Sensory motor neuropathy is associated with various inherited disorders including Charcot-Marie-Tooth disease, X-linked adrenoleukodystrophy/adrenomyeloneuropathy and Refsum disease. In the latter two, the neuropathy is thought to result from the accumulation of specific fatty acids. We describe here three patients with elevated plasma concentrations of pristanic acid (a branched-chain fatty acid) and C27-bile-acid intermediates. Two of the patients suffered from adult-onset sensory motor neuropathy. One patient also had pigmentary retinopathy, suggesting Refsum disease, whereas the other patient had upper motor neuron signs in the legs, suggesting adrenomyeloneuropathy. The third patient was a child without neuropathy. In all three patients we discovered a deficiency of alpha-methylacyl-CoA racemase (AMACR). This enzyme is responsible for the conversion of pristanoyl-CoA and C27-bile acyl-CoAs to their (S)-stereoisomers, which are the only stereoisomers that can be degraded via peroxisomal beta-oxidation. Sequence analysis of AMACR cDNA from the patients identified two different mutations that are likely to cause disease, based on analysis in Escherichia coli. Our findings have implications for the diagnosis of adult-onset neuropathies of unknown aetiology.

A specific racemase for alpha-methylacyl-CoAs, which had previously been studied in rat liver [W. Schmitz, R. Fingerhut, E. Conzelmann (1994) Eur. J. Biochem. 222, 313-323], has now been demonstrated also in human tissues. The human enzyme cross-reacts with a polyclonal antiserum against the rat liver racemase. The racemase was purified from human liver some 3600-fold. It is a monomer of 47 kDa with an isolectric point of pH 6.1 and is optimally active between pH 7-8. It acts only on coenzyme A thioesters, not on free fatty acids, and accepts as substrates a wide range of alpha-methylacyl-CoAs, including pristanoyl-CoA and trihydroxycoprostanoyl-CoA (an intermediate in bile acid synthesis), but neither 3-methyl-branched nor linear-chain acyl-CoAs. A clear difference in subcellular localization of the enzyme was found between humans and rats: the rat enzyme co-distributed exclusively with mitochondrial marker enzymes whereas in human cells, only 10-30% of the activity was found in mitochondria, the bulk activity was located in peroxisomes. Cells from patients with general deficiency of peroxisome assembly (Zellweger syndrome) showed strongly reduced racemase activity, with only the mitochondrial share being present while the peroxisomal form was absent.

2-Methylacyl-CoA racemase is an auxiliary enzyme required for the peroxisomal beta-oxidative breakdown of (2R)-pristanic acid and the (25R)-isomer of C(27) bile acid intermediates. The enzyme activity is found not only in peroxisomes but also is present in mitochondria of human liver and fibroblasts. The C terminus of the human racemase, a protein of 382 amino acids with a molecular mass of 43,304 daltons as deduced from its cloned cDNA, consists of KASL. Hitherto this sequence has not been recognized as a peroxisomal targeting signal (PTS1). From the in vitro interaction between recombinant racemase and recombinant human PTS1 receptor (Pex5p), and the peroxisomal localization of green fluorescent protein (GFP) fused to the N terminus of full-length racemase or its last six amino acids in tranfected Chinese hamster ovary (CHO) cells, we concluded that ASL is a new PTS1 variant. To be recognized by Pex5p, however, the preceding lysine residue is critical. As shown in another series of transfection experiments with GFP fused to the C terminus of the full-length racemase or racemase with deletions of the N terminus, mitochondrial targeting information is localized between amino acids 22 and 85.Hence, our data show that a single transcript gives rise to a racemase protein containing two topogenic signals, explaining the dual cellular localization of the activity.

Most proteins are targeted to the peroxisomal matrix by virtue of a peroxisomal targeting signal-1 (PTS1), a short carboxy-terminal sequence specifically recognized by the PTS1 receptor Pex5p. We had previously developed a model that allowed the estimation of the affinities of many PTS1 sequences within the human proteome for Pex5p that revealed a wide range of predicted affinities. We have now experimentally determined the affinities of the PTS1-containing peptides from 42 proteins from the human proteome for Pex5p and show that these range over 4 orders of magnitude. These affinities correlate reasonably well with the predicted values and are substantially more precise. In an attempt to provide a possible explanation for the wide range of PTS1-Pex5p affinities, we compared these affinities with mRNA levels (as a proxy for rates of protein production) of the genes encoding these proteins in 79 human tissues and cell types. We note that high affinity PTS1-Pex5p interactions tend to correspond to proteins encoded by genes expressed at relatively low levels, whereas lower affinity PTS1-Pex5p interactions tend to correspond to proteins encoded by genes exhibiting higher levels and wider ranges of expression. Further analysis revealed that these relationships are consistent with the notion that a relatively uniform pool of protein-Pex5p complexes is maintained for appropriate peroxisome assembly.

Sensory motor neuropathy is associated with various inherited disorders including Charcot-Marie-Tooth disease, X-linked adrenoleukodystrophy/adrenomyeloneuropathy and Refsum disease. In the latter two, the neuropathy is thought to result from the accumulation of specific fatty acids. We describe here three patients with elevated plasma concentrations of pristanic acid (a branched-chain fatty acid) and C27-bile-acid intermediates. Two of the patients suffered from adult-onset sensory motor neuropathy. One patient also had pigmentary retinopathy, suggesting Refsum disease, whereas the other patient had upper motor neuron signs in the legs, suggesting adrenomyeloneuropathy. The third patient was a child without neuropathy. In all three patients we discovered a deficiency of alpha-methylacyl-CoA racemase (AMACR). This enzyme is responsible for the conversion of pristanoyl-CoA and C27-bile acyl-CoAs to their (S)-stereoisomers, which are the only stereoisomers that can be degraded via peroxisomal beta-oxidation. Sequence analysis of AMACR cDNA from the patients identified two different mutations that are likely to cause disease, based on analysis in Escherichia coli. Our findings have implications for the diagnosis of adult-onset neuropathies of unknown aetiology.