MT Signal ProcessingDeBrabander

Cell biologist Marc DeBrabander has extensively studied activities of the cytoskeleton in mitosis and cellular organization in general. He and his colleagues at Belgium's Janssen Pharmaceutica Research Laboratories (DeBrat bander, DeMey, VandeVeire, Aerts and Geuens, 1975, 1986) have examined the distribution and movement of cell membrane proteins and gathered evidence that they are linked to the cortical actin filament system. In turn, the ordered activity of the filaments appears to be controlled by the microtubule system, analogous to its function in cell movement and cell shape. DeBrabander (1977) observes that in these phenomena MT act beyond their capacity as skeletal support elements and perform as signal transducers from the cell center to the periphery, and vice versa. Interactive signaling in the cytoskeleton fits with observations that local events at one site in the cell often lead to a global cellular rearrangement. DeBrabander raises several MT features which could favor signaling: MT intrinsic polarity, the existence of "centrifugal and centripetal microtubules," the conformational propagation mechanism proposed by Atema (1973), directional organelle movement and topography, and transport of ions such as calcium along microtubules.

DeBrabander and his associates (1986) have also innovated new techniques in the examination of cytoskeletal structure and dynamics. These include "Nanovid Microscopy," a nanometer scale video microscopy in which transport of labeled particles along individual MT may be visualized, and a method of labeling individual tubulin subunits within MT which are either tyrosinated or glutamated (Geuens, Gundersen, Nuydens, Cornelissen, Bulinski, DeBrabander, 1987). As described in Chapter 5, polymerized MT may be detyrosinated in the cytoplasm subsequent to DNA directed genetic input. Tyrosine, the terminal amino acid in tubulin, may be removed to expose glutamate as the new terminal amino acid. Thus all tubulins within polymerized MT are either tyrosinated, or glutamated. A variety of cytoplasmic factors determine whether or not a particular tubulin is tyrosinated. The precise significance or function of tyrosination/glutamation is unknown but could serve as a convenient programming or memory function. DeBrabander and colleagues developed a double labeling technique in which immunogold particles bind to tubulin subunits. Large immunogold particles (10 nanometers) identify glutamated tubulin and smaller particles (5 nanometers) bind to tyrosinated tubulin. Patterns of tyrosinated/glutamated tubulin within MT show heterogenous distribution, and suggest the potential for a coding mechanism. MAP attachment, tubulin conformation, calcium binding and other factors could be coupled to MT function via such patterns. Although predisposition to detyrosination may be genetically linked to tubulin isozymes, the actual patterns of tubulin detyrosination are determined by "real time" cytoplasmic factors. DeBrabander and colleagues have provided the first direct evidence of modifiable patterns of tubulin variability in intact microtubules. Consequently, MT appear capable of signal processing in addition to signal transduction.

Figure 8.1: Microtubules from PtK2 cells double labeled with antibody and immunogold under high magnification. Large circles (10 nanometer gold particles, arrows) tag glutamated tubulin subunits, small circles (5 nanometer gold particles) label tyrosinated tubulin subunits. Left: interphase microtubule, Right: spindle microtubule. With permission from Geuens, Gundersen, Nuydens, Cornellisen, Bulinski, and DeBrabander (1986), courtesy of Marc DeBrabander and Janssen Pharmaceutica Research Laboratories.

Figure 8.1: Microtubules from PtK2 cells double labeled with antibody and immunogold under high magnification. Large circles (10 nanometer gold particles, arrows) tag glutamated tubulin subunits, small circles (5 nanometer gold particles) label tyrosinated tubulin subunits. Left: interphase microtubule, Right: spindle microtubule. With permission from Geuens, Gundersen, Nuydens, Cornellisen, Bulinski, and DeBrabander (1986), courtesy of Marc DeBrabander and Janssen Pharmaceutica Research Laboratories.

8.2.5 Cytoskeletal String Processors/Barnett

Figure 8.2: Barnett's string processor model of microtubule-neurofilament memory storage. 1) a processing channel (microtubule-MT) is laterally connected to a parallel storage unit (neurofilament-NF), 2) information traverses the processing channel from right to left; on the storage channel, synonym pairs move from left to right, 3) as the word "chips" moves by its own self in storage, it is deleted, 4) "chips" is replaced by its synonym "french fries" and 5) the replacement is complete. By Paul Jablonka.

Figure 8.2: Barnett's string processor model of microtubule-neurofilament memory storage. 1) a processing channel (microtubule-MT) is laterally connected to a parallel storage unit (neurofilament-NF), 2) information traverses the processing channel from right to left; on the storage channel, synonym pairs move from left to right, 3) as the word "chips" moves by its own self in storage, it is deleted, 4) "chips" is replaced by its synonym "french fries" and 5) the replacement is complete. By Paul Jablonka.

Brooklyn College computer and information scientist Michael P. Barnett has pursued the design of computer components suited to molecular scale devices. In addition to the development of tiny binary switches to be assembled into circuits on digital architecture, Barnett (1987) has also sought an associative memory which can analyze imprecise analog data as well as conventional digital information. Fabricated systems with these features could be versatile, powerful and a boon to the goals of artificial intelligence. Like many AI oriented researchers, Barnett turned to biology for clues. However, unlike most Al researchers, Barnett scoured the subcellular biological realm in search of molecular scale information processing concepts. His fancy was captured by cytoskeletal microtubules and neurofilaments! He has proposed information representation as patterns in the subunits of cytoskeletal polymer subunits. He proposes that specific subunit states may be characterized by "electrons transferred from delocalizable molecular orbitals," but his basic premise would also be supported by other causes of conformational state variability among MT and neurofilament polymer subunits. Barnett suggests that filamentous cytoskeletal structures operate like information strings analogous to word processors.

In Barnett's conceptualization (Figure 8.2), information strings move from right to left along processing channels, which run parallel to one dimensional memory channels in which character strings can be stored. Barnett's string transformers can perform global replacements on sequences of characters like common word-processors. His model assumes the existence of processing channels (MT) along which strings of information can move, and memory channels (neurofilaments) which consist of a succession of locations, each of which can hold a single character. Parallel array and lateral interconnectedness of MT and neurofilaments could qualify these cytoskeletal elements as string processors, assuming that information may be represented in the polymers. Barnett's model is thus compatible and complementary with other models of conformational patterns within MT and the cytoskeleton.

0 0

Post a comment