R. Malcolm Brown, Jr.


Presented at the 3rd Philip Morris Science Symposium, Richmond, Virginia, 1978, November 9. Published in Structure and Biochemistry of Natural Biological Systems; Proceedings of the 3rd Philip Morris Science Symposi7m; edited by Ellen M. Walk, New York. Philip Morris Incorporated. 1979. Pp. 51-123. (Reprinted on the Internet with permission from Philip Morris, Incorporated).

Note: This review, published almost 20 years ago, was extremely predictable of what actually has transpired in the cellulose field. Thus, it was desirable to reprint this article here in its entirety because it is not readily available to the general public. If there are any questions, please contact Dr. Brown directly at ( or contact Marian Z. DeBardeleben, Section Leader, Information Services Development and Support, Philip Morris, U.S.A., Research Center, P.O. Box 26583, Richmond, Virginia 23261-6583 (tel 804 274-2000).

The author expresses his gratitude to Philip Morris for permission to publish this article in its entirety on the Web.

Important note: All photos are in separate htm files labeled the same as the figure number. When reading through the text, simply click on the figure number and you will bring up a particular figure. To return to the point in the text where you clicked for a particular figure, simply use the BACK function on your browser. If you want to return to the beginning of the document, go to the bottom of the figure page and click on the return link. All figures are in JPG format to keep the file size small. The horizontal size is set to 700, so you may want to adjust your screen width accordingly. Good Luck! RMB! October, 1997.



Ladies and gentlemen, fellow scientists, Dr. Scott, Dr. Seligman, and Dr. Wakeham, it is a distinct honor and pleasure for me to be here today to participate in this, the Third Philip Morris Science Symposium. I recall the happy incident of meeting Dr. Wakeham just over a year ago in Raleigh at the Fiber Society Meetings. It was on this occasion that Dr. Wakeham heard my presentation on cellulose and suggested that I contribute to today's symposium. I am happy, indeed, to have the opportunity to talk about a subject of great interest to me personally. But before I discuss specific examples, let me diverge for a moment and review briefly the composition and structure of cellulose.

First, cellulose is a natural biological homopolymer consisting of cellulose residues glycosidically linked in the (1 4)- configuration (Fig. 1). Bundles of glucan chains associate via hydrogen bonding to form an insoluble ultramicroscopic thread known as the microfibril) (Fig. 2). Cellulose is the most abundant macromolecule on earth. Hess' has estimated that more than 1011 tons of cellulose are produced and destroyed annually Obviously, cellulose is a major contribution to our carbon budget. The value given by Hess may be too large. My colleague, Helmut Lieth, an expert in world biomass production, has suggested that 1011 tons may exceed the total production on earth! He has proposed a lower estimation of annual cellulose biomass production of l09 tons. Thus, please lower your expectations and visions accordingly! Nevertheless, we are talking about a considerable amount of cellulose!

Where is cellulose made? The holistic approach to this inquiry would imply that cellulose is everywhere, including outer space. Imagine cellulose as the most abundant polymer in the universe! Professor Fred Hoyle and his colleagues have acquired evidence to support this hypothesis.2 they have suggested the abiotic assembly of cellulose in outer space from the polymerization of formaldehyde. Evidence for the existence of cellulose comes from the close match of an infrared absorption spectrum of celestial origin with that of cellulose. Fig. 3 shows the relationship between the infrared absorption spectra from the Trapezium Nebula and cellulose.

Returning to our galaxy, solar system, and planet, it is all the more interesting to discover that cellulose is synthesized by a diverse collection of living organisms. In fact, not only do plants synthesize cellulose, but also animals can assemble this polymer. One good example is the tunicate (Fig. 4)3 . These animals are chordates, and if one considers that multicellular animals of this evolutionary advancement have the capacity to synthesize cellulose, it is not inconceivable that mammals, including man, could also have this capacity (Fig. 5)). Only a few scattered reports have suggested cellulose synthesis in humans, and only under pathogenic conditions.4 The disease scleroderma is believed to involve activation of cellulose synthesis in humans. Why don't we know more about this fascinating diagnosis? Because medical scientists have not been interested in cellulose synthesis in plants, and botanists are largely unaware of the process in humans! We need to bring together botanists and health scientists, and at Chapel Hill, we have initiated a joint program to investigate cellulose synthesis in humans. It is too early yet to make any definitive statements on human cellulose production, but I anticipate that it will be a lively and active research topic in the near future.

