The recruitment of parkin to mitochondria in response to mitochondrial membrane depolarization1 is an important aspect of parkin biology that has attracted extensive research2. We found that parkin with the sequence ({type:entrez-nucleotide,attrs:{text:AB009973.1,term_id:3063387,term_text:AB009973.1}}AB009973.1) reported in the original discovery of the gene3 was not recruited to mitochondria following CCCP treatment, while parkin with the current Genbank Reference Sequence ({type:entrez-nucleotide,attrs:{text:NM_004562.2,term_id:169790968,term_text:NM_004562.2}}NM_004562.2) was robustly recruited to mitochondria under the same condition. The only difference between the two sequences is nucleotide 768, which is C in {type:entrez-nucleotide,attrs:{text:AB009973.1,term_id:3063387,term_text:AB009973.1}}AB009973.1 (encoding for proline at amino acid position 223) and T in {type:entrez-nucleotide,attrs:{text:NM_004562.2,term_id:169790968,term_text:NM_004562.2}}NM_004562.2 (for serine at 223). As shown in Fig. 1, Hela cells were transfected with FLAG-tagged parkin (with either P or S at amino acid 223). Cells were treated without or with CCCP (10 μM) for 3 hr and immunostained for FLAG and Tom20. CCCP induced a strong mitochondrial recruitment of FLAG-parkin (223S), but no significant recruitment of FLAG-parkin (223P). This was confirmed using GFP-parkin with either P or S at position 223 (Fig. 1C). Similar results were also obtained in COS7 or HEK293 cells using FLAG-, Myc-, HA- or GFP-tagged human parkin with S or P at amino acid 223 (data not shown). Fig. 1 Parkin sequence error has important functional consequence To test which version of parkin sequence is correct, we analyzed genomic DNA of 2102 individuals from Europe, China and the United States. None of them had C at this position; all were T. Briefly, genomic DNA from 231 normal subjects and 294 Parkinson’s disease patients in China, as well as genomic DNA from 19 normal subjects and 18 Parkinson’s disease patients in the United States were amplified by PCR using two primers in the introns flanking exon 6 of parkin (CTTGTCCAAAGAGATTGTTTACTGTGG and GGCTCGTGTGGCAGAACAATATTGGG). The PCR product (318bp) should contain a single BsrI site (CCAGT) if {type:entrez-nucleotide,attrs:{text:AB009973.1,term_id:3063387,term_text:AB009973.1}}AB009973.1 is correct. None of these samples could be cut by BsrI, while a positive control PCR product containing a BsrI site was cut at the same condition. Using a BLAST search of DNA sequences from 1540 individuals (1290 Parkinson’s disease patients and 250 healthy controls), mostly of French origin (~70%), we found that all of them had T at position 768; no one had C. Thus, the amino acid sequence reported in {type:entrez-nucleotide,attrs:{text:AB009973.1,term_id:3063387,term_text:AB009973.1}}AB009973.1 (Pro223) is wrong; it should be Serine instead, as reported in {type:entrez-nucleotide,attrs:{text:NM_004562.2,term_id:169790968,term_text:NM_004562.2}}NM_004562.2. This important error needs to be corrected, as it may affect other aspects of parkin biology, in addition to mitochondrial recruitment.