PhD (2005-2009) Thèse dirigée par S. Dukan (Université Aix-Marseille II)
Title of the thesis :
"Agrégation et carbonylation des protéines : deux facteurs impliqués dans le vieillissement cellulaire d’Escherichia Coli"(2009)
We observed that in E. coli, more than 95% of the total carbonyl content consisted of insoluble protein and most were cytosolic proteins. We thereby demonstrate that, in vivo, carbonylated proteins are detectable mainly in an aggregate state. Finally, we show that detectable carbonylated proteins are not degraded in vivo like it was previously proposed by Dukan et al., (2000). Here we propose that some carbonylated proteins escape degradation in vivo by forming carbonylated protein aggregates and thus becoming nondegradable. In light of these findings, we provide evidence that the accumulation of nondegradable carbonylated protein presented in an aggregate state contributes to the increases in carbonyl content observed in VBNC cells (Maisonneuve et al., 2008c).
Specific protein carbonylation of SN30, LP, and SP from an exponentially grown E. coli culture. Extracts from a French press CE of an exponentially grown culture of E. coli (OD600= 0.5) obtained after various centrifugation times were processed for resolution on 2D polyacrylamide. Autoradiograms were obtained after carbonyl immunoassay of proteins. Molecular masses (M) in kDa are indicated on the left.
Diagram shows the separation of the pellets and supernatant
Carbonyl derivatives are mainly formed by direct metal-catalysed oxidation (MCO) attacks on the amino-acid side chains of proline, arginine, lysine and threonine residues. For reasons unknown, only some proteins are prone to carbonylation. We used mass spectrometry analysis to identify carbonylated sites in : BSA that had undergone in vitro MCO, and 23 carbonylated proteins in E. coli. The presence of a carbonylated site rendered the neighbouring carbonylatable site more prone to carbonylation. Most carbonylated sites were present within hot spots of carbonylation. These observations led us to suggest rules for identifying sites more prone to carbonylation. We used these rules to design an in silico model (available at http://www.lcb.cnrs-mrs.fr/CSPD/), allowing an effective and accurate prediction of sites and of proteins more prone to carbonylation in the E. coli proteome. We observed that proteins evolve to either selectively maintain or lose predicted hot spots of carbonylation depending on their biological function. As our predictive model also allows efficient detection of carbonylated proteins in Bacillus subtilis, we believe that our model may be extended to direct MCO attacks in all organisms (Maisonneuve et al., 2009).
The details of the principle of detection are described in Materials and Methods section (Maisonneuve et al., 2009). A predicted HSC (grey box) is defined by two regions. A) An RKPT-enriched region (3 carbonylatable residues within a sequence of 4 amino acids (R, K, P, T ; 3 ; 4) containing at least one proline (P ; 1 ; 0). B) A specific environment around an RKPT-enriched region, enriched in various residues : (i) iron binding sites (D, E, Y, H, C, namely 1 residue within a window of 2 residues (D, E, Y, H, C ; 1 ; 2) and 8 residues within a window of 29 residues (D, E, Y, H, C ; 8 ; 29) ; (ii) hydrophobic amino acids (A, V, G, I, namely 1 residue within a window of 2 (A, V, G, I ; 1 ; 2)) ; (G namely 2 residues within a window of 14 (G ; 2 ; 14)) ; and (iii) (P, T) with 2 residues occurring within a window of 21 residues (P, T ; 2 ; 21). E, environment ; r, right ; l, left and w, window.
CO2 modulate oxidative stress
During these last four years we have developed a new axis of research dealing with CO2 impact on oxidative stress. This new axis was based on a recent paper from I. Fridovich lab’s indicating a link between Reactive oxygen species (ROS) and CO2 in vitro (Liochev and Fridovich, 2004). For this purpose we have designed and developed (in relation with Jacomex Company) a chamber allowing us to control the atmospheric CO2 level within it. Benjamin Ezraty (CR2) is in charge of this project since 2007 (in association with Maïalène Chabalier (tech.)) and he is responsible for major advances of this topic. ROS, which encompass the superoxide anion (O2•-), hydrogen peroxide (H2O2) and the hydroxyl radical (HO•), are harmful as they can oxidize all biological macromolecules (Finkel and Holbrook, 2000). in vitro metal-catalyzed reactions between CO2 and ROS can generate the harmful carbonate radical (CO3•-) (Liochev and Fridovich, 2004). In vivo, CO2 is both a major by-product of metabolism and the major pH buffer system in higher eukaryotes. CO2 is also required for the growth of many microorganisms (Walker, 1932). We thus tested whether atmospheric CO2 (current value 389 ppm) could exacerbate oxidative stress. We used E. coli as model organism to evaluate whether atmospheric CO2 influenced oxidative stress. We established that the minimal atmospheric CO2 concentration required for optimal growth was 40 ppm. We show that atmospheric CO2 (range studied : 40 to 1,000 ppm) increases the death of Escherichia coli caused by peroxide stress in a dose-specific fashion. This effect correlates with increases in H2O2-induced mutagenesis rates and DNA bases oxidation as measured by the amount of 8-oxo-guanine in the cell. Moreover, survival of mutants sensitive to aerobic growth (Hpx- dps and recA fur) (Park and Imlay, 2005 ; Touati et al., 1995), presumably due to their inability to tolerate oxygen-derived ROS, appear to be CO2 level-dependent (range studied : 40 to 1,000 ppm). The aerobic viability defect of these strains has been attributed to DNA damage caused by a Fenton reaction-based hydroxyl radical (HO•) production (Finkel and Holbrook, 2000 ; Park and Imlay, 2005 ; Touati et al., 1995). The higher oxygen toxicity at higher CO2 concentrations thus indicates that CO2 exacerbates HO•–induced DNA damage. Altogether these results indicate that atmospheric CO2 exacerbates ROS toxicity by increasing oxidative cellular lesions. This study provides the first evidence that oxidative stress is exacerbated by atmospheric CO2. CO2 have been a major point of focus, due to their contribution to the greenhouse effect (Cox et al., 2000), and increases in these levels are thought to be associated with global warming. In the light of the Special Report on Emissions Scenario, predicting that the atmosphere in 2100 will contain 1,000 ppm CO2 (Schneider, 2009) our results suggest that increases in atmospheric CO2 concentration have toxic effects over and above those of global warming (Ezraty et al., submitted).
