Integration of proteomics, bioinformatics, and systems biology in traumatic brain injury biomarker discovery.

Traumatic brain injury (TBI) is a major medical crisis without any FDA-approved pharmacological therapies that have been demonstrated to improve functional outcomes. It has been argued that discovery of disease-relevant biomarkers might help to guide successful clinical trials for TBI. Major advances in mass spectrometry (MS) have revolutionized the field of proteomic biomarker discovery and facilitated the identification of several candidate markers that are being further evaluated for their efficacy as TBI biomarkers. However, several hurdles have to be overcome even during the discovery phase which is only the first step in the long process of biomarker development. The high-throughput nature of MS-based proteomic experiments generates a massive amount of mass spectral data presenting great challenges in downstream interpretation. Currently, different bioinformatics platforms are available for functional analysis and data mining of MS-generated proteomic data. These tools provide a way to convert data sets to biologically interpretable results and functional outcomes. A strategy that has promise in advancing biomarker development involves the triad of proteomics, bioinformatics, and systems biology. In this review, a brief overview of how bioinformatics and systems biology tools analyze, transform, and interpret complex MS datasets into biologically relevant results is discussed. In addition, challenges and limitations of proteomics, bioinformatics, and systems biology in TBI biomarker discovery are presented. A brief survey of researches that utilized these three overlapping disciplines in TBI biomarker discovery is also presented. Finally, examples of TBI biomarkers and their applications are discussed.

A selective cysteine protease inhibitor is non-toxic and cerebroprotective in rats undergoing transient middle cerebral artery ischemia.

Ischemic neuronal injury mediated by cysteine proteases such as calpains and caspases has been demonstrated in various experimental models. Cathepsins B and L are also cysteine proteases which may contribute to neuronal death after ischemia. The authors measured in vitro and in vivo toxicity and post-ischemic cytoprotective effects of a cysteine protease inhibitor which does not block calpain or caspase but, rather, is relatively selective for cathepsins B and L. The compound belongs to the peptidyl-diazomethane family (cysteine protease inhibitor 1, termed CP-1). In vitro toxicity was measured using an assay of cell viability, and in vivo toxicity was measured by histological tissue analysis after infusion of CP-1 in rats. Two hours of middle cerebral artery (MCA) occlusion in rats was performed by the intravascular suture method. Immediately following reperfusion, intravenous infusion of CP-1 or vehicle was performed for 4 h at 0.9 ml/h. After a 7-day survival, the infarct volumes were measured. CP-1 was non-toxic to cultured glial cells to a local concentration of 200 microM, and relatively non-toxic to cultured endothelial cells at concentrations of 100-200 microM. No animal exhibited toxic effects at any of the doses used. Histologic comparisons revealed no signs of tissue toxicity. CP-1 significantly reduced hemispheric infarct volume compared to control (37+/-8.2%) at concentrations of 10, 50, and 250 microM [22+/-15%, P=0.008; 20+/-13%, P=0.002; 23+/-15%, P=0.022, respectively (mean+/-standard deviation; N=7-10 per group)]. CP-1, at the concentration of 50 microM, improved the functional score of the animals, but did not significantly alter cerebral blood flow. This study supports the hypothesis that the lysosomal cathepsins B and/or L contribute to cerebral injury after focal ischemia with reperfusion. Cysteine protease inhibitors which are relatively selective for cathepsins B and L, but not the calpains or caspases, are effective at reducing infarct volume after intravenous post-ischemic administration.

Functional properties of recombinant calpain I and of mutants lacking domains III and IV of the catalytic subunit.

