Like a superhero with a secret identity and hidden powers, ordinary matter has normally unseen properties. For example, all matter is both particle and wave — a consequence of quantum mechanics. However, matter’s wavelike nature is hard to detect for anything as big as a molecule because the mass-dependent wavelength is so short. Now a team from the University of Toronto, the University of Guelph in Ontario, and the University of Connecticut in Storrs has used shaped laser pulses to enhance or suppress the photoisomerization yield of the retinal molecule bacteriorhodopsin as it transforms from one state to another. University of Toronto professor of chemistry and of physics R.J. Dwayne Miller explained that the experiments tested for quantum effects such as wave coherence in biological functions. “We have shown that quantum coherences can play a role to the point that we can even manipulate them,” he said. At first glance, there would not seem to be any role for quantum effects in biology. The wavelength for a massive particle like a protein is less than a thousandth of a single bond length between two carbon atoms and, therefore, is much smaller than the functionally relevant protein motions. Likewise, the coherence time of the wavefunction describing a protein is much less than the time scale of a protein’s biological motions. So quantum effects average out and disappear. However, biological reactions ultimately often depend on the localized movement of a few active site carbon atoms. That fact allows quantum effects to play a part. The researchers picked the photoisomerization of retinal bacteriorhodopsin for this study for a number of reasons. It is a fast reaction — taking place within 500 fs — and the structural change, the photoisomerization, occurring between the 13th and 14th carbon atoms of the retinal’s carbon backbone. The bond between the atoms softens or stretches, followed by twist about the bond axis. The reaction time is comparable to the coherence time of the excited wave function, and the reaction involves only key carbon atoms, not the protein. Moreover, Miller noted that there are many retinal photoisomers in solution, each with only a few percent quantum yield. In a protein environment, on the other hand, retinal has only one photoisomer and it has a quantum yield of 65 percent. The protein structure, therefore, limits possible motions, suggesting that quantum effects won’t be overwhelmed. “This system in some sense is the acid test for any quantum coherence effects involved in a biological function,” Miller said. Technological requirements In carrying out their investigation, the researchers required some technological advances. They needed very short pulses spanning the wavelengths used by bacteriorhodopsin, and they needed the ability to shape those pulses. Most of their key equipment was custom-built, with the exception of a programmable acousto-optic modulator from Fastlite of Paris that was used for phase and amplitude manipulation of the pulses. Vital to the success of the research was the energy stability of the beam. It had to supply an excitation energy of nanojoules while maintaining stability of about half a percent. The researchers achieved this with a diode-pumped solid-state Nd:YLF laser pumping a regenerative Ti:sapphire amplifier delivering 150-fs pulses at 775 nm. These pulses pumped a noncollinear optical parametric amplifier to produce final output pulses centered at 575 nm with a full width at half maximum of 65 nm. The researchers shaped these pulses using the acousto-optic modulator and kept them at a constant, or actinic, energy. They flowed bacteriorhodopsin in solution through a flow cell past the actinic pulse, which they focused down to 150 μm. Some 20 ps later, they measured its effect by targeting a 630-nm absorption feature of the photoisomerization product. By measuring the absorption difference and weighting that by the number of photons absorbed, they had a direct readout of the isomerization yield. They used a white-light continuum generated by the laser pulse for a light source and a calibrated photodiode/spectrometer combination detector for these readings. They began with a starting set of 30 actinic pulses, with a random distribution of spectral phase and amplitude. They fired these into the flow cell, read out the isomerization yield and applied a genetic algorithm to the pulses so that they evolved through breeding and mutation toward higher yield. They repeated the process multiple times, eventually running more than 20 optimization experiments. They then selected for lower yield. Again, they did multiple runs. During all of these experiments, they kept the excitation energy constant to within 1.5 percent and ensured that the energy was too low (15 nJ) to allow different excited electronic states to be reached. The pulse intensities were such that the molecule could access the same states as those attained naturally. As described in the Sept. 1 issue of Science, the researchers used these methods to boost the yield 23 percent or decrease it by a like amount. According to Miller, the changes were consistent with constructive and destructive interference effects along the reaction coordinate of the large molecule, and the final shape of the optimization pulses wasn’t random. “The optimal pulse has a very distinctive feature that corresponds exactly to the frequency of the torsional motion that is key to isomerization,” he said. He noted that the optimal pulse drove a polarization that resonantly coupled to this motion. The optimal pulse was centered at 557 nm with regular structures spaced 6 nm apart. In contrast, the antioptimal pulse lacked these features, was centered at 577 nm and was relatively broad. Because of its characteristics, the antioptimal pulse was out of resonance with this key motion. The researchers also found that the protein’s isomerization efficiency was slightly higher for a pulse of random phase and amplitude than for one that was smoothly varying. This they attributed to a significant built-in bias for the reaction due to a protein structure. This bias, Miller said, is indicative of a protein structure that fully exploits random spectral content fluctuations in low light levels to couple into the motions involved in photoisomerization and thereby fend off the relaxation of the excited molecule back to the ground state. The scientists plan to use pulse shaping methodologies to selectively excite different types of vibrational motions. That will allow them to effectively explore the landscape of the protein and understand how the protein directs events so efficiently to one particular desired reaction out of all the many other possibilities. “Biology will teach us a great deal about how to control chemistry,” Miller said.