Abstract
The conformation dependence of protein spectra recorded by electrospray ionization mass spectrometry (ESI‐MS) is an interesting and useful phenomenon, whose origin is still the object of debate. Different mechanisms have been invoked in the attempt to explain the lower charge state of folded versus unfolded protein ions in ESI‐MS, such as electrostatic repulsions, solvent accessibility, charge availability, and native‐like interactions. In this work we try to subject to direct experimental test the hypothesis that conformation‐dependent neutralization of charges with polarity opposite to the net charge of the protein ion could play a critical role in such an effect. We present results of time‐of‐flight nano‐ESI‐MS on the peptide angiotensin II, indicating that negative charges of carboxylate groups can contribute to spectra recorded in positive‐ion mode when stabilized by favorable electrostatic interactions, which is the central assumption of our hypothesis. Comparison of horse and spermwhale myoglobin (Mb) shows that changing the total number of basic residues within a given three‐dimensional structure shifts the charge‐state distribution (CSD) of the folded protein in positive‐ion mode. This result appears to be in contrast to models in which electrostatic repulsions or availability of charges in the ESI droplets represent the limiting factor for the ionization of folded protein ions in ESI‐MS. At the same time, it suggests a role of acidic residues in conformational effects in positive‐ion mode. Furthermore, an attempt is made to rationalize those cases in which, in contrast, the main charge state observed in ESI‐MS under non‐denaturing conditions deviates considerably from the net charge expected on the basis of the amino‐acid composition. These cases usually correspond to proteins with quite balanced content in basic and acidic residues, suggesting that this might be a factor influencing their charging behavior in ESI‐MS. Experiments on mutants of ribonuclease Sa (RNase Sa) reveal that progressively reducing the excess of acidic residues, replacing them by lysine, causes almost no shift in the spectrum of the folded protein in negative‐ion mode. Analogously, variants with an excess of three or five basic residues give similar spectra in positive‐ion mode. These results indicate a lower limit to the extent of ionization observable by ESI‐MS (6− or 8+ in the case of RNase Sa in water). Below such limit of net charge, changes in the relative amount of ionizable side chains do not affect the qualitative features of the observed CSDs. A progressive loss of signal intensity caused by the mutations in negative‐ion mode suggests that low charge states might also be counterselected, even within the m/z range theoretically accessible to the instrument. Copyright © 2003 John Wiley & Sons, Ltd.