This piece of work is one of the discoveries from our lab ‘with the highest level of novelty, future significance and implications’. Here is how it came about. We were playing around with two backbone-reversed (retro) proteins that we’d made (i.e., engineered and produced from synthetic genes), both of which – we found – deposit out of aqueous solution under certain sets of similar conditions. Since we had some prior data suggesting that both retroproteins are rich in beta-sheet-forming potential, we were attempting to examine whether they could potentially co-aggregate out of solution. To do this, we labelled each protein with a different fluorophore and used a confocal microscope fitted with a fluorimetric attachment (Zeiss LSM 510 Meta) and a UV laser, to explore the interiors of protein aggregates formed from solutions containing both labelled proteins, to look for signs of either uniform colocalization of labels within all aggregates (suggestive of heteroaggregation), or differential localization of the two labels within otherwise ‘pure’ homo-aggregates. So far so good. When we performed the necessary control experiments with the unlabelled forms of these retroproteins, using the UV laser for excitation (351 and 365 nm lines) and expecting to see dark microscopic fields with aggregates of the unlabelled proteins, we saw instead that the protein aggregates glowed brightly with a brilliant blue-green radiation against a dark background field lacking such aggregates. We subsequently showed that the same radiation is seen with amorphous as well as amyloid aggregates, with protein crystals and with protein powders, that it is virtually undetectable with very small peptides, and that it can be seen in peptides and proteins lacking aromatic amino acid residues, as well as in all proteins and peptides longer than dipeptides and tripeptides irrespective of the source and manner of preparation. Initially, we thought that the UV laser was undergoing a Raman scatter of a Raman scatter, through scattering by water molecules immobilized within the aggregates and/or crystals, producing visible light through second-order scattering artefacts. However, this possibility was quickly excluded by showing that the 351 and 364 nm lines independently produce radiation with the same spectral maximum. We then showed that the radiation can be quenched by quenchers of fluorescence, photobleached and also recovered after photobleaching (FRAP), establishing that the radiation was actually fluorescence with non-aromatic origins. Through a process of rational exclusion of all other possibilities, and a leap of intuitive insight, we eventually reckoned that the fluorescence owes to peptide bonds which are hydrogen-bonded, i.e., to the C=O moiety within all peptide bonds engaged in H-bonding. The origins of the fluorescence, we argued, must lie in the excitation of the very same lone pairs of electrons which are present in the C=0 and N-H moietiesin the peptide bond (around the nitrogen atom in the N-H and the oxygen atom in the C=O). These lone pairs create a resonance within the peptide bond to give it a partially double-bonded character despite being a C-N bond, through their delocalization – away from their original nuclei. Ordinarily, in a peptide bond, or indeed within any kind of isopeptide or amide bond within a protein or organic molecule, the nitrogen-derived lone pair travels out into the C-N bond and then further out into the C=O bond, pushing the oxygen atom’s lone pair out and away from the peptide bond, to transiently create an oxyanion; this oxygen atom then reverts back to its original state, from its transiently-formed oxyanion state, with the pushing of the lone pairs occurring in the opposite direction, creating a resonance. We argued that the lone pair on the oxygen can jump across any H-bond formed by the oxygen, and that such delocalized electrons (held poorly to their original nuclei) undergo low-energy, long-wavelength electronic excitations close to the UV-visible boundary, with resulting fluorescence in the visible region. We also argued that if electrons can jump from one peptide bond to another, through the H-bonds formed between peptide bonds, the characteristics with which they do so must be governed by the type of secondary structure within a protein in which the peptide bond exists. Thus, peptide bonds in beta sheets, alpha helices, beta turns and random coils would be likely to exhibit different wavelengths of maximal fluorescence emission. We demonstrated that different proteins exhibit multi-banded fluorescence emission envelopes, with relative intensities of emission varying amongst four identifiable bands, from one protein to another. Further, we argued, if electrons from the oxygen atom of any peptide bond can jump to another peptide bond, then networks of alternating H-bonds and peptide bonds must act like ‘rivers’ along which electrons can flow, within proteins. Thus proteins must act as conductors of electricity, rather than as insulators of electricity, with the best such conductors being amyloid fibers which can act like electrical cables, and the most trivial such conductors being individual monomeric protein molecules which can allow fluxes of electron flows to occur between different parts of the same molecule, to alter electronegativity at a catalytic site; the important thing, to remember, being that the conduction occurs through networks of alternating peptide and H-bonds within secondary structures, and NOT within or through the polypeptide backbone. Linking the observation of fluorescence to the likelihood of its arising from delocalized electrons derived from C=O moieties which display a long-wavelength excitation transition close to the UV-visible boundary was not a trivial contribution; nor was the jump of insight linking the mechanism underlying the fluorescence to the possibility that proteins must conduct electricity, or the use of the fluorescence as an indicator of secondary structure. It is interesting that this work published with great difficulty in 2004 has slowly gained prominence. Today, in 2017, several groups around the world have confirmed it, agreed with the interpretation, demonstrated that proteins do indeed conduct electricity (with one group even using proteins to make MOSFETs for use in electronic devices). The observation of this fluorescence is listed on the IUPAC website as a means to non-invasively determine whether a crystal in a hanging-drop or sitting drop is a salt crystal or a protein crystal. The fluorescence has been used in FRET experiments to non-invasively detect amyloid deposits in small animals. The additional further implications of this work are quite huge. If proteins are conductors of electricity (even slightly so), instantaneous electron transfer within electron transport chains becomes immediately understandable and accessible to physical visualization, and begins to make some sense. So do the arrhythmias, and other conduction anomalies, that occur in electrical tissues such as the brain and the heart, upon the deposit of extracellular protein amyloids. So does the quantum entanglement of physically-separated organic moieties embedded within systems such as photosynthetic complexes. So does the observation that fluorescence microscopic experiments involving cells without the use of any labelled proteins or antibodies exhibit what is commonly called auto-fluorescence, even upon illumination with UV lasers, in which the entire ultrastructure of a cell (made of proteins) glows brightly without any obvious or immediate explanation. So does the auto-fluorescence that ophthalmologists routinely see and use as a measure of eye health, when they look down a slit lamp. The eye (and especially the lens) is full of proteins! The potential for the further expansion of possibilities ranges out even further. Someday, it might even emerge that the fluorescence and associated conduction of electricity is used by organisms to communicate signals (and various types of information, through the quantum world and involving quantum effects) between the different parts of an organism, or a cell. Finally, I would like to point out that this fluorescence has not been seen before (or seen and dismissed as autofluorescence) probably because it is of low quantum yield in individual protein molecules in solution, and best seen when proteins are clustered together in space, as in aggregates or crystals. Ultimately, even without networks of alternating H-bonds and peptide bonds, there could be some very low levels fluorescence seen in any molecule with C=O moieties; molecules which are capable of hydrogen bonding to water, or glycerol. Thus, any small molecule with a C=O moiety which can H-bond to an H-bonded solvent, or to any other molecule, might display very low levels of this fluorescence, because all that is needed is for the lone pair of electrons on the oxygen to travel out and stay out for a while, from where long-wavelength (low-energy) light can excite it, with deexcitation occurring through the fluorescence route (rather than through thermal or other modes of deexcitation).

Click to Read in Detail