送交者: Godfather 于 2006-1-23, 18:01:32:
With previous questions and concerns raised here in mind, I’d like to pinot out some evidence that makes Qiu's paper more suspicious for fabrication .
1. Is it possible to have desired specific activity by fusing Colicin Ia with pheromone? Based on my assessment, the answer is negative. The crystal structure of Colicin Ia was solved by M. Wiener et al in 1997 (Weiner, M. et al, 1997, Nature, 385(6615), 461-464). According to the crystal structure, in addition to the previous studies and knowledge on its structure and function, the 3 functional domains and the helices at both N- and C-termini were well presented and consistent with earlier experimental observations. In this paper, for the first time, it showed that the C-terminal channel-forming domain (called as C domain) consists of 10 helices, which interacts with membrane to form ion channels. However, helix C10 at the end of C-terminus was later identified to stay so called cis-side (outside of the membrane) by Dr. Qiu’s former lab at AECOM (Kienker P. K., Jakes K. S. and Finkelstein A. 2000, J. Gen. Physiol. 116(4), 587-597). They stated “Aside from the hydrophobic hairpin formed by helices 8 and 9 of the C domain, the rest of the molecule is highly charged (>30% of the residues)”. Now, the question is when Qiu added 8 more residues of AgrD3 (YINCDFLL) to the C-terminus, what would happen to the C10 helix? Based on the assumption and statement made by his postdoc advisor, the C10 helix should still stay at the cis side since N and D are charged residues among the 8 additional residues (25%). If that’s the case, how does the SA specific AgrD3 contact its target (SA membrane) while maintaining its channel forming function? On the other hand, if the additional 8 residues caused conformational change of the C domain, then how would C1-C7, C8 and C9 interact with membrane phospholipids to form ion channels? Obviously, from protein structure point of view, it makes no sense to have the desired specificity and activity as described and claimed by Qiu.
2. Nobody has questioned his data and statement on the identification of peptide fragments thus far. Actually, in my opinion, that’s the obvious part to be identified as fabrication. In his paper, he stated that he used LC/MS to identify 2 peptides to prove the existence of thiolactone that is essential for the activity of pheromones. However, he did not give enough experimental details for LC and MS conditions. As for the LC, nobody knows the instrument, solvent system, column (only said C18), flow rate, gradient, injection volume, etc. So does the MS, such as voltage for electron spray ionization. More importantly, nobody knows if he separated the peptide fragments by LC first or he used direct online ionization for various peptide peaks from the LC, which typically results in peptide mixtures, especially for his 2 peptides with similar hydrophobicity. Most importantly, LC/MS method could not give an accurate mass number with 2 decimals unless MALDI was used, and LC/MS generally gives larger variations than MALDI. Now, let’s take a look at his data. The first peptide (15 residues), calculated value of 889.42 vs. observed value of 889.54; the second peptide (12 residues), calculated value of 732.83 vs. observed value of 732.40. They were amazingly accurate with deviations of 0.013% and 0.059% respectively, which was better than MALDI for the determination of masses of biomolecules. I would suggest Dr. Qiu to patent such LC/MS method and submit a manuscript to Science or Nature for the accurate measurements, which may turn out to be a money tree. As a side note, since trypsin cleaves the peptide bonds at the carboxyl side of every single lysine and arginine, it’s highly impossible to obtain the peptide with 15 residues (ANKFWGIYSTCDFIM) in which there still is a lysine residue after tryptic digestion overnight at 37 degree.
In summary, combining previous speculation or suspect on the issues of streptomycin, activity spectrum, and specificity, I present new speculation on structural biology and LC/MS determination of peptides in this post to further support the my conclusion.