Volume 77, no. 12, p. 7078-7092, 2003. (Paper no. 1)
Volume 80, no. 5, p. 2291-2308, 2006. (Paper no. 2)
The following applies to both of the above articles.
From the data presented in the two papers cited above, we concluded that the YRKL sequence of influenza virus M1 functions as the L domain motif and interacts with the VPS28 and Cdc42 host proteins. However, one mutant virus containing a PTAP replacement of YRKL and another insertion mutant virus containing YRKLKRE at residues 51 and 52 and AAAA at residues 100 to 103 of M1 were sent to another lab and were found to contain only wild-type virus. Upon learning this information, we reexamined essentially all (more than 50) mutant viruses reported in these two papers. We found that, in addition to these two virus samples, the YPDL mutant (Fig. 8, paper no. 1) and mutants with the PTAP insertion mutations (4A+PTAP and WT+PTAP, Fig. 8, paper no. 1), the mutants with YRKL insertion mutations at positions 1-2, 51-52, 99-100, 103-104, 107-108, and 160-161 (Table 2, paper no. 2) as well as the mutants with YRKLKR, YRKLKRE, KLYRKLKR, and VKLYRKLKR mutations at position 51-52 (Fig. 4, paper no. 2) also contained the wild-type virus genome. In addition, immunoprecipitation data with the PTAP mutant presented in Fig. 7C of paper no. 2 could not be reproduced.
We therefore retract both papers and deeply regret any inconvenience that the errors may have caused.