2-E. coli recovery on plate, oxidative stress and VBNC phenomenon
Protein aggregation and recovery on plate
Protein aggregation is a phenomenon observed in all organisms and has often been linked with cell disorders. We demonstrated the presence of protein aggregates in an exponentially grown E. coli culture. Our results led us to speculate that protein aggregates may function as a temporary “trash organelle” for cellular detoxification (Maisonneuve et al., 2008a).
Silver-stained 2D gel analysis of proteins in pellets and supernatant. Panels show protein patterns in LP, SP, ILP, ISP, and supernatant obtained after 2D gel electrophoresis and visualized by silver staining.
In light of these observations, protein aggregates could be considered damage to cells that is able to pass from one generation to the next. Based on the assumption that the amount of aggregate protein could represent an aging factor, we monitored this amount in a bacterial culture during senescence. In doing so, we observed (i) a significant increase in the amount of aggregate protein over time,
Aggregate protein accumulates over time. (A) Coomassie-stained SDS/PAGE gel showing the relative amount of aggregate protein (per mg of soluble protein) at three time points during exponential (2, 4 and 6 hours) or stationary phase (10, 24 and 48 hours). At each point, samples were prepared from equal OD amounts of cells. Panel (B) shows representative results from relative aggregate content for each time during exponential (2, 4 and 6 hours) or stationary phase (10, 24, 30, 36 and 48 hours) quantified using Quantity One software (BioRad).
(ii) a proportional relationship between the amount of aggregate protein and the level of VBNC cells, (iii) a larger amount in VBNC cells than in culturable cells, (iv) a heterogeneous distribution of different amounts within a homogenous population of culturable cells entering stasis, and (v) that the initial amount of aggregate protein within a culturable population conditioned the VBNC cell rate of the culture. Together, the results presented in this study suggest that protein aggregates represent one aging factor leading to VBNC cell formation (Maisonneuve et al., 2008b). For this study, we got a highlight in Microbes (nov. 2008).
(A) Determination of reproductive ability (CFU) and cell integrity in the Low Density (LD, shaded bar) and High Density (HD, unshaded bar) cell populations from a 48 hours stationary-phase culture. (B) Coomassie-stained SDS/PAGE gel showing the relative amount of aggregate protein (per mg of soluble protein) from LD and HD cells prepared from an equal OD amount of cells. © Relative aggregate amount from LD an HD cells quantified using Quantity One software (BioRad). (D) Determination of reproductive ability (CFU) and cell integrity in the Low Density (LD, shaded bar) and High Density (HD, unshaded bar) cell populations from a 10 hours culture. (E) Coomassie-stained SDS/PAGE gel showing the relative amount of aggregate protein (per mg of soluble protein) from LD and HD cells prepared from an equal OD amount of cells. (F) Relative aggregate amount from LD an HD cells quantified using Quantity One software (BioRad).
Most time lapse microscopy experiments studying bacterial processes ie growth, progression through the cell cycle and motility has been performed on thin nutrient agar pads. An important limitation of this approach is that dynamic perturbations of the experimental conditions cannot be easily performed. In eukaryotic cell biology, fluidic approaches have been largely used to study the impact of rapid environmental perturbations on live cells and in real time. However, all these approaches are not easily applicable to bacterial cells because the substrata are in all cases specific and also because microfluidics nanotechnology requires a complex lithography for the study of micrometer sized bacterial cells. In fact, in many cases agar is the experimental solid substratum on which bacteria can move or even grow. For these reasons, we designed a novel hybrid micro fluidic device that combines a thin agar pad and a custom flow chamber. By studying several examples, we show that this system allows real time analysis of a broad array of biological processes such as growth, development and motility. Thus, the flow chamber system will be an essential tool to study any process that take place on an agar surface at the single cell level (Ducret et al., 2009).
Layout of the microfluidic device. Cells are confined between the coverslip and a 0.5 mm thin layer of agar. The chamber is sealed with a transparent lid containing two entries allowing flow injections. Injected molecules reach the biological specimen by diffusion through the thin layer of agar.
A) Picture of the assembly of the hybrid flow chamber. The cells are placed immediately on a coverslip and overlaid with a 0.5 mm thin layer of agar (violet). The chamber is sealed with a transparent lid containing two entries allowing flow injections ; (B) sketch of the fully automatized flow networks. The 12 valves are connected to a bubble detector that allows the triggering of 2 additional valves that isolate the flow chamber in the event of air bubbles in the flow network. The aspiration of different solution is made by a syringe pump (SP) which automatically empty into a waste (W) ; C) Overall principle of the setup. The whole setup is driven by a custom-made Visual Basic application run on a PC computer, through many RS232 interface (except the camera which has FireWire interface). The microscope (1), the control unit for the electric valves (2), the syringe pump (3), the camera (4), and the motorized stage can be remotely controlled by our custom program.
- Chatterjee I., Maisonneuve E., Ezraty B., Herrmann M. and Dukan S. (2011) Staphylococcus aureus ClpC is involved in protection of carbon-metabolizing enzymes from carbonylation during stationary growth phase. IJMM 301(4):341-6