Additive interaction of gefitinib ('Iressa', ZD1839) and ionising radiation in human tumour cells in vitro.

Cultures of human carcinoma A-431, A-549 and HeLa cells were challenged with gamma-rays without or with concomitant exposure to gefitinib, a potent inhibitor of the tyrosine kinase activity of epidermal growth factor receptor (EGFR). The outcome of treatment was determined from cell and colony count, cell cycle progression and DNA double-strand break formation and rejoining. Apoptosis was measured in parallel from hypodiploid DNA and using an annexin V assay. Gefitinib developed a cytostatic effect in all cell lines, with drug sensitivity correlating the level of EGFR expression. A weak cytotoxicity of gefitinib was observed in HeLa cells only, although the drug was unable to induce significant cell cycle redistribution in this cell line. In contrast, substantial G1 block and S-phase depletion was observed in A-431 and A-549 cells exposed to gefitinib. The drug brought about additive to subadditive interaction with radiation with regard to growth inhibition, clonogenic death and induction of apoptosis. Consistently, gefitinib did not hinder the rejoining of radiation-induced DNA double-strand breaks in any cell line. The results demonstrate that gefitinib may elicit cytotoxicity at high concentration, but does not act as a radiosensitiser in vitro in concomitant association with radiation.

Hyperfast, early cell response to ionizing radiation.

PURPOSE: To determine whether the oscillatory changes of radio-sensitivity which occur within fractions of a second to a few minutes following flash irradiation correlate with an altered incidence of apoptosis, DNA strand breaks or lipid-coupled signalling. MATERIALS AND METHODS: Human tumor cells (SQ-20B, LoVo) or Chinese hamster V79 fibroblasts were exposed to split-dose, pulse irradiation with 3.5 MeV electrons at high dose-rate (12 or 120 Gy x s(-1)) and the effects assessed by clonogenic assays, analysis of DNA cleavage and microscopic observation. RESULTS: The processes underlying oscillatory radiation response were saturable, but did not correlate with an increased incidence of DNA single- or double-strand breaks or apoptosis. N-acetylcysteine and inhibitors of lipid-derived signalling also failed to alter oscillatory response. However, this response did correlate with phenotypic alterations evoking mitotic or delayed cell death. Furthermore, high dose-rate irradiation provided a lower level of instability than protracted gamma-ray irradiation. CONCLUSIONS: It is proposed that the early steps of DNA damage recognition and repair following priming radiation exposure bring about rapid, synchronous remodeling of chromatin, evoking enhanced chromosome damage upon re-irradiation.

Radiation-induced arrest of cells in G2 phase elicits hypersensitivity to DNA double-strand break inducers and an altered pattern of DNA cleavage upon re-irradiation.

PURPOSE: To determine how radiation-induced arrest in G2 affects the response of mammalian cells to a challenging dose of radiation or to antitumour drugs producing DNA double-strand breaks. MATERIALS AND METHODS: V79 fibroblast survival to 5 Gy gamma-rays followed at intervals by 3 Gy irradiation or by contact with an equitoxic dose of neocarzinostatin or etoposide, was measured by clonogenic assays. The pattern of radiation-induced DNA double-strand breaks was determined by filter elution and CFGE (continuous field gel electrophoresis) or PFGE (pulsed-field gel electrophoresis) in G2-arrested cells as well as in nonpre-irradiated asynchronous or synchronized cells. The cell-cycle phase specificity of drug susceptibility was determined in synchronized HeLa cells. RESULTS: Cell kill by radiation-drug combined treatment varied markedly with the time elapsed after priming irradiation. Pre-irradiated, G2-arrested V79 fibroblasts demonstrated excess double-stranded DNA cleavage upon re-irradiation and hypersensitivity to drugs and radiation, although maximum resistance to both neocarzinostatin and etoposide in synchronized HeLa cells was in G2. This effect occurred in the megabase range only, with a peak around 4 Mbp; no change in the electrophoretic migration profile of DNA was observed below 1 Mbp. Moreover, the DNA migration profile and the yield of DNA cleavage in G2-arrested cells were close to those expected from S-phase cells. CONCLUSION: The available data suggest that mechanisms operating within the radiation-induced G2 block promote susceptibility to DNA double-strand break inducers at this stage. It is also proposed that the conformation of DNA in cells accumulated in G2 following irradiation bears resemblance to that for cells in S phase, due either to active repair mechanisms or to inhibition of chromosome disentanglement at the S-G2 transition.

