Ubiquitin carboxyl-terminal hydrolase L1 is required for maintaining the structure and function of the neuromuscular junction

Abstract

The enzyme ubiquitin carboxyl-terminal hydrolase L1 (UCH-L1) is one of the most abundant proteins in the mammalian nervous system. In humans, UCH-L1 is also found in the ubiquitinated inclusion bodies that characterize neurodegenerative diseases in the brain, suggesting its involvement in neurodegeneration. The physiologic role of UCH-L1 in neurons, however, remains to be further elucidated. For example, previous studies have provided evidence both for and against the role of UCH-L1 in synaptic function in the brain. Here, we have characterized a line of knockout mice deficient in the UCH-L1 gene. We found that, in the absence of UCH-L1, synaptic transmission at the neuromuscular junctions (NMJs) is markedly impaired. Both spontaneous and evoked synaptic activity are reduced; paired pulse-facilitation is impaired, and synaptic transmission fails to respond to high-frequency, repetitive stimulation at the NMJs of UCH-L1 knockout mice. Morphologic analyses of the NMJs further revealed profound structural defects—loss of synaptic vesicles and accumulation of tubulovesicular structures at the presynaptic nerve terminals, and denervation of the muscles in UCH-L1 knockout mice. These findings demonstrate that UCH-L1 is required for the maintenance of the structure and function of the NMJ and that the loss of normal UCH-L1 activity may result in neurodegeneration in the peripheral nervous system.

UCH-L1 is one of the most abundant proteins in the mammalian brain; it constitutes 1–2% of the total protein extract in human brain (1, 2). UCH-L1 is best known to function as a deubiquitinating enzyme and hydrolyzes the C-terminal esters and amides of ubiquitin (3), but it also carries ubiquityl ligase activity that recognizes α-synuclein as its substrate (4). Additionally, UCH-L1 binds to monoubiquitin in neurons and thus stabilizes the level of monoubiquitin in neurons (5). Several lines of evidence suggest that UCH-L1 is involved in the pathogenesis of a number of neurodegenerative diseases, including Alzheimer’s disease (AD) and Parkinson’s disease (PD) (6). Spontaneous mutation of UCH-L1 in mice leads to spinal gracile axonal dystrophy (gad) (7 –9). In humans, UCH-L1 is accumulated in ubiquitinated inclusion bodies of AD and PD (10). In addition, UCH-L1 undergoes oxidative modification in AD (11, 12) and PD (12), and its expression is down-regulated in sporadic AD and PD (12). Together, these studies suggest that maintaining normal levels of UCH-L1 is crucial for normal brain function.

Besides the brain, UCH-L1 is also highly expressed in the peripheral nervous system, including sensory and motor nerves (13 –15). In fact, both sensory and motor nerves degenerate in gad mice (16, 17), but the mechanisms underlying the nerve degeneration in these mutant mice remain poorly understood. This lack of understanding is due, in part, to the lack of understanding of the physiologic role of UCH-L1 in neuron. For example, it is not clear to what extent the loss of UCH-L1 may affect synaptic function. Conflicting results have been reported for the role of UCH-L1 in hippocampal synaptic function. On one hand, impairment of the maintenance of theta-burst stimulation-induced long-term potentiation (LTP) in hippocampus was reported in gad mice (18). On the other hand, the maintenance of LTP was unaffected in another line of mutant mice carrying a spontaneous deletion of the UCH-L1 gene (nm3419) (19).

In this study we have investigated the role of UCH-L1 in synapses by analyzing the structure and function of the NMJ in mutant mice carrying a targeted deletion in the UCH-L1 gene (UCH-L1 knockout mice). Using the NMJ as a synapse model, we found profound dysfunction of the NMJ, accompanied by structural defects and progressive degeneration of the motor nerves in the UCH-L1 knockout mice. These findings demonstrate that UCH-L1 is required for the maintenance of normal structure and function of the NMJ in mice.

Results

UCH-L1 Null Mutation Leads to Progressive Paralysis and Premature Death in Mice.

