electron reduction is carried out by selenoperoxidase enzymes such as the GPxs, thioredoxin reductase (TrxR), and phospholipid hydroperoxide glutathione peroxidases (PHGPxs), as well as the seleno-independent GST α (339, 356). These enzymes reduce LOOH to LOH and H₂O in the presence of two equivalents of GSH with the exception of TrxR, which instead utilizes NADPH as an electron donor (38). The LOOH-reducing enzymes use similar mechanisms apart from their cofactor differences and vary mainly in terms of size and substrate specificity. To illustrate, though both the GPxs and PHGPx detoxify H₂O₂, GPx enzymes target polar LOOH such as fatty acid hydroperoxides; whereas PHGPx reduces phospholipid hydroperoxides, cholesterol hydroperoxides, and other hydroperoxides of a lower polarity (330, 339, 340). Apolipoprotein A-I (apoAI) and apolipoprotein A-II (apoAII) have been noted to reduce cholesterol ester hydroperoxides via oxidation of critical Met residues to the corresponding sulfoxides (121). In total, the enzymatic defenses against LOOH cytotoxicity rely heavily on reductant bioavailability (GSH, NADPH) coupled with reduced active site cysteine, selenocysteine, and methionine residues of the repair enzymes. The intense disruption of cellular redox metabolism invoked by IR is sufficient to overwhelm LOOH defense mechanisms in IR-sensitive cells and tissues due to the oxidation and inactivation of detoxifying enzymes, though the timing of such modifications and the lipids most susceptible to radiation damage are not well understood. LOOH are known to persist after IR, as demonstrated in vivo in the mouse hippocampus at 2 weeks after IR exposure (10 Gy) (205).

Sphingolipid metabolism is a key pathway that is altered in response to IR. The sphingolipid ceramide, a product of sphingomyelin hydrolysis catalyzed by acid sphingomyelinase (ASMase) and neutral sphingomyelinase, has, in particular, been closely connected to the cellular IR damage. Cellular exposure to IR leads to ASMase relocalization from the lysosomes to the plasma membrane, where sphingomyelin is hydrolyzed to generate large amounts of ceramide (80). Both mutations and PTMs at Cys629 (and possibly Ser508) are known to regulate the activity and localization of ASMase at the plasma membrane, highlighting the complex interactions among IR-induced protein modifications, lipid raft microdomain rearrangement, and DNA damage repair (273, 378). In addition, DNA damage events activate ceramide synthase, the de novo source of ceramide, contributing to significantly increased levels of intracellular ceramide in response to IR (343). This large in situ generation of ceramide from sphingomyelin within membrane lipid rafts alters the membrane properties, largely because ceramide-containing lipid rafts coalesce and form large, ceramide-enriched membrane platforms. These lipid platforms not only contain membrane receptors and proteins but are also enriched in nuclear enzymes such as DNA-PK that are relocalized on irradiation. In head and neck cancer, in particular, the dynamics of lipid raft microdomains and associated signaling was shown to underlie the response to IR and targeted therapies against epidermal growth factor receptor (EGFR) (33, 101, 157). Membrane rafts have also been associated with the bystander effects of low-dose IR, where cytoplasm-targeted IR of one cell led to an increased micronuclei yield of 36–78% in surrounding cells (302). This outcome was independent of the nucleus and instead was linked to ^•NO signaling and membrane raft formation. Lipid rafts are also implicated in IR propagation via NADPH oxidase reconstitution within lipid platforms, where it functions as an additional ROS source as discussed in the section “Endogenous propagation of IR-induced ROS” (379).

Methods of detection

Many biochemical approaches have been developed to detect lipid peroxidation. Traditional techniques involve the chemiluminescence detection of LOOH that are first separated by high-performance liquid chromatography (HPLC) and then treated with isoluminol in the presence of a metal ion catalyst (e.g., heme or cytochrome c) with the detection of isoluminol oxidation induced by LOOH (233, 368). A newer approach involves the use of diphenyl-1-pyrenylphosphine (DPPP), which reacts with numerous biological oxidants and forms the highly fluorescent phosphine oxide (Fig. 3C). DPPP localizes to cell membranes, where it is oxidized by LOOH at rates much faster than oxidation by H₂O₂ or tert-butyl hydroperoxide, likely due to its hydrophobic compatibility with the lipid bilayer (252). This is an especially critical development for the fluorescence imaging of lipids in live cells, as the widely used DCF diacetate (DCF assay reagent) is prohibitively hydrophilic for lipid imaging (252).

Lipid peroxidation events have also been uncovered via the detection of the volatile, lower MW aldehyde end-products. Levels of the lipid-derived aldehyde HNE have been measured using HPLC with pre-column labeling and fluorescence detection (326). Along with HNE, MDA and acrolein have been directly quantified using GC coupled to MS; softer ionization of these species using matrix-assisted laser desorption/ionization (MALDI) and electrospray ionization (ESI) is inefficient and often requires derivatization (97, 169). 2,4-Dinitrophenylhydrazine (DNPH) readily modifies oxidized phospholipids for MS-based detection in a mechanism similar to that shown in Figure 4, but is not as successful at targeting lower MW peroxidation products in complex samples (Fig. 3C) (229). Recently described derivatization reagent 7-(diethylamino)coumarin-3-carbohydrazide (CHH) provides a means of detecting both high and low MW lipid peroxide products in the same sample (230). Furthermore, the CHH amine improves ionization (by ESI) and offers a unique footprint of reporter ions on MS² fragmentation by collision-induced dissociation (CID) (230). The biomolecules targeted by these reactive aldehydes, with increased mass and hydrophilicity, are more amenable to analytical separations by liquid chromatography (LC): Acrolein adducts of DNA in human leukocytes have been quantified by ultra high performance LC coupled to tandem MS (374). Ceramide-enriched lipid rafts are detected by imaging in live cells using fluorescent conjugates of cholera toxin subunit B (34). The detection of protein adducts of lipid-derived aldehydes is discussed in the next few sections.

FIG. 4.

Detection of reactive aldehydes and carbonylated proteins. Derivatization of protein carbonyls using DNPH and mechanistically similar Girard's P reagent and a biotinylated hydroxylamine.

Interaction of IR and IR Effectors with Proteins

In addition to nucleic acids and lipids, proteins are highly abundant cellular biomolecules and prominently targeted by IR, inducing changes in their expression and activity as well as oxidative or reductive PTMs. Single cell exposure to microbeam-directed alpha particles in the cytoplasm was shown to be mutagenic with severity of outcome dependent on the intracellular GSH and radicals (362). Importantly, these mutations were shown to occur without much, if any, cell death, illustrating the potential for transforming cell signaling along with cell persistence. In vivo evidence for the oxidative modification of proteins in response to IR has been observed in the mouse brain and rat liver as judged by the increase in protein carbonylation (99, 222).

