Figure 1.
Characterization of the anti-ADC antibody.
A ELISA assay displaying anti-ADC activity. In a competitive ELISA assay preincubation with ADC-GST and ADC-His fusion proteins at increasing concentrations competitively inhibited the specific interaction, demonstrating the reactivity of the antibody against ADC fusion protein. In contrast, pre-adsorption using the unrelated fusion protein (Arg1-GST) or GST did not show any effect on immunoreactivity. The inset shows a direct ELISA, illustrating the activity of the anti-ADC antibody against the immunogen ADC as compared to a potentially cross-reacting ODC partial fusion protein. At the same dilution as used for the competitive assay (1∶30.000), no cross-reactivity was observed. B-C Characterization of the anti-ADC antibody by Western blotting. B: Strong protein bands at 55 kDa were detected in rat liver (lane 1 and 6) as well as in rat prostate tissue (lane 3). In rat brain (lane 2), at the same molecular weight a faint but distinct band was observed. Moreover, the antibody strongly reacted with the bacterially expressed partial fusion proteins ADC-GST (36,7 kDa, lane 4) as well as ADC-His (30,7 kDa, lane 5). C: Pre-absorption of the antibody with 10 µg/ml ADC-His purified fusion protein clearly attenuated the intensity of the fusion protein bands (lane 7, 8) and led to a complete disappearance of the ADC signal in rat liver homogenate (lane 9), thus demonstrating the specificity of the antibody. Loading: Lane 1, rat liver homogenate (25 µg); lane 2, rat total brain homogenate (50 µg); lane 3, rat prostate homogenate (25 µg); lane 4 and 7, ADC-GST fusion protein (25 ng); lane 5 and 8, ADC-His fusion protein; lane 6 and 9, rat liver homogenate (25 µg).
Figure 2.
The anti-ADC antibody does not cross-react with ODC.
The partial fusion proteins ADC-GST (lane 1) as well as ADC-His (lane 2) are clearly recognized by the antibody raised against ADC. However, the antibody does not cross-react with ODC-transfected 293T cell lysate (lane 3). For control, ODC-transfected 293T cell lysate was analyzed in parallel with a commercial ODC antibody (Santa Cruz, sc-33539), resulting in a strong Western signal at the respective molecular weight (lane 6). Moreover, the ADC partial fusion proteins are not recognized by the ODC antibody (lanes 4, 5). In rat liver and prostate homogenates (lanes 7–10), protein bands of slightly different molecular weights, corresponding to the calculated molecular weights of 49 kDa (ADC) and 51 kDa (ODC), respectively, were either detected by ADC or ODC antibodies, again verifying the specificity of the ADC antibody for ADC protein. Loading: lane 1 and 4, ADC-GST fusion protein (10 ng); lane 2 and 5, ADC-His fusion protein (10 ng); lane 3 and 6, ODC (h) 293T lysate (5 µg, Santa Cruz); lane 7 and 9, rat liver homogenate (25 µg); lane 8 and 10, rat prostate homogenate (25 µg).
Figure 3.
Characterization of the anti-Arg1 antibody.
A ELISA assay displaying anti-Arg1 activity Competitive ELISA assay demonstrating the ability of different protein constructs to interfere with the anti-Arg1 immunosignal. Arg1-GST and Arg1-His at increasing concentrations competitively inhibited the specific interaction, demonstrating the antibodies reactivity against Arg1 fusion protein. In contrast, pre-adsorption using either Agm-His or -GST constructs did not show any effect on reactivity. B–C Characterization of the anti-Arg1 antibody by Western blotting. B: The anti-Arg1 antibody detected the bacterially expressed partial fusion proteins Arg1-GST (lane 1, 49,2 kDa) as well as Arg1-His (lane 2, 43,2 kDa). A strong 37kDa band was detectable in rat liver homogenate, a well known source rich of Arg1 (lane 3). C: In rat cortex homogenate (lane 7) and cytosol (lane 8) a band of about 80 kDa was detected. Pre-adsorption of the antibody with Arg1-His purified fusion protein (10 µg/ml lane 4–6, 40 µg/ml lane 9–10) lead to the disappearance of the observed bands, except for lane 4, which was loaded with Arg1-GST fusion protein. This reactivity could be attributed to residual anti-GST activity by testing the antibody against unconjugated GST protein (not shown). Loading: lane 1 and 4, Arg1-GST fusion protein (25 ng); lane 2 and 5, Arg1-His fusion protein (25 ng); lane 3 and 6, rat liver homogenate (25 µg); lane 7 and 9, rat cerebral cortex homogenate (40 µg); lane 8 and 10, rat cerebral cortex cytosol (40 µg).