The catalytic subunit (L-microCANP) of human calpain I (muCANP, the high Ca2+ affinity form) and two of its mutants were expressed in Escherichia coli or using the baculovirus Sf9 system. The mutants lacked domain III (L-mu CANPDelta3) and the calmodulin-like domain IV (L-mu CANPDelta4), respectively. The bacterially expressed proteins were solubilized from the inclusion bodies and refolded with polyethylene glycol. In Sf9 cells, co-expression of the inhibitor calpastatin was necessary to prevent autolysis of L-muCANP, whereas co-expression of the regulatory subunit enhanced it. Only very low levels of mRNA of the truncated form L-mu CANPDelta4 were found in bacmid-transfected Sf9 cells, and it proved impossible to isolate this mutant using the baculovirus expression system. While the apparent Km(Ca2+) of freshly isolated human erythrocyte muCANP was about 60 microM, the recombinant monomeric forms L-mu CANP and L-mu CANPDelta3 required 65-215 and 400-530 microM Ca2+, respectively. Bacterially expressed L-mu CANPDelta4 was Ca2+-independent; the presence of inhibitors during its renaturation was necessary to prevent its autolysis. A chimeric form (L-mu mCANP) composed by domains I-III of muCANP and domain IV of calpain II (mCANP, the low Ca2+ affinity form) was also expressed in Sf9 cells. This mutant required less Ca2+ (about 50 microM) than native erythrocyte calpain for half-maximal activity and had the highest specific activity of all calpains tested. Domain III proved unnecessary for the activity of the recombinant catalytic subunit, but its absence raised the Km(Ca2+) and removed its inactivation at high Ca2+ concentrations. All recombinant proteins were active as monomers in polyethylene glycol-containing buffers; the in vitro association with the regulatory subunit enhanced only slightly the Vmax and the Ca2+ dependence of the expressed proteins. Activation by Ca2+ promoted the separation of the two subunits of the expressed recombinant proteins.

Purification of active calpain by affinity chromatography on an immobilized peptide inhibitor.

Most purification schemes of calpain (CANP) involve a number of chromatographic steps. The final preparations often contain impurities, including degradation fragments. Two peptide-affinity columns were developed, using peptides of 27 amino acids and 30 amino acids, corresponding to the products of exons 1B and 1C, respectively, of the natural inhibitor (calpastatin) gene, coupled to CNBr-activated Sepharose 4B. Crude preparations of calpain, isolated by anion-exchange chromatography on a DEAE-Sepharose column, were incubated with a reversible or an irreversible synthetic inhibitor which blocks the catalytic subunit of the enzyme in the inactive 80-kDa form. The crude preparation was then loaded onto the peptide column in the presence of calcium. Calpain was eluted with an EGTA-containing buffer. Using the two peptide-affinity columns connected in tandem, calpain was isolated with a high degree of purity, suitable for structural and mechanistic studies, i.e. as an 80/30-kDa heterodimer or in the form of dissociated monomers.

Calmodulin binds to and inhibits GTP binding of the ras-like GTPase Kir/Gem.

Recently, a new subfamily of Ras-related GTP-binding proteins consisting of Rad (Ras associated with diabetes), Gem (immediate early gene expressed in mitogen-stimulated T-cells), and Kir (tyrosine kinase-inducible Ras-like) was discovered. The C terminus of these proteins contains an extension of approximately 30 amino acids not present in other members of the Ras family and which exhibits all the hallmarks typical for calmodulin (CaM)-binding domains. A peptide corresponding to the putative CaM-binding domain of the Kir/Gem protein was synthesized, and its affinity for CaM was determined by fluorescence spectrometry. Titration of dansyl-CaM with the Kir/Gem peptide gave an affinity constant of 1 nM. Furthermore, a single point mutation of the peptide, W269G, abolished this high affinity interaction. Gel-shift analysis showed that the complex formation between CaM and the Kir/Gem peptide is strictly calcium-dependent. We also demonstrate with a newly developed [32P]CaM overlay technique that full-length Kir/Gem and Rad proteins bind CaM in a Ca2+-dependent fashion. The binding of CaM to glutathione S-transferase-Kir and GST-Gem inhibited the binding of GTP to Kir/Gem significantly. These results suggest the existence of a direct link between Ca2+/CaM and growth factor signal transduction pathways at the level of small Ras-like GTPases.

The secondary structure of phospholamban: a two-dimensional NMR study.

Phospholamban (PLN) is an intrinsic membrane protein of 52 amino acids which regulates the Ca2+ pump of the sarcoplasmic reticulum of heart, slow-twitch and smooth muscle (SR): it is normally assumed to exist in the membrane as a homopentamer. A monomeric analogue of phospholamban PLN(C41F), in which Cys41 was replaced by a Phe, was synthesized and its conformation studied by 1H NMR spectroscopy in a 1:1 mixture of chloroform/methanol. Most of the resonances in the 1H NMR spectra were assigned. The work has shown that the C-terminal hydrophobic portion forms a very stable alpha-helix. The hydrophilic N-terminal part adopts an alpha-helix configuration which is much less stable except for the stretch containing the phosphorylation sites.

The calmodulin-binding domain of the inducible (macrophage) nitric oxide synthase.