To study the role of UCH-L1 in vivo, we obtained mutant mice carrying a targeted deletion of the UCH-L1 gene from the Mutant Mouse Regional Resource Centers (MMRRC) (see Experimental Procedures for details). The heterozygous (UCH-L1^+/−) mice were viable, fertile, and devoid of any gross phenotypic defects and were bred together to generate homozygous mutants. To verify the UCH-L1 expression in the homozygous mutants, we immunoblotted the tissue homogenates obtained from three pairs of WT and homozygous mutant mice using the anti-PGP 9.5 antibody, which recognizes UCH-L1 (1, 2). As expected, UCH-L1 was detected in both the brain and spinal cord from the WT mice but absent from the mutant mice (Fig. 1C). To examine the expression of UCH-L1 at the NMJ, we immunostained teased muscle fibers with the anti-PGP 9.5 antibody. In the WT mice (^+/+; Fig. 1F), PGP9.5 antibody intensely labeled the motor axons, including both the preterminal nerves and nerve terminals, but not the postsynaptic muscles, confirming that UCH-L1 is expressed at the presynaptic nerves. Consistent with our Western blotting results (Fig. 1C), no PGP 9.5 staining was detected in the UCH-L1 mutants (Fig. 1I). Thus, these results confirmed that these mutant mice carried null mutation for UCH-L1 (UCH-L1 ^−/−).

Fig. 1.

External phenotype and survival of UCH-L1 ^−/− mice. (A UCH-L1 ^−/− mouse (144 days) (B) compared with its WT (^+/+) littermate control (A). (C) Western blot analysis of the brain and spinal cord homogenates prepared from three pairs of WT and UCH-L1 ^−/− mice, probed with PGP9.5 antibody. UCH-L1 was detected in the WT but not the UCH-L1 ^−/− mice. (D) Survival curve for UCH-L1 ^−/− mice (UCH-L1 ^−/− mice: n = 75; controls: n = 75). (E–J) Teased hind-limb muscle fibers from WT (E–G) and UCH-L1 ^−/− (H–J), double-labeled with α-bgt (for AChRs) and PGP 9.5 antibody. UCH-L1 was detected in the WT nerve terminal (arrowhead) and in preterminal nerves (arrow) but absent in UCH-L1 ^−/− nerves (I). (Scale bar in J, 25 μm for E–J.)

To determine the external phenotype and survival rate of UCH-L1 ^−/− mice, we monitored 150 animals—75 control (including both WT and heterozygous mice) and 75 UCH-L1 ^−/− mice (Fig. 1D). For a period of 10 months, all of the 75 control mice survived. In contrast, none of the UCH-L1 ^−/− mice survived beyond 10 months. At age 2 months, however, the majority of the UCH-L1 ^−/− mice cannot be readily distinguished from their WT or heterozygous littermates, except for a small number of mice (5 of 75, or 6.7%), which died at approximately 1 month of age. The majority (61 of 75, or 81%) of the UCH-L1 ^−/− mice survived during the first 6 months of age. Among these mice, the first noticeable external phenotype appeared at ≈50 days of age, when they exhibited clasping and spastic movements in their hind limbs while suspended by their tails. By age 3 months, they had developed kyphosis and began to display a clumsy, hopping gait. Between 4 and 6 months of age, their hind limbs became paralyzed (Fig. 1B), and by age 8 months, their fore limbs became paralyzed. Most of the UCH-L1 ^−/− mice died between 7 and 10 months of age (the last one died at age 294 days, or 10 months) (Fig. 1D). These results demonstrated that although UCH-L1 is dispensable for mouse development, it is required for survival in mice.

UCH-L1 Null Mutation Leads to Marked Impairment of Synaptic Transmission at the NMJ.

The paralysis of the UCH-L1 ^−/− mice prompted us to examine synaptic transmission at the NMJ, where a disruption in synaptic transmission may result in paralysis. We carried out electrophysiologic analyses on the extensor digitorum longus (EDL) muscles, because paralysis was first seen in the hind limbs, and because EDL is one of the best-characterized fast-twitch muscles in the hind limbs (20 –22).