PTMs critically affect protein structure and function though conformational changes, modulation of enzymatic activity, degradation, and cellular trafficking (268, 275, 295, 349, 376). The paradigm for understanding cellular IR damage in the past several decades has evolved to include the elucidation of protein targets in addition to the classically investigated DNA damage events. Fueling this restructuring are investigations of cytoplasm-directed IR and a wide array of studies that revealed alterations in the cellular proteome at IR doses <10 Gy, once believed to be the threshold for affecting proteins (124, 190). For example, cellular exposure to IR targeted to the cytoplasm led to an increase in levels of RNS, LOOH, cyclooxygenase 2 expression, and activation of extracellular signaling-related kinase signaling, along with increased oxidative DNA damage (151). The modulation of protein expression and protein–protein interactions has been noted in response to doses as low as 1 Gy [see section “Endogenous propagation of IR-induced ROS” and ref. (204)]. The inactivation of redox-sensitive enzymes such as catalase, SOD, GPx, and the protein tyrosine phosphatases has been observed at 2–8 Gy (17, 25). Moreover, IR doses <6 Gy have been shown to disrupt the extent of protein PTMs, including, but certainly not limited to, S-nitrosation and carbonylation (192, 223, 386). Cumulatively, these results demonstrate that the impact of IR on the proteome was long underestimated but is now accepted to profoundly affect cellular processes. The observation that protein signaling is altered at <1 Gy of IR illustrates that the radiosensitivity of proteins is akin to other biomolecules and underscores the need for investigating protein-based IR damage (371). In this chapter, the focus will be on oxidative and reductive PTMs and IR-induced changes in protein expression.

Protein oxidation

More than 35 types of oxidative protein modifications are known currently, including direct amino-acid oxidation (e.g., for Cys and Met), oxidative cleavage of the protein backbone and/or amino acid side chains, carbonylation, and the addition of lipid oxidation products discussed earlier (219). In addition to IR therapy, protein oxidation has been implicated in the progression of many disease states, including diabetes (372), inflammation (312), sepsis (209), Alzheimer's disease (145), multiple sclerosis (37), Parkinson's disease (90), and many cancers [see (283) and references within].

Radical cleavage of the protein backbone and modification of amino-acid side chains

^•OH, generated via IR and other processes, initiates cleavage of the protein backbone and reacts with each of the twenty standard amino acids and selenocysteine, with associated rate constants for all such reactions in the 10⁷–10¹⁰ M⁻¹ s⁻¹ range (pH 7), reaching diffusion-limited rates for cysteine, methionine, and aromatic amino acids (41, 88). Radicals, including ^•OH and others, are believed to preferentially react with the protein amide backbone over the amino-acid side chains, readily abstracting hydrogen atoms and forming α-carbon-centered radicals. Similar to DNA-based radicals, protein carbon radicals subsequently react in a diffusion-controlled manner with O₂ and may be quenched by other radical species when present in sufficient concentrations. Peroxyl radicals, formed via a reaction with O₂, facilitate the formation of additional oxidation products through radical mechanisms, leading to fragmentation of the protein backbone.

Similarly, ^•OH abstracts hydrogen atoms from aliphatic amino-acid side chains at all carbons, demonstrating that its high degree of reactivity precludes chemoselectivity. Except under anoxic conditions, aliphatic amino-acid radicals are rapidly oxygenated and generate peroxyl radicals or are repaired by cysteine thiols, leading to thiyl radicals, though kinetically this reaction is not believed to compete with oxygenation (31). Aromatic amino acids are significantly more reactive with the dominant reaction pathway being ^•OH addition to the aromatic ring. In the case of Tyr, ^•OH addition and subsequent hydrogen abstraction leads to phenoxyl radicals, which in the absence of reductants, form Tyr dimers that are implicated in the formation of intra- and inter-protein linkages (107). Nitrotyrosine products are formed in response to ROS/RNS (e.g., peroxynitrite) and react with e_aq⁻ several orders of magnitude faster than tyrosine (k=3.0×10¹⁰ M⁻¹ s⁻¹ at pH 7.0 for 3-nitroTyr vs. 2.8×10⁸ M⁻¹ s⁻¹ at pH 6.6 for Tyr) (106, 304). Increased levels of nitrotyrosine have been observed in the mouse hippocampus within 2 h of IR exposure (8 Gy) relative to the control (116). Tryptophan is another target for RNS-initiated attacks to form regioisomeric nitro and hydroxyl derivatives (158, 366).

Methods of detection. Protein backbone fragmentation is classically monitored by the detection of radical products using electron paramagnetic resonance spectroscopy (129, 320).

Carbonylation

Radiation-generated ROS are noted to modify cellular proteins by carbonylation, the post-translational addition of carbonyl moieties to amino-acid side chains, particularly occurring via metal catalysis on Lys, Thr, Pro, Glu, Asp, and Arg residues (220). Along with Lys, Cys and His also undergo carbonylation by reacting with oxidized lipid products such as HNE (337). Sites of carbonylation are formed by the Michael reactivity of Cys, His, and Lys with reactive sugar and other lipid oxidation products. The carbonylation event produces a variety of acyl-modified amino acids and is capable of inducing a lesion in the protein backbone; see (219) for chemical structures. Carbonylation is widely considered an irreversible process and appears to target specific amino acids within a protein rather than many residues indiscriminately; recent evidence suggests that the sites susceptible to carbonylation within a given protein appear to be modulated by the identity of the oxidant as well as structural aspects of the protein (327). Repair processes for proteins modified by carbonylation have not been identified, and, instead, degradation is believed to be the primary mechanism for cellular regulation of proteome carbonylation (201). Increased levels of protein carbonyls have been noted after IR exposure (10 Gy) (310, 317), and in addition, have been implicated in the pathophysiology of many human disease states, including chronic lung disease (357), neurodegenerative diseases (133), diabetes (2), and ischemia-reperfusion injury [(84) and references therein].

Methods of detection. Protein carbonyls, though formed irreversibly, are unstable and readily captured in the Schiff base form by lysine residues, even in frozen samples, making their detection challenging after long-term storage of biological samples (219). Detection of nascent or more stable carbonyls is, however, possible by derivatizing with DNPH (also called Brady's reagent). Nucleophilic carbonyl capture affords hydrazone products (Fig. 4) that are readily detected spectrophotometrically (360 nm) or via immunoblotting using anti-DNP antibodies (369). DNP derivatization is rendered irreversible on a mild reduction of the resulting Schiff base hydrazone product by borohydride reductants; the use of NaB³H₄ provides a means for radioactive quantitation of carbonylation (199). Another method to identify and quantify cysteine carbonylation products employs the Raney nickel-catalyzed thioether reduction (338).

Newer techniques for uncovering protein carbonylation have been developed with a focus on compatibility with MS. Sites of His and Cys carbonylation on DNPH-modified peptides have been detected by MS using MS² with CID for precursor ion fragmentation, though the issue of neutral loss during the acquisition of MS² spectra complicates the detection of sites carbonylated by HNE (136). A more often utilized approach couples gel electrophoresis and MS, where protein carbonylation is first detected using a DNP-targeted antibody followed by the digestion of selected proteins and MS analysis of their resulting peptides (30). Hydrazine-based detection has been extended using carbonyl-labeling compounds that are appended to biotin for affinity capture or fluorophores such as boron-dipyrromethene, Cy3, and Cy5 for in-gel fluorescence detection (219, 324).

The detection of DNP-derivatized peptide carbonyls has also been achieved by MS using negative ESI with MS² acquired with pulsed-Q dissociation fragmentation (43). One such carbonyl labeling compound, Girard's P reagent (GPR), utilizes analogous hydrazine chemistry but is equipped with a pyridinium moiety amenable to enrichment using strong cation exchange (SCX) chromatography and is beneficial for peptide ionization (231). Isotope-coded GPR has been used to identify 41 carbonylated peptides in H₂O₂-treated yeast using SCX followed by MALDI-time-of-flight (TOF)-TOF analysis (232). Alternatively, label-free approaches have been used for the analysis of HNE-modified peptides, using neutral loss-triggered MS³ to pinpoint carbonylated residues (276).