Poly(ADP-ribose) polymerase, a major determinant of early cell response to ionizing radiation.

PURPOSE: To determine whether DNA-dependent protein kinase (DNA-PK) and poly(ADP-ribose) polymerase (PARP-1) are involved in eliciting the rapid fluctuations of radiosensitivity that have been observed when cells are exposed to short pulses of ionizing radiation. MATERIALS AND METHODS: The effect of DNA-PK and PARP-1 inhibitors on the survival of cells to split-dose irradiation was investigated using Chinese hamster V79 fibroblasts and human carcinoma SQ-20B cells. The responses of PARP-1 proficient and PARP-1 knockout mouse 3T3 fibroblasts were compared in a similar split-dose assay. RESULTS: Inactivation of DNA-PK by wortmannin potentiated radiation-induced cell kill but it did not alter the oscillatory, W-shaped pattern of early radiation response. In contrast, oscillatory radiation response was abolished by 3-aminobenzamide, a reversible inhibitor of enzymes containing a PARP catalytic domain. The oscillatory response was also lacking in PARP-1 knockout mouse 3T3 fibroblasts. CONCLUSION: The results show that PARP-1 plays a key role in the earliest steps of cell response to ionizing radiation with clonogenic ability or growth as endpoint. It is hypothesized that rapid poly(ADP-ribosylation) of target proteins, or recruitment of repair proteins by activated PARP-1 at the sites of DNA damage, bring about rapid chromatin remodelling that may affect the incidence of chromosomal damage upon re-irradiation.

[DNA-dependent protein kinase (DNA-PK), a key enzyme in the re-ligation of double-stranded DNA breaks]. / La protéine kinase dépendante de l'ADN (DNA-PK), une enzyme clé de la religation des cassures double-brin de l'ADN.

Repair pathways of DNA are now better defined, and some important findings have been discovered in the last few years. DNA non-homologous end-joining (NEHJ) is a crucial process in the repair of radiation-induced double-strand breaks (DSBs). NHEJ implies at least three steps: the DNA free-ends must get closer, preparation of the free-ends by exonucleases and then a transient hybridisation in a region of DNA with weak homology. DNA-dependent protein kinase (DNA-PK) is the key enzyme in this process. DNA-PK is a nuclear serine/threonine kinase that comprises three components: a catlytic subunit (DNA-PKCS) and two regulatory subunits, DNA-binding proteins, Ku80 and Ku70. The severe combined immunodeficient (scid) mice are deficient in DNA-PKCS: this protein is involved both in DNA repair and in the V(D)J recombination of immunoglobulin and T-cell receptor genes. It is a protein-kinase of the P13-kinase family and which can phosphorylates Ku proteins, p53 and probably some other proteins still unknown. DNA-PK is an important actor of DSBs repair (induced by ionising radiations or by drugs like etoposide), but obviously it is not the only mechanism existing in the cell for this function. Some others, like homologous recombination, seem also to have a great importance for cell survival.

Synthesis and in vitro evaluation of novel 2,6,9-trisubstituted purines acting as cyclin-dependent kinase inhibitors.

Novel C-2, C-6, N-9 trisubstituted purines derived from the olomoucine/roscovitine lead structure were synthesized and evaluated for their ability to inhibit starfish oocyte CDK1/cyclin B, neuronal CDK5/p35 and erk1 kinases in purified extracts. Structure activity relationship studies showed that increased steric bulk at N-9 reduces the inhibitory potential whereas substitution of the aminoethanol C-2 side chain by various groups of different size (methyl, propyl, butyl, phenyl, benzyl) only slightly decreases the activity when compared to (R)-roscovitine. Optimal inhibitory activity against CDK5, CDK1 and CDK2, with IC50 values of 0.16, 0.45 and 0.65 microM, respectively, was obtained with compound 21 containing a (2R)-pyrrolidin-2-yl-methanol substituent at the C-2 and a 3-iodobenzylamino group at the C-6 of the purine. Compound 21 proved cytotoxic against human tumor HeLa cells (LD50-6.7 microM versus 42.7 microM for olomoucine, 24-h contact). Furthermore, unlike olomoucine, compound 21 was effective upon short exposure (LD50= 25.3 microM, 2-h contact). The available data suggest that the affinity for CDKs and the cytotoxic potential of the drugs are inter-related. However, no straightforward cell cycle phase specificity of the cytotoxic response to 21 was observed in synchronized HeLa cells. With the noticeable exception of pronounced lengthening of the S-phase transit by 21 applied during early-S in synchronized HeLa cells, and in striking contrast with earlier reports on studies using plant or echinoderm cells. olomoucilnc and compound 21 were unable to reversibly arrest cell cycle progression in asynchronous growing HeLa cells. Some irreversible hlock in GI and G2 phase occurred at high olomoucine concentration, correlated with induced cell death. Moreover, chmronic exposure to lethal doses of compound 21 resulted in massive nuclear fragmentation, evocative of mitotic catastrophe with minour amounts of apoptosis only. It was also found that olomoucine and compound 21 reversibly block the intracellular uptake of nuicleosides with high efficiency.