First, we asked whether spontaneous synaptic activity was compromised in the absence of UCH-L1. We analyzed miniature end-plate potentials (mEPPs) from UCH-L1 ^−/− and age-matched controls (WT or UCH-L1 ^+/−). In control muscles (1.5–5 months of age), we observed that the mEPP frequencies were typically within the range of 1.0–1.5 Hz and amplitudes of 0.5–0.6 mV (Fig. 2); these recordings are consistent with those of previous studies in WT EDL muscles (20, 21). In UCH-L1 ^−/− mice, at 1.5 months of age, mEPP frequencies were similar to those recorded in control (1.05 ± 0.12 Hz and 0.93 ± 0.08 Hz, respectively; P = 0.398). However, by age 2 months, mEPP frequencies were significantly reduced in UCH-L1 ^−/− mice compared with control mice (0.61 ± 0.08 Hz and 1.23 ± 0.20 Hz, respectively; P = 0.002). A similar disparity was observed at age 3 months (0.57 ± 0.11 Hz and 1.25 ± 0.15 Hz, respectively; P < 0.001) and at age 5 months (0.62 ± 0.08 Hz and 1.31 ± 0.15 Hz, respectively; P < 0.001) (Fig. 2B). Interestingly, whereas mEPP frequencies were greatly reduced in the UCH-L1 ^−/− mice, mEPP amplitude remained unaffected; no significant difference was found for mEPP amplitude between the control and UCH-L1 ^−/− mice at each age group (Fig. 2C). These data suggest that spontaneous synaptic transmission through the NMJ is impaired in UCH-L1 ^−/− mice and that these defects likely reside in the presynaptic sites.

Fig. 2.

Reduced spontaneous synaptic activity in UCH-L1 ^−/− mice. (A) Sample traces of mEPPs from UCH-L1 ^−/− and their littermate control (WT or UCH-L1 ^+/−) mice at 1.5, 2, 3, and 5 months of age. Each trace is a set of 20 superimposed 200-ms sweeps representing a 4-s continuous recording from an individual muscle fiber. (B) Quantitative analysis of mEPP frequencies (mean ± SEM): MEPP frequencies were significantly reduced in UCH-L1 ^−/− mice at 2, 3, and 5 months. (C) No significant difference was found for mEPP amplitudes between control and UCH-L1 ^−/− mice at any age. The numbers of muscle fibers (n) and mice (N) at each age groups were as follows: 1.5 months (control: n = 51, N = 4; UCH-L1^−/−: n = 30, N = 3), 2 months (control: n = 20, N = 3; UCH-L1^−/−: n = 29, N = 3), 3 months (control: n = 23, N = 3; UCH-L1^−/−: n = 21, N = 3), and 5 months (control: n = 51, N = 7; UCH-L1^−/−: n = 52, N = 5). **P < 0.01, Student's t test.

We next analyzed the evoked neurotransmitter release. We recorded end-plate potentials (EPPs) in control and UCH-L1 ^−/− mice at 1.5, 2, 3, and 5 months of age (Fig. 3). Again, we observed age-dependent reduction of EPP amplitudes in UCH-L1 ^−/− mice. Initially, EPP amplitudes were similar for both control and UCH-L1 ^−/− mice at age 1.5 months (12.96 ± 0.55 mV and 13.73 ± 0.84 mV, respectively) and at age 2 months (12.96 ± 0.66 mV and 13.09 ± 0.88 mV, respectively), but they were significantly reduced in mutant mice compared with control mice at age 3 months (13.10 ± 1.00 mV vs. 16.99 ± 0.86 mV, respectively; P < 0.005) and 5 months (9.30 ± 0.66 mV vs. 16.11 ± 0.42 mV, respectively; P < 0.001) (Fig. 3B). Similarly, the mean quantal content was also significantly reduced at 3 and 5 months of age in the mutant mice (Fig. 3C). Interestingly, when EPP latencies and the rising slopes (10–90%) were compared between control and UCH-L1 ^−/− mice at each age group, we noticed that the EPP latencies were significantly prolonged and that the rising slopes were significantly decreased in UCH-L1 ^−/− mice at age 2, 3, and 5 months (Fig. 3 D and E). The increases in EPP latencies suggest that the rate of action potential propagation may be reduced, because the length of the nerves aspirated into the suction electrode was kept approximately the same between the control and UCH-L1 ^−/− mice to avoid fluctuations of the latencies due to varied length of the nerve. On the other hand, the reduced rising slopes of the EPPs may also be attributed to either decreased amount of neurotransmitter ACh released from the presynaptic nerve terminal, or decreased number of ACh receptors (AChRs) available on the postsynaptic membrane (23). We therefore measured the fluorescence intensity of postsynaptic endplates labeled by α-bungarotoxin (α-bgt). We found that α-bgt labeling intensity at the endplates was similar between control and UCH-L1 ^−/− mice (Fig. S1). Together with the reduction of EPP amplitude and quantal content in the UCH-L1 ^−/− mice, our results demonstrate that the amount of neurotransmitter release is reduced at the NMJs of these mutants.