Similar to other PTMs, the detection of carbonylation events is challenged by their low abundance, owing to the relatively low intracellular concentrations of many signaling proteins and the fact that only a small number of amino acids per protein are targeted for oxidation. A complicating factor in distinguishing the two primary types of protein oxidation events, carbonylation and cysteine oxidation, is the recently noted cross-reactivity of DNPH and hydrazine probes with cysteine sulfenic acid (83, 370). Biotin-conjugated hydroxylamines capture protein carbonyl Michael adducts (e.g., those formed via amino-acid modification by HNE) and afford stable oxime adducts at physiological pH (59, 70). Their selectivity for carbonyls over a direct reaction with cysteine sulfenic acid has not been investigated, though such products should be readily distinguished using high-resolution MS techniques. Care should also be taken to avoid nucleic acid contaminants, which generate false positives in DNPH-based assays (215).

Methionine oxidation

A wide array of one-electron and two-electron oxidants target methionine, including the water radiolysis products. The products of methionine reaction with ^•OH, H^•, e_aq⁻, and H₂O₂ have been reported in a recent study using aqueous solutions of free methionine (10 mM) that are exposed to a prolonged IR dose (10 h at 6.5 Gy min⁻¹) (15). Characterization of the Met-based products revealed that in the presence of molecular oxygen, (i) the addition of ^•OH to the sulfur atom of Met did not afford methionine sulfoxide (MetO) but exclusively 3-(methylthio)propanal; (ii) H₂O₂, derived from water radiolysis or O₂^•− disproportionation, oxidized Met to MetO (Fig. 5A) (15). Reactions of Met with IR-associated reductants H^• and e_aq⁻ are discussed in the section “Protein modification by reducing radical species.”

FIG. 5.

Radiation-induced sulfur oxidation in proteins, affecting (A) methionine and (B) cysteine residues. Oxoforms in gray are readily reduced to thiol by DTT and other reductants in vitro.

Other IR-relevant species such as ONOO⁻ and other secondary IR oxidants readily oxidize free methionine and protein methionines to sulfoxide products (Fig. 5A). MetO is significantly more polar than unoxidized Met, and its high degree of thermodynamic stability precludes its reduction by the majority of non-enzymatic intracellular reductants (GSH, ascorbate, and protein thiols) and enables its direct detection. Hyperoxidation of MetO generates methionine sulfone, an irreversible oxidation product identified on irradiation of free methionine (305). Though the fundamental chemical details regarding Met oxidation products are well established, large-scale investigations of IR-induced MetO formation in proteins and the resulting intracellular consequences are still ongoing.

Specific to IR exposure, MetO formation has been noted in vitro only for cytochrome c (329), hemoglobin (363), and bovine α-crystallin (110). MetO has also been observed on ROS treatment of latent TGF-β, an established target of IR (98, 162), providing indirect evidence for the potential formation of MetO in this protein on treatment with IR. Numerous examples have shown, however, that MetO formation occurs in proteins under a range of conditions (non-IR or indirectly linked to IR) with site specificity, suggesting a role of Met oxidation as a regulated mechanism for redox protection and modulation of protein function, though the molecular factors underpinning selectivity are not well established (214, 314). For example, sulfoxide formation at 2 of 9 calmodulin methionine residues was shown to disrupt calmodulin-mediated activation of plasma membrane Ca-ATPase (18, 32). Other examples include MetO formation in the protein Parkinson disease protein 7 (DJ-1), an event that is closely associated with the development of neurodegenerative diseases (65), and in apoAI and apoAII enzymes, in which the reduction of high-density lipoprotein-associated LOOH to LOH by these enzymes was shown to occur with the concomitant formation of MetO at Met112 and Met148 (human apoAI) and Met26 (human apoAII) (121). In total, methionine oxidation selectively modulates numerous cell signaling pathways that are regulated by calcium, phosphorylation, and others (71, 142, 211).

The repair of MetO is catalyzed by the methionine sulfoxide reductases (Msr), which restore reduced methionine from MetO. Chemically, oxidation of the Met thioethers generates the chiral sulfoxide in a racemic mixture of R and S isomers (k=0.006 M⁻¹ s⁻¹ for free Met+H₂O₂) (285). MsrA selectively reduces the S epimer of MetO [Met(S)O] and MsrB reduces only the R epimer [Met(R)O]. MsrA resides in the mitochondria, cytosol, and nucleus; while three forms of MsrB are known: MsrB1 (cytosol; nucleus), MsrB2 (mitochondria), and MsrB3 (mitochondria; endoplasmic reticulum) (167, 175, 237). MsrA, MsrB2, and MsrB3 have cysteine-based active sites, while MsrB1 contains an active site selenocysteine (Sec) nucleophile, though exceptions have been noted (150). Stereospecific MetO reduction is characterized by oxygen-atom transfer from MetO to Msr, forming an active site with sulfenic acid or selenenic acid (187). Capture by a resolving cysteine with subsequent disulfide (or selenosulfide) reduction by the thioredoxin/TrxR system restores active Msr, reducing MetO at the expense of NADPH (50, 253). Thus, the Msr enzymes help defend cells and tissues from oxidative damage. For example, the expression of MsrA is up-regulated in the epidermis that is exposed to UVA radiation and H₂O₂, suggesting MetO reduction to be a crucial component of the cellular oxidative stress response (251). To date, the impact of IR on Msr expression and activity has not been elucidated.

Methods of detection. Detection of MetO is accomplished with relatively inexpensive techniques such as Fourier transform infrared spectroscopy with monitoring for vibrational bands at 1044 and 1113 cm⁻¹ or biochemically using a commercial MetO-targeted antibody (246, 277). MS is another major avenue for protein MetO detection, helping identify MetO formation in keratins (195), prion proteins (306), and to correlate MetO with glaucoma (328) and sepsis (122). Both ESI and MALDI were utilized to identify MetO formation in hemoglobin of IR-treated primary human erythrocytes (363). Tandem MS methods are the most widely used for the identification of MetO sites using CID-mediated neutral loss of methanesulfenic acid (loss of 80 Da) or milder precursor ion fragmentation methods such as electron capture or transfer dissociation (135). An essential consideration for all approaches to MetO detection is the possibility for adventitious Met oxidation during sample handling and processing (68, 135). Non-physiological routes to MetO by means of ESI and gel electrophoresis have been uncovered and optimized protocols have been suggested, though many additional avenues likely exist (62, 236, 318, 377).

Cysteine oxidation

Among the amino acids, protein cysteine residues have the lowest redox potential, rendering their thiol groups particularly sensitive to oxidative modification by ROS/RNS. Protein thiols have diverse chemical responses to ROS/RNS, and in a very general sense, cysteine residues with a lower thiol pK_a tend to be more readily modified by oxidants and electrophiles and are often termed “redox active” or “redox sensitive” cysteines. However, a range of analyses have uncovered the varying reactivity of cellular thiols with electrophiles, including the endogenous electrophile HNE, and are discussed in greater detail in a recent review (117). It is now well established that decreased thiol pK_a does not always correlate with increased reactivity toward H₂O₂, particularly among redox-regulated enzymes [see ref. (358) for a thorough analysis]. In addition to the cysteine thiol pK_a, the differential reactivity of thiols toward ROS/RNS and the products of oxidation are affected by factors, including residue accessibility, polarity and basicity of adjacent amino acids, and identity and concentration of the ROS/RNS source (244, 245, 293). The chemical interplay between thiols and ROS/RNS is further governed by protein subcellular localization, organelle redox potential, and proximity to ROS/RNS sources.

