Photoreactions and photocycles

In free retinal, illumination will produce a mixture of several single- and doublebond isomeric configurations. Protein-bound retinal is more restricted in its motional possibilities in the rhodopsins the binding pocket is tailored to fit only a few specific retinal configurations. Indeed, the light-initiated reaction cycles of the bacterial rhodopsins are based exclusively on dlX-trans to 13-cw isomerization of their retinal chromophores. When under some conditions other isomers are produced upon illumination, such as 9 cis, they accumulate in limited quantities [90-92]. 

Absorption of a photon by the W-trans chromophore produces an excited state which relaxes within 500 femtoseconds to the J intermediate (in the case of bacteriorhodopsin at least), in which rotation aroimd the Cb-Ch double bond had occurred [93,94]. It has been suggested that there is isomerization also around the Cu-Cb single bond [95], but others dispute this [96]. All further intermediates are thermal (dark) products of J. Some reports distinguish, on the basis of their absorption maxima, between an early K 

Many reports show this photocycle as a linear unidirectional scheme. 

This is the simplest explanation for the observation that when L and M have come to an equilibrium which contains these species in comparable amounts, the concentration of L decreases to near zero even while M remains at its maximal accumulation. Recent measurements of the quasi-equilibrium which develops in asp96asn bacteriorhodopsin before the delayed reprotonation of the Schiff base confirm this kinetic paradox [115]. Two M states have been suggested also on the basis that the rise of N did not correlate with the decay of M [117]. In monomeric bacteriorhodopsin the two proposed M states in series have been distinguished spectroscopically as well [115]. It is well known, however, that kinetic data of the complexity exhibited by this system do not necessarily have a single mathematical solution. Thus, assurance that a numerically correct model represents the true behavior of the reaction must come from testing it for consistencies with physical principles. It is encouraging therefore that the model in Fig. 5 predicts spectra for the intermediates much as expected from other, independent measurements, and the rate constants produce linear Arrhenius plots and a self-consistent thermodynamic description [116]. 

The proposed halorhodopsin photocycle is shown in Fig. 6. In this protein the Schiff base of L does not deprotonate (except as a side reaction, cf. below), but a chloride-dependent equilibrium is established, within a few milliseconds, between L and a red-shifted 0-like state [120-123]. From the kinetics it is inferred that the 0-+L back-reaction is accompanied by uptake of a chloride ion [120,121,123]. Decay of O by a non-reversible reaction regenerates halorhodopsin in tens of milliseconds, but to balance the loss of the chloride it is assumed that first the chloride-free form, HR, is formed, which produces the original state after uptake of chloride on a time scale too rapid to measure. In the absence of chloride the L state is not observed [55,121] it appears that O is directly produced in a truncated photocycle. In the presence of nitrate, an anion poorly transported by halorhodopsin from H. halobium, both photocycles are seen simultaneously. Halorhodopsin from Natronobacterium pharaonis transports nitrate as well as chloride, and in this system O cannot be detected in the photocycle [54]. 

---