Fig. 3.

Reduced evoked synaptic transmission and impaired short-term synaptic plasticity in UCH-L1 ^−/− mice. (A) Sample EPP traces from UCH-L1 ^−/− and their littermate control (WT or UCH-L1 ^+/−) mice at 1.5, 2, 3, and 5 months of age. (B–E) Bar graphs of average EPP amplitudes (mV), quantal content, latency (ms) and 10–90% rising slope (mV/ms). Numbers of muscle fibers (n) and mice (N) analyzed for EPPs were as follows: 1.5 months (control: n = 52, N = 4; UCH-L1 ^−/−: n = 31, N = 3), 2 months (control: n = 25, N = 3; UCH-L1 ^−/−: n = 33, N = 3), 3 months (control: n = 24, N = 3; UCH-L1 ^−/−: n = 27, N = 3), and 5 months (control: n = 99, N = 7; UCH-L1 ^−/−: n = 76, N = 5). (F) A twin-pulse stimulus induced PPF at the NMJs of control mice at both 1.5 and 2 months of age. In contrast, in UCH-L1 ^−/− mice, twin-pulse stimulation induced PPF at 1.5 months but PPD at 2 months of age. (G) Paired-pulse ratio (EPP₂/EPP₁) plotted against interpulse intervals at 20, 30, 40, and 50 ms. Numbers of muscle fibers (n) and mice (N) analyzed for PPF were as follows: 1.5 months (control: n = 47, N = 4; UCH-L1 ^−/−: n = 40, N = 2), 2 months (control: n = 24, N = 3; UCH-L1 ^−/−: n = 26, N = 3). **P < 0.001, Student's t test.

To further characterize the dysfunction of the neuromuscular synapses in UCH-L1^−/− mice, we examined short-term synaptic plasticity by applying paired-pulse stimulation and trains of repetitive, high-frequency stimulation within physiologic ranges (24). Recordings of the postsynaptic potentials evoked by paired-pulse stimulation revealed that facilitation was elicited in both control and UCH-L1^−/− mice at age 1.5 months (Fig. 3G). However, paired-pulse facilitation (PPF) was not elicited in UCH-L1 ^−/− mice at age 2 months; instead, paired-pulse depression (PPD) was induced (Fig. 3 F and G). Thus, short-term plasticity of the NMJs in UCH-L1^−/− mice was impaired at 2 months of age.

We next delivered a train of stimulation at increasing frequency (from 10 to 50 Hz). At 10 Hz, every stimulus evoked a postsynaptic EPP in both control and UCH-L1 ^−/− mice, but the rundown of EPP amplitude was faster in the mutant after 2 months of age (Fig. S2). Similarly, the rate of EPP rundown increased in the mutant mice at 2, 3, and 5 months of age, in responding to 30 Hz train stimulation (Fig. S3). Interestingly, at 50 Hz, faster rundown rate of EPP amplitudes was observed at all ages, including 1.5 months, in UCH-L1 ^−/− mice (Fig. S4); that is, the rate of decline in EPP amplitude was considerably faster in mutant mice compared with control mice. Increased EPP rundown in UCH-L1 ^−/− mice (compared with age-matched WT mice) worsened with age (Fig. S4 A and B). Furthermore, incidents of synaptic transmission failure were observed in UCH-L1 ^−/− mice in responding to repetitive stimulation at 30 Hz or 50 Hz (Fig. S4C). The incident of synaptic transmission failure correlated with the age of the mice. At 30 Hz, synaptic transmission failures were observed only at 5 months of age, in 14% of cells (9 of 64). At 50 Hz, the failures were also observed in 2- and 3-month-old UCH-L1 ^−/− mice (10–15%) and 5-month-old mice (30% of the cells, 19 of 64). These extreme responses to rapid stimulation patterns suggest that UCH-L1 is essential to maintain the plasticity and thus optimal functioning of the NMJ, and that a lack of UCH-L1 contributes to the progressive, age- and use-dependent loss of normal neuromuscular activity.

Loss of UCH-L1 Results in Progressive Degeneration of Presynaptic Terminals at the NMJ.