Cellulose synthesis is found not only among plant and animal systems but also among eukaryotic and prokaryotic cells. A beautiful example of prokaryotic cellulose synthesis is the Gram-negative bacterium, Acetobacter xylinum. Anyone who has attempted to produce homebrew wine without first treating the must with sulfur dioxide knows that a leathery film will appear on the surface and that fermentation to acetic acid usually will destroy the wine. The surface film is a pellicle of cellulose.

Cellulose synthesis has reached its greatest diversification among eukaryotic cells. Non-photosynthetic cells can synthesize cellulose (Fig. 6). These include certain fungi, amoeba cysts, and the cellular slime molds, of which Dictyostelium discoideum is a famous example.5 Cellulose synthesis has achieved a great diversity among the photosynthetic eukaryotic cells.6 The algae have experimented with many different natural polymer systems, and today we shall examine two unique and beautiful examples of cellulose synthesis in Oocystis apiculata and Glaucocystis nostochinearum

Progressing from the algae to the more complex multicellular land plants, cellulose has assumed the primary role as a major reinforcing polymer. The mixture of lignin with cellulose has given great strength and allowed development of extremely large multicellular plants. The redwoods are perhaps the most famous of the giants. What is the relationship of the cellulose microfibril to the structural framework of the plant cell? I mentioned earlier that the basic molecular unit is the glucan chain and that these aggregate into the microfibril. Microfibrils are synthesized at the cell surface, and together with noncellulosic polysaccharides and proteins7 they comprise a cell coat or extracellular matrix known as the cell wall. Long after the plant dies, the cell wall remains. In 1665, Robert Hooke 8 examined such dead products of cork and named them 'cellulae" (Fig. 7). Cellulosic walls comprise some of our major products of commerce. Cotton fibers are epidermal cells of the ovule, dead at maturity. Secondary wall thickenings of xylem tissue comprise the products so familiarly known as wood. The lignin can be removed from xylem tissue, and the cellulose pulped, bleached, and pressed into paper. Cellulose is of major interest in tobacco research, not only as a part of the tobacco leaf cells, but also as a component of filter material.

The central enigma of cellulose revolves around the mechanisms of its biosynthesis. For all that we know about the structural and compositional aspects of the cellulose microfibril, we know essentially nothing about how the living cell assembles this polymer. We know even less about how the cell controls and regulates the polymerization of glucan chains, the size of microfibrils, the orientation of microfibrils, and the expansion of microfibrils during cell differentiation and growth. I t is these aspects which I would like to deal with today.

During the past thirteen years, I have used various model systems in plant cell biology (Table I). There are many questions to ask about cellulose synthesis, and no ideal single system can provide answers to all of these questions. The old adage, "you cannot have your cake and eat it", holds true for the study of cellulose synthesis. While Acetobacter xylinum may be an ideal subject for time-lapse cinematography of the in vivo process of cellulose synthesis, this organism is not particularly well suited for the study of the role of membranes in microfibril assembly. While the cotton fiber is a prime example of a highly specialized cell for cellulose synthesis, the very nature of its thick secondary wall is a barrier to effective fixation. While the green alga Oocystis apiculata is a beautiful example demonstrating highly ordered microfibrils, it cannot be used to make extrapolations of cellulose microfibril orientation in higher plants. The same holds true for the haptophycean alga Pleurochrysis scherifelli which produces cellulosic scales of marvelous geometrical complexity, then transports them through the Golgi apparatus pathway to the external surface of the cell.9

Today, I will review the mechanisms by which living cells assemble cellulose. I will use a variety of organisms to demonstrate various phases of this process. You will see how certain technological advances have assisted in opening new areas of investigation. You will have a better understanding of why no one has succeeded in making cellulose in vitro, even though there have been numerous reports.10 Hopefully, you will gain a better understanding of the biosynthetic process of cellulose production and an appreciation for the ramifications for future industrial and commercial utilization.