Pulse treatment of interphasic HeLa cells with nanomolar doses of docetaxel affects centrosome organization and leads to catastrophic exit of mitosis.

In order to investigate the role of centrosome duplication in mitotic spindle morphogenesis, we designed a 1 hour pulse treatment protocol on synchronized HeLa cells with nanomolar doses of taxoids that might impair centrosome biogenesis but would allow the recovery of normal microtubule (Mt) dynamics before mitosis. We were prompted to use this approach as docetaxel (DOC; taxotereTM), a taxoid known to promote Mt polymerization, was shown to be more cytotoxic when applied during S phase. We show that pulse drug exposure is most efficient in late S and in G2 and results in a marked disorganization of the centrosome in G2, the pericentriolar material (PCM) being dissociated from centrioles. Separation of centrosomes at the G2-M transition is also impaired and mitotic spindle morphogenesis is grossly abnormal: although in most spindles chromosomes align in a metaphase plate, the two centrosomes stay most often unseparated at one pole and most of the NuMA protein accumulates at the other. Interestingly, we find that the centrosomes' ability to duplicate is not abolished as they are still able to trigger parthenogenetic development of frog eggs. Despite spindle asymmetry, the progression through mitosis is not blocked. This results in a catastrophic exit from mitosis, each mitotic cell generating several micronucleated cells linked together by multiple midbodies. Lack of mitotic block appears therefore as the prime cause of cell lethality. These experiments suggest that NuMA redistribution at the onset of mitosis depends upon the correct redistribution of PCM between centriole pairs. They also indicate that the presence of aberrant spindle poles does not alert the surveillance mechanism controlling the exit of mitosis.

[Early cell response to radiation]. / Réponse cellulaire précoce à l'irradiation.

The early steps of cellular radiation response have been investigated using a linear electron accelerator operated in a split-dose mode, in such a way that the time intervals between pulse exposures to relativistic electrons ranged from fractions of a second to a few minutes. The initial dose brought about large, synchronous changes in radiation sensitivity and generated a tetraphasic, W-shaped time-dependent profile of cell survival upon the second radiation exposure. While this time-related process was observed in most cell lines investigated, its kinetic parameters varied significantly from one cell line to the other. The number of DNA strand breaks (neutral and alkaline DNA filter elution) and the level of apoptosis (gel electrophoresis and flow cytometry) induced at the different phases of the time-dependent profile showed no relationship with the W-effect. It is presently hypothesized that mechanisms involved in molecular recognition of radio-induced lesions and initiation of genomic instability play a major role in this effect. Whatever the mechanism involved, the split-dose irradiation in the range of seconds enables dissecting the early steps of radiation response. The relevance of the W-effect to radiation therapy and technical drawbacks are discussed.

Pulse exposure to ionizing radiation elicits rapid changes in cellular radiosensitivity.

A linear electron accelerator, operated in a recurrent chopped mode, was used for time-resolved investigation of split-dose radiation recovery in 3 mammalian cell lines in vitro. The time intervals separating the sequential radiation exposures in this study ranged from fractions of a second to a few minutes. The primary pulse brought about rapid, synchronous oscillations of cellular radiosensitivity giving rise to a tetraphasic, W-shaped time-dependent profile whose first phase was accomplished by a large decrease of cell survival. Only the last phase correlated with sub-lethal damage repair determined by gamma-ray irradiation. The same profile was observed for the 3 cell lines investigated. However, the kinetics of the whole process varied extensively from one cell line to another. The first phase lasted 1 s only for Chinese hamster V79 fibroblasts, 6 s for human squamous carcinoma SQ20B cells, and as much as 25 s for human colon adenocarcinoma LoVo cells. The relative amplitude of this first phase grew with both the first and second radiation doses in the range explored. It is hypothesized that rapid oscillation of the cytotoxic potential of radiation may result from various mechanisms such as molecular recognition of radio-induced lesions, changes in chromatin structure, or differential activation of phospholipid-dependent transduction pathways.