Loss of transmission may occur when synaptic structures are altered. To determine whether UCH-L1 contributes to the structural integrity of the NMJ, we examined the morphology of the neuromuscular synapses in the EDL muscle that had been previously analyzed electrophysiologically as described above, and in triangularis sterni (TS) muscle (which is ideal for whole-mount staining of the NMJs) (25). We double-labeled these muscles with anti-syntaxin or anti-synaptotagmin-2 antibodies to label presynaptic nerves and with α-bgt to label postsynaptic AChRs (end-plates). In control mice, presynaptic nerve terminals were closely aligned with their respective postsynaptic end-plates (Fig. 4 A–C, upper rows), whereas in UCH-L1^−/− mice some presynaptic terminals appeared to have retracted from their corresponding end-plates (Fig. 4 A and B, lower rows), and some end-plates were totally vacant (denervated) (Fig. 4C, lower row). The degeneration of presynaptic nerves in UCH-L1^−/− mice became progressively worse over time. At age 1.5 months, all of the end plates in the TS muscles of the mutant animals were fully innervated (Fig. S5). At 2 and 3 months of age, however, 1% and 2% (8 of 643 and 13 of 548, respectively) of those end-plates were denervated; by age 5 months, 28% (168 of 596) were denervated. Similarly, denervation of end-plates in EDL also progressed with age: ≈11–12% of EDL end-plates had been denervated by 2 to 3 months and 57% by 5 months of age in UCH-L1^−/− mice (Table S1). Thus, in both TS and EDL muscles in UCH-L1^−/− mice, the majority of denervation took place after 3 months of age and coincided with the onset of paralysis.

Fig. 4.

Denervation of the end-plates in UCH-L1 ^−/− mice. EDL muscles (A, 3 months) or TS muscles (B, 3 months; C, 5 months) were double-labeled with Texas-Red conjugated α-bgt and antibodies against syntaxin (A) or synaptotagmin-2 (B and C). In control mice, presynaptic nerve terminals (green) and postsynaptic end-plates (red) are adjacent to each other, as illustrated in the merged images. In UCH-L1 ^−/− mice, some nerve terminals retracted from their corresponding end-plates and developed swollen ends (arrowheads in A and B). Some postsynaptic sites were completely denervated (vacant) (arrows in C). (Scale bar in A, 20 μm; in C, 25 μm for B and C.)

To determine the spatial pattern of presynaptic degeneration in UCH-L1^−/− mice, we examined cross-sections of the ventral roots of the spinal cord and the distal nerves innervating the EDL muscles in both animal groups. The ventral roots in the mutant mice were normal, even at age 8 months (Fig. S6C), but the density of the large-caliber myelinated axons in their distal nerves was greatly reduced at age 8 months (Fig. S6F) compared with age-matched WT mice (Fig. S6D). A similar loss of myelinated axons was also evident in the distal nerves in younger mutants (age 3.5 months), but considerably fewer axons (Fig. S6E) were lost compared with the distal nerves from 8-month-old mutant (Fig. S6F). These observations suggest that the degeneration of axons in UCH-L1^−/− mice progressed in a distal-to-proximal (“dying back”) pattern.

Loss of UCH-L1 Leads to Ultrastructural Defects of Presynaptic Nerve Terminals at the NMJ.

To further understand the structural basis of synaptic transmission impairments at the NMJ of UCH-L1 ^−/− mice, we carried out EM studies. Like the majority of the higher vertebrate skeletal muscles, each muscle fiber in mice receives single innervation at a single site that is less than 0.1% of the surface of individual muscle fiber (26). Thus, the most frequently encountered structures we observed under EM were sarcomeres, the basic unit of striated myofibrils. Strikingly, the fine structures of muscle fibers of the UCH-L1 ^−/− mice were similar to those of the WT mice. The Z lines, M lines, I-bands, and A-bands were all readily identifiable in both genotypes (Fig. S7). At the NMJs, the postsynaptic muscle membrane formed characteristic junctional folds, which apposed with the presynaptic nerve terminal (Fig. 5A). These postsynaptic junctional folds also appeared normally in UCH-L1 ^−/− mice (Fig. 5 B and C, white arrowhead).

Fig. 5.

