An NF-kappaB-like transcription factor in axoplasm is rapidly inactivated after nerve injury in Aplysia

Abstract

We found a protein in Aplysia neurons that has many characteristics of the transcription factor NF-kappaB. Thus, the protein recognized a radiolabeled probe containing the kappaB sequence from the human interferon-beta gene enhancer element (PRDII), and the binding was not affected by PRDIV, an ATF-2 enhancer sequence from the same gene. Binding was efficiently inhibited, however, by nonradioactive oligonucleotides containing H2, the kappaB site from the major histocompatibility complex I gene promotor. In addition, recombinant mammalian IkappaB-alpha, which associates specifically with the P65 subunit of NF-kappaB, inhibited the binding to the PRDII probe in a dose-dependent manner. The nuclear form of the Aplysia protein was constitutively active. Axoplasm, however, contained the constitutively active form as well as a latent form. The latter was activated by treatment with deoxycholate under the same conditions as mammalian NF-kappaB. Based on these findings, we believe the protein to be a homolog of NF-kappaB. To investigate the role of apNF-kappaB in the axon, we crushed the peripheral nerves to the body wall. Surprisingly, there was a rapid loss of apNF-kappaB binding at the crush site and, within 15 min, as far as 2.5 cm along the axon. In contrast, exposing either the intact animal or the nervous system in situ to levels of 5-HT that induce synaptic facilitation did not affect apNF-kappaB activity.

Fig. 1.

Characterization of Aplysia κB binding; representative results from two to five experiments are shown.A, Radiolabled probes were added to nuclear extracts and after EMSA, DNA-binding complexes were visualized using radioautography. Radiolabeled probe PII was added to a nuclear extract from HeLa cells to determine the position of NF-κB (p50/p65) (arrow). When radiolabeled PII was incubated with 5 μg of a nuclear extract, two proteins were detected (lane 1). Binding of one of the proteins (arrow) to the probe was not competed by unlabeled oligonucleotide PIV (lane 2), but was efficiently blocked by H2 (lane 3). The other protein (arrowhead) was unaffected by the presence of competitor DNAs. B, Nuclear κB-DNA binding to radiolabeled PII (arrow, lane 1) was inhibited by 1 μg (lane 2) and 2 μg (lane 3) of recombinant mammalian IκB-α. The other protein (arrowhead) was unaffected by IκB-α. In addition, binding to a radiolabeled AP-1 enhancer sequence (CTAGATGACTCAGCGCT; asterisk) was not affected by IκB-α at either concentration (lanes 4–6). The portion of the film containing the unbound probe was removed in this and subsequent figures. C, The κB binding was enriched in axoplasm extruded from the pedal nerves (lane 1, arrow) relative to that which remains in the sheath (lane 2). In contrast, most of the AP-1 (asterisk, lanes 3, 4) and CRE-binding (small arrows, lanes 5,6) activities were found in the sheath.D, The κB binding to PII in extruded axoplasm (lane 1) was not affected by excess PIV (lane 2), but was specifically disrupted by excess H2 oligonucleotide (lane 3). E, Latent κB activity was unmasked when extracts from nerve segments were treated with DOC (lanes 1, 2). The activity in the extracts was stimulated by a low concentration of DOC (lanes 3, 4), but was inhibited by 0.8% DOC (lane 5). Activity was recovered when excess DOC was removed by 0.85% NP-40 (lanes 2, 6).

Fig. 2.

Loss of apNF-κB DNA-binding activity after nerve crush. A, Schematic showing the crush site on nerves p8 and p9 to the tail. Crush refers to the nerve segment adjacent to the crush site, and P1 and P2 refer to segments more proximal. Each segment was 0.5 cm long. Segments of equal size on the contralateral nerves served as controls. Extruded axoplasm, or intact segments, from two to three animals was pooled for each experiment, and representative results are shown. Tissue fractions were assayed for κB binding by EMSA using the PII probe. B, Soluble axoplasm (10 μg) from control and crush segments was analyzed 20 hr after nerve crush. Constitutive apNF-κB-DNA binding was present in the control (lane 1) but absent in the crush segment (lane 2). Activity was reduced in nuclear extracts (10 μg) pooled from the pleuro-pedal neurons of four animals 20 hr after nerve crush, relative to that in control nuclear extracts (lanes 3,4), and activity was not recovered with DOC (lane 5). All of the pedal nerves were crushed in the experimental and none in the control. C, Five minutes after crush injury, 4 μg of protein from the segments was analyzed. Comparisons showed a loss of apNF-κB activity in the crush segments (lane 2) relative to controls (lane 1).D, To examine the loss of apNF-κB along the axon after crush injury, 4 μg of protein from the crush, P1, P2, and control segments was analyzed at 15 min (lanes 1–4) and 45 min (lanes 5–8). Constitutive apNF-κB binding was present in the controls 15 min (lanes 1) and 45 min (lane 5) after injury, but was absent in the crush (lanes 2, 6), P1 (lanes 3, 7), and P2 (lanes 4, 8) segments at both times. E, DOC treatment of 4 μg of extract from control segments unmasked latent activity (lanes 1, 2) but had no effect on the crush segments 15 min (lanes 3, 4) or 45 min (lanes 5, 6) after injury. F, Forty-five minutes after nerve crush, the sensory cell cluster from the crush-injured and control sides were excised and homogenized, and 600 ng of protein was analyzed. There was a dramatic loss of apNF-κB (lanes 1, 2) that was not recovered by DOC (lanes 3, 4).

Fig. 3.

5-HT does not affect apNF-κB activity in the axon. To assess the effect of 5-HT on apNF-κB activity, an intact animal was exposed to 250 μm 5-HT for 2 hr. Nerve segments (2 cm) were homogenized, and 6 μg of the extract was removed for EMSA using the PII probe. An equal amount of extract from segments from an animal incubated in seawater was used as a control. There was no difference in binding between the animal treated with the 5-HT (lane 2) and the control (lane 1). Likewise, exposure of half of the nervous system in situto 20 μm 5-HT for 2 hr did not alter apNF-κB activity in 4 μg of extracts (lane 4) relative to a similar extract from the contralateral control half incubated for the same time in seawater (lane 3). There was also no difference in the latent form (lanes 5,6).