Interaction of ionizing radiation with paclitaxel (Taxol) and docetaxel (Taxotere) in HeLa and SQ20B cells.

Altered gamma-ray response by brief (1 h), concomitant exposure to paclitaxel (Taxol) or docetaxel (Taxotere) was investigated in growing HeLa and SQ20B human tumor cells in vitro. For both cell lines, both taxoids were able to reduce or enhance radiation cell killing, depending on the drug concentration. Large reduction of radiosensitivity (up to 3.3-fold reduction relative to radiation alone) was observed in HeLa cells over a wide range of drug concentrations, extending to 1.5- (paclitaxel) or 3.3- fold (docetaxel) the IC50s determined for drug alone. This antagonistic effect was also observed with SQ20B cells. It disappeared for drug concentrations exceeding 0.9 (SQ20B), 1.6 (HeLa; paclitaxel), and 3.4 (HeLa; docetaxel) IC50 equivalents, above which a drug dose-dependent, supra-additive radiation-drug interaction was observed. Reduction of radiation susceptibility in the low-drug dose range also held for mid-G1 synchronized HeLa cells, i.e., in the cell cycle compartment characterized as the most resistant one to docetaxel (C. Hennequin et al., Br. J. Cancer, 71: 1194-1198, 1995). In the case of SQ20B cells, the cytotoxicity of either drug or radiation alone was primarily dependent on the state of growth, with quiescent (G(0)) cells showing increased radiosensitivity and reduced drug toxicity compared to the growing fraction. The effect of taxoids (1-h contact) was finally investigated in sequential treatment as a function of the time elapsed between radiation and exposure to drugs. In HeLa cells, the postirradiation time-dependence of the response to combined treatment was biphasic. The radioprotecting potential of either taxoid disappeared in approximately 1.5 h following radiation. At longer postirradiation delays, radiation-induced redistribution in the cell cycle appeared to be the major determinant of HeLa cell survival, in relation to the differential cell cycle phase specificity of each drug. Pronounced paclitaxel recovery versus increased sensitivity to docetaxel occurred over 8 h after irradiation. SQ20B cells showed monophasic radiation recovery with both drugs over the same time range.

S-phase specificity of cell killing by docetaxel (Taxotere) in synchronised HeLa cells.

Cell viability following short (1 h) contact with paclitaxel or docetaxel was assayed using synchronised HeLa cells. Docetaxel proved almost totally lethal against S-phase cells. Its toxicity was only partial against cells in mitosis, and declined to a minimum with progression to G1. For paclitaxel, cytotoxicity increased with progression through S and G2, peaked at the time of mitosis, and decreased thereafter. Maximum resistance to paclitaxel was in early S. Although lethal, brief exposure to docetaxel in S-phase did not delay progression through S and G2. Gross damage was detectable immediately after mitosis, with dysfunction in cytokinesis and accumulation of multinucleated, non-viable cells. Arrest of cells at prometaphase required continuous contact with lethal amounts of docetaxel or reintroduction of drug shortly before mitosis following pulse-chase treatment in mid-S-phase. Paclitaxel at moderate doses presumably acts mostly via damage to the mitotic spindle. In contrast, the available data suggest that docetaxel primarily targets centrosome organisation, leading to abortive mitosis and cell death.

Interaction of ionizing radiation with the topoisomerase I poison camptothecin in growing V-79 and HeLa cells.