Mechanistically, redox-sensitive cysteine thiols undergo a two-electron oxidation in the presence of H₂O₂, hydroperoxides (ROOH), peroxynitrite, and other oxidants, forming sulfenic acid (CySOH) as the initial cysteine oxidation product (Fig. 5B). The subsequent capture of CySOH by protein and low MW thiols (e.g., GSH) generates disulfides, which are readily reduced to regenerate the thiols in the presence of cellular reductants or by the activity of disulfide reductases. In addition to its role as a thiol-disulfide intermediate, CySOH may be captured by adjacent lysine residues or an amide nitrogen in the protein backbone and form sulfenamides [also called sulfenyl amides (−SN)] as in the well-defined example of Cys-based protein tyrosine phosphatase 1B (PTP1B) (161, 292). Without sufficient thiol content proximal to the CySOH site, further oxidation results in the formation of sulfinic (CySO₂H) and sulfonic (CySO₃H) acids, an irreversible process except for the special case of Prx sulfinic acid repair by sulfiredoxin (36, 165). Cysteine thiols react with NO-generated RNS, form S-nitrosothiols, and also undergo transnitrosation, in which the nitroso moiety is passed among cysteine residues in a site-selective manner that is dependent on the local environment and the pK_as of thiols involved (49, 149, 309). S-Nitrosation of cysteine occurs through three major pathways: (i) thiol capture of nitrosonium ion (via acidic nitrite) (144); (ii) ^•NO capture by cysteine thiyl radical; and (iii) directly through N₂O₃ modification (126). The transition metal-mediated formation of nitrosonium ion from ^•NO also facilitates S-nitrosation (172). For additional chemical details, including mechanisms of cysteine oxidation and nitrosation, the reader is referred to a recent review on this topic (259).

Methods of detection. Despite only a few examples of physiological cysteine sulfinic acid formation, this modification attracts attention, as it indicates a sharply increased intracellular oxidant status and profoundly impacts redox signaling (39, 114, 243, 274). As a largely irreversible cysteine product, CySO₂H may be directly detected by MS, though high-resolution detection is required to distinguish this modification from others with a similar mass (e.g., sulfinamide −SONH₂; persulfide −SSH). A recent report details the capture of aryl nitroso electrophiles by small-molecule sulfinic acids, leading to stable sulfonamide conjugates, though this chemistry has not yet been applied to protein CySO₂H detection (210).

Among the cysteine modifications by ROS/RNS, cysteine nitrosation and sulfenylation are of special interest due to their reversibility and emerging roles in regulating protein function and communication (146, 160, 260, 268, 349). For example, S-nitrosothiol formation on mitogen-activated protein kinase phosphatase-1 is shown to increase phosphatase activity and reduce radiosensitivity in head and neck cancer cells (134). Further evidence linking redox signaling and phosphorylation pathways has emerged with the observations that cysteine oxidation to sulfenic acid (Cys124) inactivates Akt2 kinase (349), while sulfenic acid formation at redox-sensitive Cys797 of EGFR stimulates its kinase activity (260). With the rapid advancements in ESI- and MALDI-based MS methods and newly designed strategies for sub-proteome identification, a host of proteins are now widely accepted ROS/RNS targets (45, 67, 216, 219, 231, 258, 303, 311), though these methods have not yet been applied to investigate the effects of IR. The methods pertaining to the high-throughput identification of the CySOH and CySNO proteins and their cysteine sites of modification are discussed in the section “Redox and phosphoproteomics.” In a research article included in this Forum, we demonstrate the successful application of biotin-tagged probes to monitor the cellular redox status in tumors with various degrees of response to radiation treatment.

Protein modification by reducing radical species

In addition to IR-generated oxidants, the radiolysis of water is a major source of the competent intracellular reductants e_aq⁻ and H^•. Though reductive stress has been exceedingly understudied in comparison to its oxidative counterpart, the disruption of cellular redox balance toward reduction has been linked to several pathologies, most notably of the cardiovascular system (48). The reductants e_aq⁻ and H^• readily target the sulfur-containing amino acids, reacting at or near diffusion-controlled rates (51). Methionine serves as a principal target of H^•, adding sulfur to generate a sulfuranyl radical (Fig. 6) (15). Subsequent desulfurization forms a carbon-centered radical that may be terminated by a hydrogen atom donor and form α-aminobutyric acid (Aba) or by molecular oxygen to generate homoserine (58). The exposure of methionine-containing peptides to e_aq⁻ results in peptide cleavage through a deamination mechanism (316).

FIG. 6.

Impact of IR-induced reductive stress on protein sulfur, including methionine, cysteine, and cystine.

The reaction of e_aq⁻ with reduced cysteine generates hydrosulfide ion (HS⁻) and the corresponding alkyl radical (R^•); a reaction with cysteine/protein disulfides (RSSR) provides the thiyl radical (RS^•) and the thiolate (RS⁻) ion (Fig. 6). In contrast, the reactivity of H^• is marked by radical abstractions: A reaction with reduced CySH occurs via hydrogen abstraction to generate RS^• or through homolytic cleavage to form R^• and hydrogen sulfide (H₂S). Cystine reactivity with H^• induces homolytic disulfide cleavage, forming either RS^•+RSH or RSS^•+RH, depending on the site of cleavage. Thiyl radicals are also generated by the abstraction of thiol hydrogen by ^•OH, illustrating how this species, among others, is common to both reductive and oxidative processes (226).

Biochemically, thiyl radicals are gateways to amino-acid conversion and the formation of additional redox-critical species. Thiyl radicals are efficiently captured by thiolate ions (R′S^-), yielding a disulfide radical anion (R′SSR^•−), which readily reduces O₂ to O₂^•− (Eq. 1) (118, 353).

Though short lived, RS^• reactivity with ^•NO is a known source of protein S-nitrosothiols, an oxidative cysteine modification whose role in redox signaling has gained increasing attention in recent years (309). Thiyl radicals are quite reactive toward hydrogen abstraction: Cysteine thiyl radical is known to abstract hydrogen from its α-carbon, facilitating the loss of sulfur and cysteine conversion to dehydroalanine (239).

A significant challenge in the investigation of the reductive effects of radiation on cellular proteins is the required quenching of ^•OH formed in situ, which enables monitoring of reductive processes but may distort the view of the wider cellular implications. A recent study of bovine ribonuclease A (RNase A) using ESI-Q-TOF MS revealed that high-dose IR generated e_aq⁻ and H^• induced Cys>Ala conversion at three sites along with one Met>Aba desulfurization (108). Interestingly, reductively stressed RNase A affected only three of the eight Cys and one of four Met residues, illustrating a degree of site selectively for reductive protein desulfurizations. The chemoselectivity of these reductive processes, though not well understood, has been observed for the radiation-induced reduction of human serum albumin as well, though at 50–300 Gy doses (294). For instance, perthiyl radical (RSS^•) was found to be the major reductive product in lysozyme radiolysis because of the lack of competition reactions for the thiyl radicals (105). It has also been reported that thiyl radicals formed in the Aba production contributed to isomeric formation of trans fatty acid from the cis residues in lipids, which again illustrates the interplay between protein-lipid damage orchestrated by H^• atoms (109).

