Duane L. Marcy

Volumetric optical memory based on bacteriorhodopsin

An architecture for a three-dimensional optical computer memory based on the photoactive protein bacteriorhodopsin (BR) is described, utilizing a branch-off of the BR photocycle to access a long-lived photointermediate capable of serving as an active element for memory storage. This intermediate (the Q-state) is accessed as a result of the sequential absorption of two photons, the first to initiate the BR photocycle, and the second to drive the protein into the branched photocycle from the O-state several milliseconds later. The stability of the Q-state arises predominantly from the fact that it is strongly blue-shifted with respect to other intermediates in the photocycle, making it invisible to the laser wavelengths used to write and read information in the memory. Both proof of principle and second-generation prototypes are currently being developed in the W.M. Keck Center for Molecular Electronics, in collaboration with Critical Link, LLC of Syracuse, NY. The article will focus on the BR-branched photocycle memory architecture, the remaining challenges to fabrication of a commercially viable device, and the ongoing efforts in prototype development, optimization and protein characterization at Syracuse University.

Monolithically integrated bacteriorhodopsin-GaAs field-effect transistor photoreceiver.

We have applied the large photovoltage developed across a layer of selectively deposited bacteriorhodopsin to the gate terminal of a monolithically integrated GaAs-based modulation-doped field-effect transistor, which delivers an amplified photoinduced current signal. The integrated biophotoreceiver device exhibits a responsivity of 3.8 A/W. The optoelectronic integrated circuit is achieved by molecular-beam epitaxy of the field-effect transistors heterostructure, photolithography, and selective-area bacteriorhodopsin electrodeposition.

Optimization of protein-based volumetric optical memories and associative processors by using directed evolution

The potential use of proteins in device applications has advanced in large part due to significant advances in the methods and procedures of protein engineering, most notably, directed evolution. Directed evolution has been used to tailor a broad range of enzymatic proteins for pharmaceutical and industrial applications. Thermal stability, chemical stability, and substrate specificity are among the most common phenotypes targeted for optimization. However, in vivo screening systems for photoactive proteins have been slow in development. A high-throughput screening system for the photokinetic optimization of photoactive proteins would promote the development of protein-based field-effect transistors, artificial retinas, spatial light modulators, photovoltaic fuel cells, three-dimensional volumetric memories, and optical holographic processors. This investigation seeks to optimize the photoactive protein bacteriorhodopsin (BR) for volumetric optical and holographic memories. Semi-random mutagenesis and in vitro screening were used to create and analyze nearly 800 mutants spanning the entire length of the bacterio-opsin (bop) gene. To fully realize the potential of BR in optoelectronic environments, future investigations will utilize global mutagenesis and in vivo screening systems. The architecture for a potential in vivo screening system is explored in this study. We demonstrate the ability to measure the formation and decay of the red-shifted O-state within in vivo colonies of Halobacterium salinarum, and discuss the implications of this screening method to directed evolution.

Evaluation of blue and green absorbing proteorhodopsins as holographic materials.

Transient holographic diffraction is observed for the green (GPR) and blue (BPR) absorbing proteorhodopsins (BAC31A8 and HOT75M1, respectively), as well as the GPR E108Q and BPR E110Q variants. In contrast to bacteriorhodopsin, where the metastable bR-M pair is responsible for generating diffraction, the pR and red-shifted N-like states fulfill that role in both the green and blue wild-type proteorhodopsins. The GPR E108Q and BPR E110Q variants, however, behave more similarly to their bacteriorhodopsin analogue, D96N, with diffraction arising from the PR M-state (strongly enhanced in both GPR E108Q and BPR E110Q). Of the four proteins evaluated, wild type (WT) GPR and GPR E108Q produce the highest diffraction efficiencies, etamax, at approximately 1% for a 1.7 OD sample. GPR E108Q, however, requires 1-2 orders of magnitude less laser intensity to generate eta equivalent to WT GPR and BR D96N under similar conditions (as compared to literature values). WT BPR requires lower actinic powers than GPR but diffracts only about 30% as well. BPR E110Q performs the most poorly of the four, with etamax < 0.05% for a 1.4 OD film. The Kramers-Kronig transformation and Kogleniks coupled wave theory were used to predict the dispersion spectra and diffraction efficiency for the long M-state variants. To a first approximation, the gratings formed by all samples decay upon discontinuing the 520 nm actinic beams with a time constant characteristic of the appropriate intermediate: the N-like state for WT GPR and BPR and the M-state for GPR 108Q and BPR E110Q.