Interaction between gamma-rays and camptothecin (CPT) was investigated in vitro, using log phase HeLa-S3 cells and Chinese hamster V-79 fibroblasts. A plateau of toxicity was rapidly reached for both cell lines upon exposure to CPT alone, consistent with S-phase specificity of CPT. With synchronized HeLa cells, however, CPT proved cytotoxic after the middle of G1 phase. This effect was abolished by cotreatment with the DNA polymerase alpha inhibitor aphidicolin. CPT enhanced the initial slope of the radiation survival curves for asynchronous cells by a factor of 1.1 (HeLa) to 2.3 (V-79). This apparent radiation sensitization correlated with the intrinsic radiosensitivities of the drug-surviving fractions within the different compartments of the cell cycle. There was no evidence of mutual potentiation of CPT and radiation in terms of survival as well as with regard to the formation and rejoining of DNA double-strand breaks. V-79 cells exhibited pronounced postirradiation recovery. In contrast, the response of HeLa cells to drug did not vary appreciably for over 8-h following radiation. This difference proceeded from differential cell cycle redistribution. In both cell lines, acute irradiation produced depletion of the G1 compartment, accumulation at the S-G2 junction, and G2 arrest in proportion to the gamma-ray dose. However, while radiation brought about rapid depletion of the S-phase compartment in V-79 cells, it induced accumulation of HeLa cells into S phase. As a result, the CPT-sensitive HeLa cell population, consisting of the early-S, mid-S, and late G1 fractions, did not vary very much after irradiation. Exposure to CPT under conditions of low dose rate irradiation (1 Gy/h) selectively reversed the radiation effect on S-phase progression in V-79 cells; i.e., it induced accumulation of cells in S phase in the same way as found with HeLa cells. Isobolic analysis of survival data consistently showed supraadditivity of cell killing in both cell lines upon concomitant exposure to CPT and low dose rate irradiation. Cytokinetic cooperation appears to be the major determinant of cell survival in treatments associating CPT and radiation in growing cells. Attempts to predict the outcome of such a combined modality thus should take into consideration the response of the growing fraction of each cell line in terms of cell cycle-regulatory processes and redistribution.

DNA repair and cell cycle interactions in radiation sensitization by the topoisomerase II poison etoposide.

The interactions between ionizing radiation and etoposide were probed using asynchronous growing V-79 fibroblasts. Synergistic cell kill was observed as gamma-rays were applied prior to or concomitantly with the drug. Three major determinants of enhanced cytotoxicity in combined treatment were identified. The kinetic analysis of radiation recovery bore evidence of two repair interaction mechanisms. First, rapidly repairable radiation-induced DNA damage was fixed into lethal lesions by etoposide, thus giving rise to marked supra-additive interaction under concomitant radiation-drug exposure. Second, cells arrested in G2 phase following radiation proved hypersensitive to the cytotoxic effect of etoposide. It is proposed that either topoisomerase II alpha is closely involved in some rapid DNA repair pathway operating during all phases of the cell cycle, and even further involved in DNA repair acting within the radiation-induced G2 block, or that the lesions induced by etoposide are able to impair these processes. The shoulder of the radiation survival curve was abolished as gamma-rays and drug were applied at 1-h intervals. This effect, corresponding to mode II additivity from isobologram determinations, appeared to be correlated with a differential sensitivity of the various phases of the cell cycle to drug and radiation.

Additive and supraadditive interaction between ionizing radiation and pazelliptine, a DNA topoisomerase inhibitor, in Chinese hamster V-79 fibroblasts.

The cytotoxic effect of the 9-azaellipticine derivative pazelliptine in combination with gamma-ray irradiation was investigated using Chinese hamster V-79 cells in culture. gamma-ray irradiation and drug treatment (1-h drug exposure) were applied at 1-h intervals for partial DNA damage recovery in growth medium. Isobologram analysis of the clonogenic potential gave evidence of supraadditive interaction in the radiation----drug sequence with 10% survival as an endpoint. No synergistic potentiation was observed at higher survival or as pazelliptine was applied first. Pazelliptine abolished the low-dose shoulder characteristic of asynchronous cell response to gamma-rays. Although rejoining of radiation-induced DNA strand breaks was completed at the time of drug exposure, pazelliptine brought about a larger amount of DNA strand breaks in preirradiated than in nonirradiated cells. The time and dose dependencies of DNA strand break formation and repair with radiation and/or pazelliptine were analyzed by neutral and alkaline filter elution. Pazelliptine in the micromolar range showed the same pattern of double-stranded cleavable complex formation as expected of a DNA topoisomerase II-targeting agent. At a low concentration of pazelliptine, however, protein-concealed breaks were mostly in the form of single-stranded adducts. Such single-stranded complexes have been reported to occur with some topoisomerase II-targeting drugs; their properties are also reminiscent of those induced by the topoisomerase I poison, camptothecin. It is proposed that topoisomerase poisoning interacts with the repair of radiation-induced lesions.