Protein expression

Global profiling of protein expression changes in response to IR is an active area of radiation research. Such investigations include, for example, the comparative proteomics analysis of IR-induced changes in the expression of mouse liver proteins that are associated with specific subcellular fractions. Such findings pointed to the activation of antioxidant and inflammatory responses by IR (206). Another study looked at the alteration of protein expression in different tissues such as the brain, lung, spleen, and intestine in an IR-treated (1 Gy) mouse using MS-based proteomics (204). Transaldolase 1 and phosphoglycerate kinase 1 were assigned as potential tissue-specific biomarkers of IR exposure in the brain and intestine, respectively. Further applications of MS-based proteomics to detect and quantify protein expression and IR-relevant PTMs are reviewed in the section “Redox and phosphoproteomics.”

Emerging Fields and Technologies for Mapping Cellular Responses to IR

Genomics and epigenomics

Differential responses to IR can be partly attributed to the combined effects of genetic and epigenetic variations among individuals, and it is well established that the diversity in “allelic architecture” relates closely to the biological response to IR (6). The field of radiogenomics has thus emerged in recent years to investigate patient-to-patient variability in normal tissue reactions after radiotherapy. Single nucleotide polymorphisms account for 90% of this naturally occurring sequence variation. Over the years, a number of clinical studies have been carried out to test the association of various genetic factors with radiosensitivity with the goal of developing algorithms that predict the normal tissue toxicity after radiotherapy (5). A study of Danish breast cancer patients showed a relationship between the ATM codon Asp1853Asn (5557 G>A) single nucleotide polymorphism and radiosensitivity in which the homozygous (AA) and heterozygous (AG) genotypes of the ATM G5557 G>A variant were more radiosensitive and at an increased risk of radiation-induced subcutaneous fibrosis (7, 63).

Epigenetics, defined as “the structural adaptation of chromosomal regions so as to register, signal, or perpetuate altered activity state,” involves modifications such as DNA methylation, histone methylation and acetylation, ADP-ribosylation, ubiquitination, non-coding RNA-mediated gene silencing, and others (35, 82). Epigenetic modifications are mediators that translate a combination of genetic, metabolic, and environmental factors into a characteristic gene expression pattern. Over the past several years, a number of groups have investigated changes in DNA methylation that were associated with IR and revealed differences which were dependent on the tissue analyzed, gender, and radiation dose (125, 182, 186, 212). Studies using animal models have shown that IR treatment (5 Gy) results in significant DNA hypomethylation in radiation-treated tissues (181). DNA lesions, including strand breakage, deletions, and nucleobase oxidation, are shown to adversely affect the action of DNA methyltransferase enzymes, leading to hypomethylation (345). The presence of 8-oxodG, a prominent oxidation product correlated with IR, reduces the methylation efficiency of nearby cytosines in a proximity-dependent manner, illustrating one example of how IR-generated ROS impacts the epigenome (334, 355). Interestingly, low-dose IR (up to 0.03 Gy in mice) during gestation was linked to an increase in DNA methylation in offspring; this effect was negated with maternal dietary antioxidants, suggesting oxidation as a link between IR and DNA methylation (29).

DNA methylation status was also found to correlate with the response and resistance to IR. For example, the treatment of MCF-7 cells with fractionated IR (2 Gy fractions for a total of 10–20 Gy) decreased DNA methylation at the Forkhead box C1 and Trafficking protein particle complex 9 loci, which was linked to decreased apoptosis and increased post-IR survival (186). In another study, microarray analysis of the CpG methylation profiles of radiosensitive H460 and radioresistant H1299 human non-small-cell lung cancer cell lines identified 1091 differentially methylated genes (174). Interestingly, twice as many hypermethylated genes were identified in the H1299 radioresistant cell line compared with the H460. The function of hypermethylation of the SERPINB5 and S100A6 genes in the radiation resistance phenotype was further validated by down-regulating their expression in H460 cells, which increased their resistance to radiation. Hypomethylation of the gene encoding catalase was observed, consistent with its increased expression in the H1299 cells and a general function of the antioxidant system in regulating the resistance to radiation.

The connection between the central redox metabolism (citric acid cycle and mitochondrial ETC) and epigenetic modulation in cancer was covered in detail in a recent review (148). To the best of our knowledge, there are no studies that present direct evidence of epigenetic modulation as a result of IR-induced modifications of metabolic or signaling proteins, but certainly these could be linked. Many epigenetic enzymes are directly regulated by phosphorylation and oxidative PTMs or use substrates (e.g., S-adenosyl methionine—a methyl donor; acetyl coenzyme A—an acetyl donor) that are products of redox-regulated metabolic pathways. Thus, connective networks that link signaling, metabolism and epigenetics exist and are being discovered using a combination of the emerging technologies discussed here.

Such discoveries would not have been possible without the development of high-throughput technologies such as microarray platforms, ChIP-on-chip, next generation sequencing, and MS. To strengthen these technologies, the development of sophisticated statistical methods is required for an accurate interpretation of the high volume of data generated from such experiments. Often a combination of the aforementioned techniques is used for better results. For example, ChIP from paraffin-embedded patient samples (PAT-ChIP) can be coupled with high-throughput sequencing (PAT-ChIP-seq) for the genome-wide analysis of distinct chromatin modifications (104). Sometimes, ancillary techniques are a prerequisite for the subsequent high-throughput microarray or sequencing analysis. For example, in studies involving the analysis of methylated cytosines, the first step often involves techniques such as Methylated DNA immunoprecipitation, Methylated-CpG island recovery assay, and bisulfite conversion for the enrichment of methylated cytosine DNA followed by an array-based approach (such as tile array, promoter array, or bead array) or a sequence-based approach for epigenetic profiling (180).

Redox and phosphoproteomics

Redox proteomics

Cysteine oxidation, in particular, attracts greater attention in the redox signaling field than other amino acids because of its active site and regulatory roles in numerous signaling and metabolic proteins, including phosphatases, kinases, and transcription factors. Here, we focus on MS-driven proteomic methods for detecting protein sulfenic acids and S-nitrosothiols. Since CySOH and CySNO species are often trapped by proximal thiols and form disulfides, the assessment of cellular thiol-disulfide status is a widely utilized approach for studying cysteine oxidation (140, 198, 280, 373). Isotope-coded affinity tag for detecting oxidized cysteine (OxICAT) is a method for quantifying the cellular thiol and disulfide species in complex samples using an isotopically light electrophile to first capture thiols, followed by global disulfide reduction and nascent thiol capture by an isotopically labeled form of the electrophile (197, 301). An important caveat for OxICAT and related methods is that the reductant employed, tris(2-carboxyethyl)phosphine (TCEP), reduces S-nitrosothiols and sulfenic acids in addition to disulfides. Quantification using this method, therefore, determines the extent of reversibly oxidized cysteines, unless more chemoselective reductants, such as ascorbate or arsenite, are used in lieu of TCEP (52, 295). Slight modifications to the OxICAT approach, such as care to avoid protein acidification, enable the use of this method to detect oxidative and nitrosative protein modifications in the same sample (isotope-coded affinity tag for detecting nitrosated and oxidized cysteine) (207).

