Mitosis in the pennate diatom Surirella ovalis

Mitosis in Surirella is described; this organism displays a number of unusual features including an unorthodox method of chromosome attachment to the spindle, and the differentiation of an extranuclear central spindle from a large spherical organelle named the microtubule center (MC). The MC, present during interphase, breaks down at late prophase as the central spindle is formed. Later, the spindle enters the nucleus; the chromatin, in association with microtubules (MTs) from the poles, increasingly aggregates around the middle "overlap" region of the central spindle, and by metaphase completely encircles it. Throughout, MTs usually associate laterally with the chromatin. We were not able to identify kinetochore MTs with confidence at either metaphase or anaphase. Instead, at anaphase the leading point of the chromosomes is embedded in a ring of electrondense material, named the "collar," which encircles each half spindle and extends from the chromatin to the pole. Anaphase separation of the chromosomes is achieved by at least three separate mechanisms: (a) between metaphase and late anaphase the central spindle increases in length by addition of MT subunits; (b) at late anaphase the central spindle elongates concurrent with a reduction in the overlap; this apparently results from an MT/MT sliding mechanism; (c) each set of chromosomes moves to the poles by a thus far unknown mechanism; however, we anticipate some interaction of the collar and central spindle. At telophase, the polar complexes, (i.e., structures at the spindle pole) separate from the spindle, and later a new MC is formed near each polar complex, after which the polar complexes break down. Aspects of the complex differentiation of the MC, spindle formation, and some unusual characteristics of the diatom spindle as they relate to anaphase motion and spindle function are discussed. In 1896, Lauterborn (12) published a detailed account of nuclear and cell division in various diatoms, with emphasis on the pennate diatom Surirella. He noted several unusual features characteristic of mitosis in these organisms, which were not later confirmed by workers using the light microscope (see Discussion). More recently, Manton et al. (14-17) observed with the electron microscope that the spindle of the centric Lithodesmium does indeed have unusual characteristics. They illustrated in considerable detail the formation of the extranuclear "central spindle" which runs from one pole to the other, and indicated that by metaphase it consists essentially of two halfspindles, each a highly ordered set of microtubules (MTs) which interdigitate in the central "overlap" region. They also demonstrated that the centriole involved in flagellum formation during sexual reThE JOURNAL OF CELL BXOLOGY 9 VOLUlaE 73, 1977 9 pages 705-727 705 production forms de novo at each pole. We have documented all stages of mitosis and cytokinesis in two diatoms; several further interesting observations were made, including that the central spindle of Diatoma (20) elongates during anaphase, concurrent with a reduction in the overlap region. Additionally, we have shown in Melosira (28) that the overlap region during prophase consists of a very precise arrangement of MTs; one MT from one pole is usually surrounded by four from the other pole. These observations together clearly hinted that sliding of parallel/antiparallei MTs in the overlap could be responsible for generating that part of anaphase motion due specifically to that spindle elongation. Further work on the structural analysis of the Diatorna spindle 1 confirms that sliding generated by such MT/MT interaction provides the most logical and satisfying explanation for the reduction of the overlap during spindle