[Restoration of DNA supercoiling in fibroblasts in the rat subjected to gamma ray or fast neutrons]. / Restauration du surenroulement de l'ADN de fibroblastes de rat soumis aux rayons gamma ou aux neutrons rapides.

Nucleoid sedimentation analysis was applied to the study of DNA supercoiling repair in cultured FR 3T3 fibroblasts exposed to low doses of fast neutrons or gamma-rays. Supercoiling was fully restored in both instances upon post-irradiation at 37 degrees C, but the rate of repair of neutron-induced lesions was lower than that for gamma-rays. Non-repairable breaks were not evidenced at the neutral pH used. We suggest that the non-repairable alkali-labile sites evidenced by others in neutron-irradiated DNA do not prevent strand break rejoining and subsequent recovery of the tertiary DNA structure.

Study of several factors in RNA-protein cross-link formation induced by ionizing radiations within 70S ribosomes of E. coli MRE 600.

The induction of RNA-protein crosslinks in E. coli 70S ribosomes by gamma-irradiation was studied by measuring the dependence of cross-link formation on ribosome concentration. The inverse dependence of cross-link percentage upon concentration up to at least 20 A260 nm units ml-1 indicate that indirect effects seem to play a more major part than direct effects for these ribosome concentrations. The effect of various gases and free radical scavengers was used to determine the roles of the radicals H., CO2-., OH. and e-aq and to estimate their relative efficiencies for cross-links. They were found to be: 7.2(H.), 6(CO2-.), 2(OH.) and 1(e-aq). The extent of RNA-protein cross-link production in 70S ribosomes induced by gamma-rays and neutrons in the presence and absence of oxygen was also investigated. Cross-link formation was estimated by separation of linked and unlinked material on nitrocellulose filters or after separation by SDS-sucrose gradient centrifugation under dissociating conditions. Oxygen inhibited cross-link formation by both neutrons and gamma-rays. However, very few cross-links were formed in de-aerated solutions by exposure to neutrons, compared to those produced by gamma-rays under the same conditions. This suggests that molecular oxygen generated along the secondary particle track can reduce the formation of RNA-protein cross-links.

Ribosomal RNA-protein cross-links, induced by gamma-irradiation of 40S and 60S ribosomal subunits of L cells.

The 40S and 60S ribosomal subunits of L cells are gamma-irradiated in the absence of oxygen at low radiation doses to keep the integrity of the ribosomal structure. We show that under these experimental conditions, specific cross-links are induced in situ between rRNA and ribosomal proteins due to close contact between their reactive groups. We found that about 15 proteins are cross-linked to the 28S RNA. Most of them belong to the core proteins of the 60S ribosomal subunits. A few high-molecular mass proteins which might be factors are also bound to 28S RNA. Between 8 and 11 proteins are covalently linked to 18S RNA; 8 of these have been previously described as transferable proteins from one subunit to the other. Only 3 are core proteins of the small subunit. The contribution of these results to the study of the three-dimensional ribosomal structure is also discussed.

Radiochemical cross-linking between proteins and RNA within 70 S ribosomal particles from E. coli MRE600.

In the absence of oxygen, gamma-irradiation produces covalent links between some ribosomal proteins and 16 S RNA to 23 S RNA, within 70 S ribosomes from E. coli MRE600. Under optimal conditions minimizing the structural modifications induced by radiations, in situ formed cross-links appear specific and reflect close RNA-protein contacts. In view of these results, the spatial organization of the 30 S, 50 S subunit interfaces is discussed. In addition, the gamma-irradiation technique reveals that subunit association induces modifications of some protein--RNA interactions.

Radiochemical cross-linking of proteins to RNA within ribosomal subunits from E. colil MRE 600.

Irradiation in vitro of 30S and 50S ribosomal subunits from E. coli MRE 600 by gamma rays from cobalt 60, in the absence of oxygen, results in the formation of covalent links between the RNAs and some ribosomal proteins. At low radiation doses, just sufficient to keep the integrity of the ribosome structure, the phenomenon appears highly specific. In the 30S irradiated particle, protein S1 attaches to 16S RNA; in the 50S irradiated particle, the proteins L3, L13, L19, L21, L22 and L24 are linked to 23S RNA. The mechanism of formation of these cross-links and their contribution to the study of the tridimensional ribosome structure are also discussed.