Investigations of the CySNO or CySOH proteome are challenged by the low cellular abundance of these PTMs in proteins, necessitating the use of enrichment strategies and therefore requiring the use of biotin-conjugated reagents for labeling the oxidized/nitrosated cysteine in proteins. Unfortunately, the MS signal intensity of a biotin-containing peptide ion is often significantly lower than the ion of the corresponding non-biotinylated peptide. As a result, several groups, including ours, have developed CySOH–, CySNO–, and other oxoform-targeted probes with cleavable biotin tags to facilitate MS compatibility (272, 301, 333). A number of methods developed to selectively detect and quantify the CySNO or CySOH proteomes are discussed next.

Proteomic studies of protein nitrosation (CySNO)

Due to their thermal instability and decomposition in light, nitrosated protein cysteines are not directly detected and require chemical derivatization. A common method for detecting S-nitrosocysteine in biological samples is the biotin-switch technique (BST), a three-step sequence in which (i) free thiols are captured by an electrophile (blocking); (ii) S-nitrosothiols are reduced by ascorbate in the presence of a labeled electrophile; and (iii) labeled thiols are quantified using fluorescence imaging, Western blot, or MS. This approach, though indirect, is widely used and fairly controversial, especially with regard to the chemoselectivity of thiol blocking and ascorbate reduction. Careful controls are paramount to avoid the misinterpretation of BST results. Alternative methods include CySNO decomposition, accompanied by chemiluminesence detection of nascent ^•NO, and phosphine-based reagents that react directly with the S-nitrosothiol and form chemically unique conjugates (19, 21, 347, 381). Nevertheless, numerous methods have been developed that take advantage of the ascorbate-based reduction of CySNO. Site mapping of an S-nitrosated cysteine in PKBα/Akt1 from rat soleus muscle was determined using trypsin digestion and derivatization of CySNO-containing peptides with isotopically labeled iodoacetic acid (2-¹²C/¹³C, 50% ratio) in the presence of ascorbate (213). To globally identify S-nitrosated cysteine sites, the S-nitrosothiol site identification method was adapted from the BST approach but incorporates protein digestion before affinity capture, resulting in an enriched CySNO-containing peptide mixture (132, 141). Alternatively, the His-Tag switch approach uses a highly MS-compatible His-tag linked to iodoacetamide for tagging nascent thiols generated from CySNO reduction by ascorbate (52). A more recent technique is resin-assisted capture (resin-assisted capture for S-nitrosothiols [SNO-RAC]), where a thiol-reactive resin captures ascorbate-generated thiols. SNO-RAC is more streamlined than the classical BST, enables on-resin digestion, is compatible with isotope tagging for relative and absolute quantification (iTRAQ) quantification, and offers higher sensitivity than the BST for high MW proteins (111). These and other CySNO detection methods have been recently reviewed (20).

Proteomic studies of protein sulfenylation (CySOH)

Similar to the S-nitrosothiol, several direct and indirect methods were developed for the detection of CySOH proteome. These are thoroughly reviewed in a recent article (117), and only a brief summary is presented here. The indirect BST approach is used for the detection of CySOH with the difference that arsenite is used instead of ascorbate to achieve selective reduction of CySOH; this is followed by a reaction of the nascent thiol with a biotin (or fluorescein) tagged thiol electrophile (e.g., biotin iodoacetamide) followed by enrichment and detection using MS or imaging technologies (295). A potential caveat is the chemoselectivity of thiol-blocking and oxoform reduction, which are not unequivocally established (281, 348, 383). More direct methods include detection using a CySOH-specific antibody, which relies on the CySOH modification remaining intact during sample processing (300). Another approach uses in vivo labeling by a genetically engineered C-terminal cysteine-rich domain of the transcription factor Yap1 (Yap1-cCRD) (321, 322). The single Cys of the Yap1-cCRD mutant forms mixed disulfides with CySOH in vivo, enabling co-enrichment using nickel affinity columns. Although not yet extensively investigated, this strategy may suffer from false identifications via protein cross-links in vivo.

The lability of sulfenic acids and controversies involving the potential for false positive results with the BST has fueled the shift toward the direct detection of protein sulfenic acids using 1,3-dicarbonyl (often dimedone-based) probes (Fig. 7A). To accommodate the many experimental techniques used for profiling sulfenic acid formation, both in cells and in lysates, a diverse library of 1,3-dicarbonyl probes has been developed with functionalities for affinity capture, MS, and fluorescence imaging. In addition, alkyne- and azide-containing dicarbonyl probes enable their customization using click chemistry (57, 178, 200, 267, 271, 278, 299, 333).

FIG. 7.

Capture of cysteine sulfenic acid by 1,3-dicarbonyl compounds. (A) Dimedone (top), Alk-β-KE (bottom); click chemistry (not shown) installs a biotin affinity tag; the ketoester enables hydroxylamine-mediated ester cleavage, generating MS-friendly isoxazolone. (B) MS² spectrum for the Alk-β-KE-labeled peptide after Asp-N digestion reveals MS compatibility. (C) Accumulation of intracellular ROS in response to probes is lower for Alk-β-KE compared with dimedone (10 mM each). (D) Click chemistry attaches biotin tag; labeling by Alk-β-KE is specific to CySOH and abolished in the presence of TCEP. (E) Addition of Alk-β-KE to cells, followed by lysis, click to attach the biotin moiety and immunoblotting illustrates membrane permeability. (F) Cellular toxicity of Alk-β-KE is reduced compared with dimedone as measured by the DCF assay. Adapted from Qian et al. (272). DCF, dichlorodihydrofluorescein; MS, mass spectrometry; TCEP, tris(2-carboxyethyl)phosphine. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

MS-based CySOH proteomics studies have been achieved using a biotinylated dimedone analog in heart tissue (57), in HeLa cells using dimedone-based azide DAz-2 subsequently conjugated to biotin via Staudinger ligation (200), and in tumor necrosis factor α-stimulated HEK-293 cells using biotin-containing DCP-Bio1 (248). Despite that fact that a plethora of CySOH-containing proteins have been identified through MS-based proteomics using a diverse set of chemical probes, global mapping of redox-sensitive cysteines has not been successfully completed. As discussed earlier, biotin is essential for enrichment but often problematic for the detection of low-abundance peptides by MS. A recent 1,3-dicarbonyl CySOH probe, Alk-β-KE, incorporates a β-ketoester linkage that is susceptible to chemoselective cleavage by hydroxylamine (NH₂OH), enabling on-bead cleavage of labeled proteins or peptides and generating a highly MS-compatible isoxazolone tag (Fig. 7B). This reagent is cell-membrane permeable and has reduced cell toxicity compared with dimedone (Fig. 7C–F) (272). The use of Alk-β-KE and other CySOH probes in large-scale proteomics studies is ongoing with the ultimate goal of uncovering IR-dependent protein oxidation along with the modified cysteine site(s) (Fig. 8). The direct labeling approaches with chemical probes can easily be integrated with relative or absolute quantitative proteomics methods such as stable isotope labeling of amino acids in cell culture (SILAC), iTRAQ, tandem mass tags, absolute quantification, and others, enabling studies of time course or radiation dose-dependence profiles of protein oxidation, phosphorylation, and other PTMs discussed next (103, 203, 254).

FIG. 8.