Figure 4.
In the cerebral cortex, both, Arg-like and ADC-like immunoreactivity were predominantly localized to neuronal cell bodies and main dendrites of all neuronal cell types (A, B), but were also detected in the neuropil. Here, labelling intensity varied with respect to different cortical layers. With both antibodies, cortical layers II/III were comparatively most prominently labelled (arrowheads in A, B). In contrast to neurons, immunoreactivity in glial cells was less pronounced, but still clearly evident in white matter (asterisk) oligodendrocytes (also compare Fig. 3 H). Within neuronal cell bodies, Arg-like immunoreactivity mostly displayed a punctate labelling pattern suggesting a localization to subcellular compartments roughly measuring 0,5 µm in diameter (inset in A). By contrast, ADC-like immunoreactivity was strongly but diffusely distributed throughout the cytoplasm (inset in B). For comparison, the downstream enzymes of the classical and alternative pathway for polyamine synthesis, SpdS (C) and Agm (D) clearly differed with respect to neuronal labelling, since here subpopulations of interneurons were most strongly labelled (arrows in C, D; black arrowhead in inset in D). By contrast, with anti-SpdS-antibodies, principal neurons were not clearly delineated from the surrounding neuropil. With anti-Agm antibodies, however, principal neurons like pyramidal cells showed an intermediate level of labelling intensity when compared to interneurons and neuropil (inset in D; white arrowheads in D). Scale bars represent 400 µm in A–D, 10 µm in insets in A–D.
Figure 5.
In the hippocampal formation, similarly to the cerebral cortex, immunoreactivity for Arg (A, B, C) and ADC (F, G, H) was generally present in principal neurons (A, B, F), many interneurons (A, C, F, G), and stacks of white matter oligodendrocytes (H). Similarly, the neuropil was also differentially labelled in the dentate gyrus (DG, asterisks in A) and ammons horn (CA, white arrowheads in F). With this respect, Spds and Agm also displayed different labelling intensities (D, I black arrowheads in I and white asterisks in J). Cell body labelling for SpdS and Agm was most pronounced in interneurons (compare Fig. 2 C, D). With anti-Arg labelling most principal neurons and interneurons displayed the typical cytosolic punctate labelling pattern, in contrast to ADC labelling. The boxed areas in A, D, F, I are shown at higher magnifications in B, E, G, J, respectively. ml = molecular layer, gl = granule cell layer. Scale bars represent 400 µm in A, D, F, I; 100 µm in E; 50 µm in C,D; 20 µm in G; 10 µm in H.
Figure 6.
Pre-absorption of affinity-purified anti-Arg1-antibodies with Arg2- fusion protein (C-terminal 223 amino acids).
To verify the potential cross reactivity between the two isoforms Arg1 and Arg2, the anti-Arg1 antibody was incubated with Arg2-recombinant protein-loaded nitrocellulose membranes and subsequently eluated (A). The resulting fractions, supernatant (C) and eluate (D), were compared with the original affinity-purified reagent (presumed “pan-arginase”-antibody; B). In the hippocampal CA1 region, two populations of immunoreactive cells became evident. The Arg II membrane-eluated fraction labelled the majority of neuronal cell bodies in both, principal neurons and interneurons (D) with interneurons located mostly outside the pyramidal cell layer. In contrast, the presumed Arg I-representing supernatant predominantly displayed a separate population of putative interneurons (C), several of them among CA1 pyramidal cells. In contrast to the typical punctate labelling pattern as observed with the eluate (inset in D), the scattered interneurons displayed in the inset in (C) were strongly but diffusely labelled. so = stratum oriens; pl = pyramidal cell layer; sr = stratum radiatum. Scale bars represent 100 µm in B–D; 10 µm in insets.
Figure 7.