Ultrastructural defects in the nerve terminals of EDL muscles in UCH-L1^−/− mice. (A) An NMJ in a 5-month-old WT mouse. The nerve terminal is filled with synaptic vesicles (SV) and mitochondria (M). Junctional folds are present at the postsynaptic muscle membrane (white arrowhead). (B) An NMJ in a 5-month-old UCH-L1 ^−/− mouse. The SV population is markedly reduced, and an excessive number of tubulovesicular structures (black arrowheads) and MLB are seen in the terminal. (C–E) Another NMJ from a 5-month-old UCH-L1 ^−/− mouse. Branched tubulovesicular profiles have accumulated within the nerve terminal (magnified in D and E, black arrowheads). An MLB is evident, but few SVs are present. The axoplasm appears electron-dense compared with the electron-lucent WT terminal in A and appears fibrous under higher magnification (asterisk in E). Note the wider gap between the nerve terminal and the postsynaptic membrane (black arrow), indicating nerve detachment. DCV, dense-core vesicle. Asterisk (*) denotes the accumulation of neurofilaments in the nerve terminal.

At the presynaptic nerve terminals of the NMJs, however, ultrastructural defects were evident in UCH-L1 ^−/− mice. In contrast to WT mice, in which the presynaptic nerve terminals typically appeared electron lucent and were filled with abundant synaptic vesicles and mitochondria (Fig. 5A), the presynaptic nerve terminals contained a marked reduction of synaptic vesicles (Fig. 5B). In addition, an excessive number of branched tubulovesicular profiles (Fig. 5 B, D, and E, black arrowheads) and multilamellar bodies (MLBs) (Fig. 5 B and C) were often present in the presynaptic nerve terminals of UCH-L1 ^−/− mice. The ultrastructural defects of the presynaptic nerve terminals were evident at 2 months of age, despite the fact that the UCH-L1 ^−/− mice had no visibly detectable symptoms at this age. The density of synaptic vesicles was significantly (P < 0.001) reduced compared with that from the WT littermate mice. In rodents, the normal ratio of presynaptic contact length to postsynaptic contact length is approximately 80% (27). This ratio was significantly reduced in the UCH-L1 ^−/− mice compared with the WT littermate control, suggesting detachment of presynaptic nerve terminal and/or Schwann cell invasion to the synaptic cleft. These alterations of ultrastructure were consistently observed in UCH-L1 ^−/− mice at 5 and 8 months of ages (Table S2).

Consistent with the synaptic degeneration phenotype we observed under the light microscope (Fig. 4), degeneration of presynaptic nerve terminals was also evident under EM. The degenerating nerve terminals were typically wrapped by Schwann cells, whose processes invaded into the synaptic cleft (Fig. S8B, arrowhead), whereas the nerve terminals in the WT mice were typically capped by the Schwann cell (Fig. S8A, arrowhead). In some cases, a structure enclosing a few clear synaptic vesicles (Fig. S8C, arrow) was found within the Schwann cell at the synaptic site marked by postsynaptic junctional folds (Fig. S8 C and D). This seemed to be the remains of the presynaptic nerve terminals engulfed by the Schwann cell; these abnormal structures were not present in WT Schwann cells. These ultrastructural defects are consistent with the observed synaptic dysfunction at the NMJs of UCH-L1 ^−/− mice.

Discussion

In this study, we have investigated the role of UCH-L1 in neuromuscular synaptic structure and function by examining mutant mice deficient in the UCH-L1 gene. We found that UCH-L1 is required for the maintenance of the structure and function of the NMJ in mice. In the absence of UCH-L1, neurotransmitter release at the motor nerve terminals was markedly reduced, and the presynaptic nerves exhibit progressive degeneration, starting from the nerve terminals and progressing to the proximal nerves, in a “die-back” pattern. This accompanied paralysis and premature death, probably as a result of respiratory failure caused by neuromuscular dysfunction and degeneration. These results are consistent with those of previous studies in the spontaneous mutation of UCH-L1 in gad mutant mice (16, 17). However, those studies of gad mutants do not indicate the extent to which NMJ transmission is affected. We were able to determine that UCH-L1 plays a critical role in the maintenance of normal neuromuscular synaptic transmission.