General workflow for in vitro labeling of CySOH using biotinylated 1,3-dicarbonyl probes. Proteins are precipitated to remove unreacted probe; then, labeled proteins are enriched using avidin-based methods. After proteolytic digestion, protein identifications are obtained by LC-MS/MS analysis of the full peptide mixture, while site mapping requires a second round of enrichment. LC, liquid chromatography. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

Phosphoproteomics and other targeted proteomics

SILAC-based quantitative proteomic analyses were recently applied to identify the dynamics of human nuclear protein acetylation and phosphorylation during IR-induced DDR (26, 27). Cumulatively, these two studies have revealed a number of regulatory pathways linking acetylation and phosphorylation signaling to DNA methylation and the regulation of gene expression. In another study, phosphoproteomic methods were employed in conjunction with microarray analyses of mRNA and miRNA expression to quantify changes induced in these species by fractionated radiation versus single-dose radiation. Significant differences were identified in DDR pathways, apoptosis, and immune response (221). Quantitative phosphoproteomics was also applied to identify previously unknown ATM-dependent and independent changes in the phosphorylation of nuclear proteins after DNA damage (28).

Metabolomics

Nuclear magnetic resonance (NMR) and MS-based metabolomics has emerged as a powerful approach for identifying and quantifying biomarkers of IR exposure in both cell culture and in vivo animal studies, including in nonhuman primates. However, the high complexity of the metabolome, the dependency on factors such as gender, diet, age, and genetic background, and the variability of metabolite composition between cell types and tissues make it difficult to employ metabolomics as a stand-alone approach. Nevertheless, this technology is extremely powerful when used in defined systems and in conjunction with genomics, epigenomics, and/or proteomics studies.

Metabolomics studies in radiation biology are divided into two categories: (i) those aimed at the mechanistic understanding of the biological response to IR, and (ii) biodosimetry studies focused on quantifying the damage induced by IR and identifying the respective biomarkers in easily accessible biological fluids (urine, saliva, and serum). For an excellent discussion on the use of radiation metabolomics to identify small-molecule biomarkers that respond to radiation in a dose-dependent manner, the reader is directed to several recent publications (163, 164, 188, 335, 336). These include a description of current and emerging metabolomics technologies, computational methods for data analysis, as well as specific studies that have led to the identification and, ultimately, validation of a number of urinary biomarkers in mice and non-human primates. Analytical methods used to identify metabolite biomarkers and to quantify the extent of metabolite modulation by IR include NMR spectroscopy and MS. One-dimensional and two-dimensional NMR provide structural information for metabolomics studies and are, therefore, critical tools in the characterization of new metabolites. NMR spectroscopy is, however, hampered by low sensitivity, requiring larger amounts of sample than other analytical techniques, but despite this limitation, it has provided valuable information about IR biomarkers (61, 171). MS-based metabolomics methods, in addition to increased sensitivity, also incorporate online analytical separation of metabolites using GC, LC, or capillary electrophoresis.

Examples of radiation biomarkers from studies in non-human primates include 2′-deoxyxanthosine, 2′-deoxyuridine, thymidine, xanthine, and xanthosine (163). Other studies in cell culture showed a good correlation between the changes in small molecule metabolites (e.g., AMP) and the expression of proteins involved in their metabolism as determined by a proteomic analysis (e.g., increase in the adenylate kinase that converts AMP to ADP and a concomitant decrease in the AMP-generating enzymes such as phosphodiesterase (cAMP to AMP) and hypoxanthine-guanine phosphoribosyltransferase, which converts adenine to AMP) (156, 256). In addition, studies of skin exposure to low doses of radiation (3–10 cGy) found significant changes in the metabolites involved in the DNA damage, damage repair, and bioenergy metabolism (155).

In a cell culture, a number of studies have reported metabolomic (small or large scale) changes by IR, and a few of these have applied integrated “omics” analyses (53, 60, 341). Using integrated “omics” analyses, ATM was found to regulate purine, pyrimidine, and urea cycle metabolism through the activation of AMP-activated protein kinase, a crucial sensing enzyme in the regulation of cellular energy pathways (60, 341). In another study and using NMR-based metabolomics, the inhibition of small ubiquitin-like modifier-mediated protein–protein interactions was connected to decreased mitochondrial function and impairment of GSH synthesis after exposure to IR (4 Gy) (53).

GSH content was also shown to undergo dynamic time- and dose-dependent changes in response to IR in human lymphoblast TK6 and human telomerase reverse transcriptase (hTERT)-transformed fibroblast BJ cells (256). In this study, cells were treated with 0.5–8 Gy, and the water-soluble metabolome fraction was analyzed at 1–16 h after IR exposure using ultra performance liquid chromatography-ESI-TOF MS. Extensive metabolomic changes were observed as early as 1 h after IR in a dose-dependent manner. Certain metabolites were depleted at 1 h post-irradiation with 1 Gy, including GSH, AMP, nicotinamide, 5-oxoproline (intermediary metabolite in GSH biosynthesis), phosphocholine (lipid metabolism), uridine monophosphate, proline, NAD⁺, and spermine (protects against DNA damage). At 4 h, the levels of these metabolites were partially restored to baseline, and by 8 h, there was no significant difference between IR-treated and control. At 16 h (after 4 and 8 Gy only), a reduction of GSH, NAD⁺, and spermine was only observed in the cells treated with 4 and 8 Gy.

Interestingly, there was a difference in the time- and dose dependence of GSH in the hTERT-transformed fibroblast BJ cells compared with the TK6 cells (256). First, there was an increase in GSH at 1 h after 1 Gy, a depletion of GSH at 4 h and 4 Gy, followed by an increase again at 16 h at the 4 Gy IR dose. The authors attributed these observations to differences between normal and transformed phenotypes, but certainly, additional studies are needed to thoroughly profile crucial metabolites across a range of cell lines, radiation treatments, and time elapsed after radiation treatment.

Within cells, mitochondria are an established target of IR damage (166), and studies have linked the response to IR with changes in specific metabolites. Investigations of IR-induced micronucleus formation have linked the functional status of mitochondrial DNA (mtDNA) to ATP production and radiosensitivity (375). Cells lacking mtDNA showed a decreased sensitivity to radiation and a lower rate of micronucleus formation compared with cells harboring intact mtDNA, which was correlated primarily with decreased ATP production.

Mitochondria-generated ROS were also found to be involved in micronucleus formation (66), and more recent studies have investigated the function of mitochondrial SOD (manganese superoxide dismutase [MnSOD]) in the response to radiation. Increased expression of MnSOD was associated with increased resistance to IR both in vivo (102) and in vitro (152). The function of MnSOD extends beyond protection against IR-induced ROS: A study investigating the function of MnSOD in the adaptive response after low-dose radiation (10 cGy X-ray) using human skin keratinocytes (HK18) found an association of MnSOD with metabolic proteins (100). Thus metabolomic changes associated with altered MnSOD expression are expected but have not been fully elucidated.

IR-induced changes in the serum of rats exposed to approximately 8 Gy of IR showed a decreased prevalence of citric acid cycle intermediates such as isocitrate (3 and 8 Gy) (208), citric acid, and 2-oxoglutaric acid (α-ketoglutarate) (188). Knockdown of the mitochondrial isocitrate dehydrogenase, which catalyzes the oxidative decarboxylation of isocitrate to α-ketoglutarate, induced G1 cell cycle arrest and autophagy after 5 Gy of IR (177). The inhibition of autophagy using chloroquine in conjunction with IR treatment resulted in apoptotic cell death. Cumulatively, these studies establish a connection between the metabolic intermediates of the citric acid cycle and the biological response to IR. Furthermore, levels of creatine, another mitochondrial metabolite, were increased in human leukocytes exposed to 4 Gy (γ-rays) along with arginine, glutamine, proline, and GSH (measured as reduced GSH) (194). Circulating citrulline, a biomarker of intestinal function and a mitochondrial metabolite, was decreased in mice at day 4 after 8–12 Gy of proton IR (217).