A: Arg-like immunoreactivity was observed in neurons and neuropil in all cerebellar cortical layers but was most obvious in cell bodies (white arrowheads) and processes (black arrowheads) of Bergmann glial cells. GABAergic interneurons including basket, stellate, and Golgi cells were also observed displaying a prominent cytosolic punctate labelling pattern, whereas Purkinje and granule cells displayed considerably less label. B, C: By contrast, ADC-like immunoreactivity was most prominent in cell bodies and branched dendritic trees (black arrow) of Purkinje cells and inhibitory interneurons in the molecular and granule cell layer. However, Bergmann glial cell bodies (white arrowhead) and processes (black arrowhead) and granule cells (white arrow) also were positive. With increasing antibody concentrations (compare 1∶5000 in B and 1∶1000 in C), the diffuse immunoreactivity in the neuropil became clearly evident. D: With SpdS labelling, cell bodies were hardly delineated. Instead, the neuropil of the molecular layer was strongly but diffusely immunopositive, while in the granule layer numerous giant mossy fiber-like boutons (arrowheads) were detected. E: Agm labelling was also markedly expressed in the molecular layer but most prominently observed in Purkinje cell bodies and denritic processes. Comparatively strong labelling was evident in the regionally occuring unipolar brush cells (arrows). ml = molecular layer, pl = Purkinje cell layer, gl = granule cell layer. Scale bar represents 50 µm.
Figure 8.
Basal forebrain at the level of the anterior commissure.
In this area containing the nucleus accumbens, the “classical” and “alternative” pathway enzymes Arg/SpdS (A, C) and ADC/Agm (B, D), respectively, clearly differed with respect to the homogeneity of labelling in the neuropil. With anti-Arg and anti-SpdS antibodies, patches of varying labelling intensity became evident in the Acb (arrowheads in A; double arrowheads in C) and striatum (asterisks in A, C). By contrast, ADC- and Agm-like immunoreactivity was comparatively homogeneous in these neuropil regions, most obviously displaying neuronal cell bodies. However, in the dorsal accumbens shell area, Agm labelling was pronounced in the neuropil (asterisk in D) and surrounded by stronger staining for Arg- and SpdS-like immunoreactivities. Note the prominent ADC-like immunoreactivity in the neighbouring major island of Calleja (arrow in B) resulting from the high density of labelled cell bodies in this area. LV = lateral ventricle ac = anterior commissure. Scale bar represents 400 µm.
Figure 9.
In the brain stem, the overall labelling patterns were very similar for all tested antibodies. In survey micrographs (A–D), several brain stem nuclei like the medial vestibular nucleus (MeV), the nucleus prepositus hypoglossi (Pr), and the inferior olive (IO) appear clearly pronounced. In addition to neuropil labelling in these areas, with anti- Arg and anti-ADC antibodies, numerous neuronal cell bodies were also evident (E, F) in the MeV. By contrast, in the Pr, cell body labelling was evident with all antibodies, though less pronounced with anti-SpdS antibodies. Scale bars represent 1000 µm in A–D; 200 µm in E–H.
Figure 10.
Virtual pre-embedding and electron microscopy.
To obtain more information about the abundant diffuse neuropil labelling as observed with standard immunoperoxidase/DAB light microscopy, here we used the high sensitivity VirP labelling for light microscopy (A–C) and DAB-based immune-electron microscopy (D–G). With VirP, a clearly structured neuropil displaying numerous punctate profiles became evident in cerebral cortex (A) and hippocampus (B, C). The areas labelled by arrows are displayed at higher magnification in the insets in A–C. The labelled punctate profiles partly delineated dendritic profiles (asterisk in inset in B; arrowheads in inset in C), thus suggesting a synaptic localization of the antigens. With electron microscopy, ADC was localized to dendrites (D) and dendritic spines (E; compare unlabelled spines marked by asterisks), while Arg-like immunoreactivity was detected in presynaptic spine terminals (F). Postsynaptic densities of asymmetrical spine synapses are indicated by arrowheads. In neuronal cell bodies (G), the frequently displayed Arg-like punctate immunoreactivity was localized to numerous Golgi compartments including pre-Golgi buds at the endoplasmic reticulum (arrowhead in G). Rarely, immunoreactivity could also be attributed to ruptured mitochondria. Scale bars represent 20 µm in A–C; 5 µm in insets in A–C; 200 nm in D, E, inset in G; 100 nm in F; 500 nm in G.
Table 1.
Bacterial fusion proteins used for antibody generation, purification and characterization.