This impairment of synaptic transmission at the NMJs of the UCH-L1 ^−/− mice may be attributed to presynaptic defects. mEPP frequencies are significantly reduced, but mEPP amplitudes remained unaffected in the UCH-L1 ^−/− mice. Because the quantal content is reduced but the mEPP amplitude is unaffected, it suggests that the amount of neurotransmitter released from an individual vesicle is similar but the total output of the neurotransmitter release is reduced. Consistent with this idea, we found marked reduction of synaptic vesicles in the presynaptic nerve terminals but no detectable alternation of the postsynaptic endplates at the NMJs of the UCH-L1 ^−/− mice.

In UCH-L1^−/− mice, twin-pulse stimulation induced PPD, instead of PPF, which normally occurs at the NMJs of WT mice. A number of mechanisms have been proposed for the physiology of PPF, including alternation of probability of release (28, 29), which could be attributed to accumulation of residual Ca²⁺ that enters the nerve terminal after the first nerve impulse (30), or alternated Ca²⁺ channel/release-site distribution (31). However, the most straightforward explanation for PPD at the NMJs of UCH-L1^−/− mice is the depletion of synaptic vesicle pools, as revealed by EM. Similarly, the rapid decline in EPP amplitude during repetitive stimulation is likely due to a decrease in the quantal content resulting from depletion of synaptic vesicle pools (32, 33).

Our studies of synaptic transmission impairment at the NMJs of UCH-L1 knockout mice are consistent with previous studies demonstrating a role for UCH-L1 in synaptic function in the brain. For example, Ap-uch (the Aplysia homolog of UCH-L1) is crucial for the induction of long-term synaptic facilitation (34, 35) and depression (36). Overexpression of UCH-L1 in mice rescues β-amyloid-induced synaptic dysfunction (37). Furthermore, a spontaneous deletion in the UCH-L1 gene leads to profound defects of theta-burst-induced LTP in hippocampus (18).

The molecular mechanisms underlying UCH-L1’s involvement in regulating synaptic function remain to be further elucidated. One plausible mechanism is that UCH-L1 may be required for endocytosis and synaptic vesicle recycling. This hypothesis is based on the types of ultrastructural defects we observed in the presynaptic nerve terminals in UCH-L1^−/− mice. These defects included reduction of synaptic vesicle number and accumulation of branched tubulovesicular profiles in the mutant presynaptic nerve terminal and resemble those seen in nerve terminals undergoing intense exocytosis (38) and in presynaptic terminals that are defective in vesicle endocytosis/recycling, such as those seen in a temperature-sensitive shibire mutant in Drosophila (39), in dynamin-1–null mice (40, 41), and in nerve terminals poisoned with the disulfide reducing agent 2,4-dithiobiuret (42). Additionally, the synaptic transmission impairment we observed in UCH-L1^−/− mice seems similar to that produced in shibire mutants (39, 43). Several lines of evidence suggest that UCH-L1 is involved in stabilizing the level of ubiquitin in neurons (5, 44). Because ubiquitin plays a critical role in endocytosis of membrane proteins (45), it is conceivable that in the absence of UCH-L1, synapses fail to maintain normal levels of ubiquitin required for vesicle endocytosis or vesicle regeneration after endocytosis. Thus, further studies of vesicle endocytosis in UCH-L1^−/− mice may provide insights into the role of UCH-L1 in the maintenance of synaptic structure and function at the NMJ.

Experimental Procedures

Mice.

UCH-L1 mice (B6; 129-UCH-L1^tm1Dgen/Mmnc, 011642-UNC, mouse genome informatics: 3604452) were obtained from MMRRC, a strain repository funded by the National Institutes of Health (NIH), which received them as a generous donation from Deltagen. The murine UCH-L1gene consists of nine exons. The UCH-L1 mutant mice were generated by targeted deletion of a region containing exons 6 through 8 and the first 6 base pairs of exon 9. The heterozygous mice were viable, fertile, and devoid of any gross phenotypic defects. The homozygous mutants were generated by breeding heterozygous mice together. The following primer sets were used for genotyping: WT allele forward, CCT TGC CTC CGT CCT CTA TTA AAG C, and reverse, CTC TCC CCA GAC TTA AGC TGC TTT G (product size: 213 bp); UCH-L1 mutant allele forward, GGG TGG GAT TAG ATA AAT GCC TGC TCT, and reverse, CTC TCC CCA GAC TTA AGC TGC TTT G (product size: 447 bps). The mice were analyzed in littermate pairs, including control mice, either WT or heterozygous, and homozygous mutant (UCH-L1^−/−) mice. All experimental protocols followed NIH Guidelines and were approved by the University of Texas Southwestern Institutional Animal Care and Use Committee.