Radiation metabolomics is still an emerging area of research, and efforts to uncover organelle-specific metabolic changes are ongoing. The current chromatography/MS approaches to metabolomics will be enhanced by the growing use of high-resolution mass spectrometers, which offer sharply increased sensitivity and mass-based molecule identification. Thus, metabolomics is expected to become a critical “omics” tool in radiation research and shows significant promise in the discovery of small-molecule biomarkers of IR damage and response that could be exploited for radiation biodosimetry or cancer therapeutics.

Potential caveats of “omics” approaches

Though high-throughput “omics” technologies are providing a wealth of biological information and have proved valuable to identify critical networks affected by radiation, “omics” analyses and the interpretation of their results have a number of limitations. First, the sheer complexity of the human metabolome and inability to predict its composition makes metabolomics and follow-up studies in this area of research highly challenging. Proteomic studies are complicated by the potential for a multitude of protein modifications with varying degrees of contribution to protein activity and as a result, the overall IR effect. Technologies that would enable a high-throughput assessment of protein activity in vivo which could then be used to build computational models of a system (cell, tissue, and body) would be ideal. To date, it is only for selected classes of proteins that these technologies exist (16, 202, 249). “Omics” approaches have traditionally been viewed as “fishing expeditions” and criticized as simply providing snapshots of biological systems. The acquisition of time course and dose-dependence “omics” analysis increased sophistication of data integration and computational analysis, and follow-up studies both in cells and in vivo have made significant strides toward alleviating these concerns and are now a mainstay in the field.

Conclusion and Perspectives

From the studies reviewed here and many others, it has become clear that in order to achieve an understanding of biological systems at the level which would enable a prediction of targets to manipulate the biological outcome, a thorough integration of “omics” and classical technologies through novel computational approaches is needed.

Given the wide range of PTMs induced by IR (Table 1) with their effects on the activity of individual molecules and on the metabolic flux and propagation of signaling networks mostly unknown, future efforts will require the inclusion of protein activity profiles to generate intricate computational models of biological response to IR. Too often in scientific literature, the altered phosphorylation status of a protein is equated with gain or loss of activity. Evidence from our group and others shows that oxidation can easily override the effects of phosphorylation, activating or inhibiting the activity of enzymes (260, 349). This is extremely important in the context of IR-dependent signaling and in studies that integrate metabolomics, proteomics, and epigenetics. While indirect methods for the detection of oxidative PTMs have been applied with various degrees of success, many caveats are associated with these approaches, particularly when a quantitative analysis is attempted. For example, the cross-reactivity of protein sulfenic acids with DNPH and other hydrazines used for detecting protein carbonylation is established, and recently, the cross-reactivity of sulfenic acids with commonly used thiol-blocking electrophiles has been described (281). Therefore, the arsenal of probes that chemically target each of the oxidative and other PTMs should be expanded. As discussed in the current review, the development of robust and selective detection strategies for modified biomolecules is a vibrant area of research both within and outside the radiation biology field. The use of chemical probes in conjunction with high-throughput technologies and multi-scale computational methods opens exciting opportunities for radiation research by facilitating the discovery of new targets and biomarkers of IR and the prediction of long-term effects of IR. Such advances will also enable more detailed investigations of the mechanisms underlying the biological response to IR, such as understanding the mechanistic connections among epigenetic changes (e.g., DNA methylation), DNA repair systems, and metabolism. Further elucidation of the biological mechanisms of IR will impact multiple areas of medicine, from prevention of disease induced by IR exposure to selection of IR-based treatment schedules for cancer.

Abbreviations Used

γH2AX

phosphorylated histone H2AX

8-oxodG

8-oxo-2′-deoxyguanosine

Aba

α-aminobutyric acid

ACEI

angiotensin-converting-enzyme inhibitors

Ang II

angiotensin II

ApoAI

apolipoprotein A-I

ApoAII

apolipoprotein A-II

ASMase

acid sphingomyelinase

AT1R

angiotensin II type 1 receptor

BST

biotin-switch technique

CHH

7-(diethylamino)coumarin-3-carbohydrazide

Chk1/2

checkpoint kinases 1/2

CID

collision-induced dissociation

CT

computed tomography

DCF

dichlorodihydrofluorescein

DDR

DNA damage response

DJ-1

Parkinson disease protein 7

DNA-PK

DNA-dependent protein kinase

DNPH

2,4-dinitrophenylhydrazine

DPPP

diphenyl-1-pyrenylphosphine

DSB

double-strand break

e_aq⁻

hydrated electron

EGFR

epidermal growth factor receptor

ELISA

enzyme-linked immunosorbent assay

ESI

electrospray ionization

ETC

electron transport chain

FISH

fluorescence in situ hybridization

GC

gas chromatography

GC-MS

gas chromatography coupled to mass spectrometry

GPR

Girard's P reagent

GPx

glutathione peroxidase

GSH

glutathione

GST

glutathione S-transferase

Gy

gray

H^•

hydrogen radical

H₂O₂

hydrogen peroxide

HNE

4-hydroxy-2-nonenal

HPLC

high-performance liquid chromatography

hTERT

human telomerase reverse transcriptase

IHC

immunohistochemistry

iNOS

inducible nitric oxide synthase

IR

ionizing radiation

iTRAQ

isotope tagging for relative and absolute quantification

LC

liquid chromatography

LOOH

lipid hydroperoxide

MDA

malondialdehyde

MetO

methionine sulfoxide

MnSOD

manganese superoxide dismutase

MS

mass spectrometry

Msr

methionine sulfoxide reductase

mtDNA

mitochondrial DNA

N₂O₃

dinitrogen trioxide

NADPH

nicotinamide adenine dinucleotide phosphate, reduced form

NF-κB

nuclear factor kappa-light-chain-enhancer of activated B cells

NMR

nuclear magnetic resonance

^•NO

nitric oxide

NO₂^•

nitrogen dioxide

NOX

NADPH oxidase

NOxICAT

isotope-coded affinity tag for detecting nitrosated and oxidized cysteine

O₂^•−

superoxide

^•OH

hydroxyl radical

ONOO⁻

peroxynitrite anion

OxICAT

isotope-coded affinity tag for detecting oxidized cysteine

PHGPx

phospholipid hydroperoxide glutathione peroxidase

Prx

peroxiredoxin

PTM

post-translational modification

PUFA

polyunsaturated fatty acid

RAS

rennin–angiotensin system

RNase A

ribonuclease A

RNS

reactive nitrogen species

ROS

reactive oxygen species

SCX

strong cation exchange

SILAC

stable isotope labeling of amino acids in cell culture

SNO-RAC

resin-assisted capture for S-nitrosothiols

SOD

superoxide dismutase

SSB

single-strand break

Sv

sievert

TCEP

tris(2-carboxyethyl)phosphine

TGF-β

transforming growth factor β

TrxR

thioredoxin reductase

TUNEL

terminal transferase

UHPLC-MS/MS

ultra high-performance LC coupled to tandem MS.

UV

ultraviolet

Acknowledgments

Financial support was provided by funds from NIH R01 CA136810 (C.M.F.) and pilot funds from the Center for Molecular Signaling and Communication at Wake Forest University (C.M.F.). The authors thank Nelmi O. Devarie Baez, PhD, and Jeffrey G. Kuremsky, MD, for helpful feedback on this work.