A domain in the inducible, macrophage nitric oxide (NO) synthase has been selected as the putative calmodulin-binding site. The domain was synthesized as a peptide of 29 residues [P29, NO synthase-(504-532)-peptide], having the accepted hydrophobic/basic composition of calmodulin-binding domains and containing, like most of them, an aromatic amino acid at its N-terminus and a long chain aliphatic residue 12 amino acids downstream of it. A 34-residue peptide from the synthase sequence [P34, NO synthase-(499-532)-peptide], consisting of peptide P29 and of the five extra N-terminal amino acids, three of them basic, was also synthesized. Both peptides bound calmodulin in the presence as well as in the absence of Ca2+ (i.e. in the presence of excess EGTA). The KD of the binding in the presence of Ca2+ was < or = 1 nM. The binding affinity was lower, but still remarkably high in the presence of EGTA. The peptides counteracted the stimulation by calmodulin of a classical calmodulin-target enzyme, the Ca2+ pump of the plasma membrane.

PEST sequences do not influence substrate susceptibility to calpain proteolysis.

Mutations lowering the PEST score of domains surrounding the calmodulin (CaM)-binding region of the plasma membrane Ca(2+)-ATPase failed to influence the susceptibility of the enzyme to mu-calpain (mu-CANP). Synthetic peptides corresponding to the high PEST score C-terminal sequences A18 and B28 had no effect on the rate of pump proteolysis by mu-CANP, i.e. the peptides did not compete for a putative high PEST score recognition site for mu-CANP in the pump molecule. An accessible CaM-binding region appears to be critical for substrate (i.e. the Ca2+ pump) proteolysis and probably also for its recognition by mu-CANP; phosphorylation of the CaM-binding domain of the pump or its occupation by CaM significantly decreased the rate of proteolysis.

Ca(2+)-activated neutral protease is active in the erythrocyte membrane in its nonautolyzed 80-kDa form.

The aim of this study was to investigate the process leading to Ca(2+)-activated neutral protease (CANP) activation in vivo. The unautolyzed form of CANP has been targeted to the erythrocyte membrane by increasing, in a controlled way, the Ca2+ concentration in the cells; this was achieved by incubating erythrocytes with the Ca2+ ionophore A23187 and fixed Ca2+ concentrations. After isolation of the CANP-bearing erythrocyte membrane, we could observe that CANP remained bound to the membrane in the 80-kDa unautolyzed form in the presence of low Ca2+ concentrations (1.75 microM); under these conditions, the preferred CANP substrates (the Ca(2+)-ATPase and Band 3) were cleaved. That the cleavage was due to CANP was shown by the finding that the two substrates were not degraded in the presence of a membrane-permeable irreversible CANP inhibitor, Cbz-Leu-Leu-Tyr-CHN2, nor when the free Ca2+ concentration was decreased to sub microM levels with EDTA. The findings suggest an activation mechanism of CANP based on its translocation to the membrane rather than on its autolysis. In this mechanism, CANP would become reversibly activated on the membrane and would return to the quiescent state after dissociating from it when the cell Ca2+ concentration has returned to the physiological, submicromolar level.

Phosphorylation of the calmodulin binding domain of the plasma membrane Ca2+ pump by protein kinase C reduces its interaction with calmodulin and with its pump receptor site.

Two versions of the calmodulin binding domain of the plasma membrane Ca2+ ATPase, a 24-amino acid peptide, C24W (Q-I-L-W-F-R-G-L-N-R-I-Q-T-Q-I-R-V-V-N-A-F-R-S-S-NH2), and the corresponding phosphothreonine containing peptide, C24W-P (Q-I-L-W-F-R-G-L-N-R-I-Q-T(phospho)-Q-I-R-V-V-N-A-F-R-S-S-NH2), were synthesized. They were used to investigate the effect of threonine phosphorylation by protein kinase C on the binding of calmodulin by the calmodulin binding domain and on the inhibitory role of the domain on the activity of the Ca2+ pump. The phosphopeptide C24W-P was obtained after global phosphorylation of the free Thr side chain on the protected resin bound peptide. The phosphorylated calmodulin binding domain failed to bind calmodulin; this was shown by gel shift experiments, by fluorescence energy transfer studies and by competition experiments against calmodulin stimulation of the pump. The inhibition of the Ca2+ pump activity by the calmodulin binding domain in the absence of calmodulin was also affected by the phosphorylation of the threonine; the inhibition of the fully active calpain-truncated pump by the phosphothreonine containing peptide was lower than that by the unphosphorylated synthetic domain.