Electrophysiology.

Electrophysiologic analysis was carried out in acutely isolated EDL muscle with the common peroneal nerve attached. All recordings were made at room temperature (22°C) and in oxygenated Ringer’s solution (46). The end-plate regions were impaled with glass micropipettes (30–40 MΩ). Intracellular voltage recording was acquired by a high-impedance microelectrode amplifier (AxoClamp-2B). Evoked end-plate potentials were triggered by suprathreshold stimulation of the nerve using a suction electrode. Muscle contraction was prevented by bath application of μ-conotoxin GIIIB (1 μM). The signal was digitized at 10 kHz and analyzed using pClamp 9.0. The amplitudes of mEPP and EPP were normalized as described previously (47). Quantal content was estimated by the direct method (quantal content = EPP amplitude/mEPP amplitude) (48).

Western Blot.

Tissues were collected from three pairs of 2-month-old mice (3 UCH-L1^+/+ and 3 UCH-L1^−/−). Brain and spinal cords were dissected and homogenized in Tris buffer containing 50 mM Tris-NaOH, pH 7.4, 150 mM NaCl, 2 mM EDTA, 1 mM PMSF, and protease inhibitor mixture (Roche Applied Science). Tissue homogenates were separated by SDS-PAGE, transferred to a nitrocellulose membrane, and blocked in 5% milk in Tris-buffered saline. The membrane was probed with anti-PGP 9.5 (Biomol, catalog no. PG 9500) or anti-valosin-containing protein (VCP, as an internal loading control) (gift from Dr. Cezary Wójcik, Indiana University), followed by peroxidase-conjugated secondary antibodies, and visualized with enhanced chemiluminescence.

Morphologic Analysis of the Neuromuscular Synapse.

Light microscopy and quantitative analysis of the NMJs in EDL.

Whole-mount TS muscles and EDL muscles from control (WT or UCH-L1 ^+/− mice) and UCH-L1^−/− mice at various ages (1.5, 2, 3, and 5 months) were analyzed. EDL muscles were serial-sectioned at 30 μm sagittally. Muscle sections were then incubated with 2 nM Texas-red conjugated α-bgt (Invitrogen) and with antibodies against syntaxin or synaptotagmin 2 (49), followed by FITC-conjugated secondary antibodies to label the motor nerves, as previously described (50). Additionally, teased muscle fibers were double-labeled with a PGP 9.5 antibody and α-bgt to examine the expression pattern of UCH-L1 at the NMJ. Images were acquired by a Zeiss confocal microscope.

EM and EM morphometry.

Mice were perfused with a freshly prepared fixative (1% glutaraldehyde and 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4). Tissues were then dissected and remained in the same fixative overnight. The tissues were postfixed with 1% osmium tetroxide and embedded in Epon 812. Semithin sections were cut at 0.5 μm thickness and stained with toluidine blue for light microscopic observation of the ventral roots and distal nerves. Ultrathin sections (70–80 nm) were stained with uranyl acetate and lead citrate and observed using a Tecnai transmission electron microscope. Four UCH-L1^−/− mice (one each at 2, 3.5, 5, and 8 months of age) and three WT mice (littermates of the UCH-L1^−/− mice at 2, 5, and 8 months) were analyzed. For EM morphometry studies, we measured the following parameters from each presynaptic nerve terminal profile using NIH ImageJ: nerve terminal area, nerve terminal perimeter, synaptic contact length, and postsynaptic membrane length (excluding junctional folds) (27). The synaptic contact was defined as the length of the presynaptic plasma membrane that was apposed to the postsynaptic muscle membrane at a distance of 50–80 nm (51, 52). All synaptic vesicles in each nerve terminal profile were counted to give rise to the total synaptic vesicle number. The synaptic vesicle density was calculated from total synaptic vesicle number divided by nerve terminal area and expressed per square micrometer (μm²). The pre/postsynaptic ratio was calculated from synaptic contact length divided by postsynaptic membrane length and expressed as percentage. A two-tailed t test was used to determine statistical significance between the WT and UCH-L1^−/− mice at the same age.