Biomolecular Electronic Device Applications of Bacteriorhodopsin

Over the last thirty years, bacteriorhodopsin has become one of the most actively researched proteins in biochemistry and biophysics. Interest in this protein stems not only from its unique photochemistry as a light-driven proton pump, but also from its potential as an active component of biomolecular device applications. Architectures for devices that range in form and function from polymer film based holographic interferometers to three-dimensional optical memories have been proposed and built. Bacteriorhodopsin’s propensity for biomolecular optics and electronics is due to three main light-induced responses: (1) A light-induced photocycle that consists of a variety of photochemically-distinct intermediates characterized by shifted absorption maxima, (2) A branched photocycle that results in a permanent blue shifted intermediate, and (3) A complex time-dependent photoelectric response that varies in both sign and amplitude over several orders of magnitude. These properties will be discussed below, as relevant to the development of two specific bacteriorhodopsin applications: the branched photocycle three-dimensional optical memory and the holographic associative memory.

An FPGA-based MOS circuit simulator

A novel method for performing SPICE-like circuit simulations on a field programmable gate array (FPGA) is presented. An RC filter and an NMOS transistor have been simulated as demonstration examples. The presented technique utilizes a digital signal-processing object (SPO) that is analogous to an analog operational amplifier. An array of SPOs solves nonlinear difference equations in the same way that an array of opamps solves nonlinear differential equations. The SPO is first designed in Simulink and then emulated in the Xilinx system generator. This model is then programmed onto an FPGA, Spartan-3, e.g. The simulation result is virtually identical to a standard SPICE simulation, but simulation speed is increased by more than an order of magnitude, as compared to simulations running on typical workstations. This is because a typical FPGA holds many SPOs, all running in parallel at a high clock rate. Advantages are: 1) speed of execution with little or no degradation in the accuracy, 2) no parallel programming step since the SPO interconnections are specified by the difference equations and 3) the resulting FPGA can act as a high speed attached processor to the workstation.

Three-dimensional data storage using the photochromic protein bacteriorhodopsin

A three-dimensional computer memory architecture based on the optical properties of the integral membrane protein bacteriorhodopsin has been previously proposed. However, confirmation of the very important branched photocycle property of this protein has not been shown. Data confirming the existence of a branched photocycle by writing a long term stable bit of information within the center of a cube of protein is presented. The stable photoproduct is formed by pulsing a 635 nm laser followed by an intersecting 690 nm laser pulse. At the point of intersection a stable branched photo-product appears. We found that at elevated temperatures the rate of production of the stable photoproduct is greatly increased and therefore the experiments are performed at 40/spl deg/C. Data representing absorbance changes at 690 nm due to the bacteriorhodopsin photocycle at 40/spl deg/C is presented. This information is required to set the timing of the light pulses required to write a bit. Then the pulse-pulse experiment is performed with a corresponding increase in transmission observed during the 690 nm light pulse. This confirms the production of the stable photoproduct. The experiment is then repeated with both lasers pulsing simultaneously. If the photoproduct is formed from a branch off of the photocycle, the simultaneous pulses should form much less photoproduct, despite an equivalent number of photons. Data showing this result is presented confirming the branched production of the stable photoproduct, thus proving the three dimensional computer memory architecture.