Affinity labelling of the Ca(2+)-activated neutral proteinase (calpain) in intact human platelets.

Two irreversible calpain inhibitors, benzyloxycarbonyl (Cbz)-Leu-Leu-Tyr-Ch2F and Cbz-Leu-Leu-Tyr-CHN2, were shown earlier [Anagli, Hagmann and Shaw (1991) Biochem. J. 274, 497-502] to penetrate intact platelets and to inactivate calpain. This permitted an evaluation of certain functions attributed to this proteinase. For example, in platelets pretreated with these inhibitors, talin and actin-binding protein were protected from subsequent degradation when the Ca2+ level was raised. On the other hand, additional properties of stimulated platelets attributed to calpain remained unaffected by this treatment, and such hypotheses may be dismissed. Radioiodinated inhibitors permitted confirmation of the labelling of calpain by the procedures used. Although Cbz-Leu-Leu-Tyr-CHN2 is more effective in vitro than the corresponding fluoromethyl ketone, we now show that the latter penetrates more readily. These two inhibitors, and two additional ones, t-butyloxycarbonyl-Val-Lys(Cbz)-Leu-Tyr- CHN2 and Cbz-Leu-Tyr-CH2F, have been radioiodinated to permit a comparison of their intracellular labelling patterns in activated platelets. Calpain is the major target of all four inhibitors. Although they are closely related peptide structures, variations with respect to the labelling of additional proteins were observed. These were minor in the case of the peptidyl diazomethyl ketones, but were major in the case of the fluoromethyl ketones. However, in contrast to calpain, this labelling was neither time-dependent nor Ca(2+)-dependent. Radiolabelling and cellular fractionation studies were used to localize active calpain during platelet activation. Calpain appears to be activated in the cytosol and translocated to the membrane or cytoskeletal sites.

Additional peptidyl diazomethyl ketones, including biotinyl derivatives, which affinity-label calpain and related cysteinyl proteinases.

Calpain, the calcium-activated cysteinyl proteinase, can be irreversibly inactivated by peptidyl diazomethyl ketones in which the peptide portion contains a penultimate leucine residue. Some new derivatives of this type have been synthesized and examined for their rates of inactivation of chicken gizzard and human platelet calpain. Two derivatives containing a C-terminal biotin residue, Biot-Aca-Leu-TyrCHN2 and Biot-Aca-Leu-Leu-TyrCHN2, have also been prepared in the expectation that their application to the study of the function of calpain and related proteases will prove fruitful.

Inactivation of calpain by peptidyl fluoromethyl ketones.

The syntheses of Z-Leu-Leu-Tyr-CH2F (1) and Z-Tyr-Ala-CH2F (3) are described. The ability of Z-Leu-Leu-Tyr-CH2F (1) and Z-Leu-Tyr-CH2F (2) to inactivate in vitro calcium-activated proteinase from chicken gizzard are compared. Like the analogous diazomethyl ketones 4 and 5, these inhibitors were also found to inactivate cathepsin L in common with other inhibitors under current investigation. However, other specific inactivators for cathepsin L are available, for example, the fluoromethyl ketone 3 and diazomethyl ketone 6 of Z-Tyr-Ala-OH, which have no effect on the calcium-activated proteinase and therefore provide control inhibitors for observations made with Z-Leu-Leu-Tyr-CH2F (1).

Investigation of the role of calpain as a stimulus-response mediator in human platelets using new synthetic inhibitors.

A series of peptidyl diazomethanes and monofluoromethane with structures specific for calpain have been synthesized and tested for their ability to inhibit calpain activity in vivo, using human platelets as a model system. Calpain activity in vivo was determined by observing proteolysis of actin-binding protein and talin, two known substrates of calpain. Very potent inhibitors, which emerged from this study, were used to investigate the role of calpain in some platelet response processes. Our results show that calpain-mediated proteolysis in platelets is not an obligatory event leading to change of cell shape, adhesion to glass and spreading, aggregation and 5-hydroxytryptamine release. Two of the inhibitors were iodinated with 125I and used to radiolabel the enzyme in vivo. To our knowledge, this work also represents the first report describing the affinity labelling of calpain in human platelets using irreversible radioactive inhibitors.