Oocystis apiculata is a unicellular green alga (Fig. 8). The cell is elliptical and has prominent polar thickenings (Fig. 9 Fig. 10, Fig. 11 Fig .13 . When examined in polarized light with the first order red compensator inserted, the cell walls exhibit a typical + birefringence (Fig. 8). At the ultrastructural level, the cell wall is observed to consist of a crossed lamellar pattern of microfibrils (Fig. 14, Fig. 15). Robinson and Preston11 investigated this organism and found that the microfibrils were cellulose. They also examined this organism with a technique known as freeze-etching. We too have used this technique, and with my former student, Dr. David Montezinos, we prepared cells for freeze-etching without any cryoprotectants or chemical fixations.12, 13, 14 The results were striking (Fig. 16) . The procedure is as follows: living cells are rapidly frozen in Freon 22 or in supercooled nitrogen slush at - 208 degrees C. Then, they are transferred to liquid nitrogen at - l96 degrees C. They are placed on a stage at - 150 degrees C, pumped to a vacuum of 10-6 torr, and fractured with a microtome in vacuo at a temperature of - 100 degrees C. A platinum-carbon replica of the fractured surface is made. Then, the replica and specimen are returned to atmospheric pressure and room temperature, and the replica is cleaned. The replica is then examined with a transmission electron microscope. This procedure has the remarkable advantage that the fracturing process tends to split membranes at the central hydrophobic interface. This allows a view of the membrane from the interior. If the fracture face is viewed as though one were inside the cell, this exposure is termed the "E-fracture face."15 F-fracture faces of Oocystis reveal microfibrillar impressions on the outer leaflet of the bimolecular layer of the plasma membrane (Fig. 16) . That these are true impressions is demonstrated by the visualization of microfibrillar "tears" through the fractured membrane (Fig. 17) . During the fracture process, the shearing tends to pull wall microfibrils back through the frozen outer bimolecular leaflet. We noticed that microfibrillar impressions end abruptly with decorated termini (Fig. 18) . We have named these structures "terminal complexes." While we assume that they have something to do with cellulose microfibril assembly, we have no direct evidence to substantiate this point; however, the circumstantial evidence for their involvement in cellulose synthesis is strong. Terminal complexes are comprised of three rows of subunits, hence they may be more specifically termed linear terminal complexes. There are approximately thirty subunits in each row. If one examines the number of glucan chains which can be packed into the observed microfibrillar dimensions in Oocystis, the conclusion that each subunit is involved with the assembly of a single glucan chain is plausible. Another point of interest relates to the position of the linear terminal complex in relation to the plasma membrane. Note that the subunits are deeply embedded within the membrane. Thus, the probability of subunit contact with the inner leaflet of the plasma membrane is high, and it is not unlikely that a transmembrane complex may be present. This is important in considering the movement of metabolic intermediates from the cell to the site of polymerization. Note that the linear complexes are associated with the ends of microfibrils. T hus, it is feasible to suggest that microfibril growth occurs by a terminal addition of glucose residues. If this happens, the terminal complex must move in the plane of the membrane as polymerization and crystallization proceed. This seems reasonable since cell biologists are now finding compelling evidence to substantiate the fluidity of membranes.16 Thus, as the microfibril grows, its distal end becomes entangled in the cell wall, while its proximal synthesizing end moves over the cell surface. It is possible by this mode of synthesis to create an "endless" belt of microfibrils wrapped around the cell surface. The situation is very much like a ball of string.

Is this true for Oocystis? F-fracture faces of the plasma membrane in the polar region of the cell reveal microfibril impressions in pairs which come to an abrupt end as a single impression after partially circling the pole (Fig. 20) . An examination of the pole with a polarization microscope reveals a characteristic Maltese cross pattern and simultaneous presentation of the slow and fast orders. Examination of wall microfibrils at the pole demonstrates that they are arranged in a circular orientation (Fig. 19) .

Another feature of the linear terminal complex will be described. Any given F-fracture view will demonstrate that during active microfibril synthesis, only one lamellar axis is assembled at a time (Fig.16). Furthermore, linear terminal complexes are associated with microfibrils headed toward both poles. This is not to imply that these complexes are found on both ends of the same microfibril. T hey are always found on opposite ends of microfibril pairs (Fig. 21) . Note that some linear complexes also are found oriented just as all the others, but they are free from any associations with microfibrils. These are associated in pairs (Fig. 21) . They have been termed "paired linear complexes" and are believed to be the incipient cellulose-synthesizing complexes. Thus, when cellulose synthesis is initiated by a paired complex, the direction of growth will be toward the poles opposite each member of the pair (Fig. 22) . This ensures the equal distribution of microfibrils to both hemispheres of the cell. When the terminal linear complex approaches the pole and circles partially, synthesis is believed to come to an abrupt halt and then to resume, but in reverse. This conclusion is supported by the tuning fork" images of microfibril impressions as they approach the pole (Fig. 20) . The reversal of synthesis of a microfibril by a linear terminal complex may be hard to conceptualize, but most certainly the process must involve: (1) cessation of cellulose synthesis, (2) disconnection from the product, (3) reversal of the active site of the subunits, and (4) initiation of new synthesis of microfibrils (Fig.23) . I will diverge for the moment and discuss several experiments which have been performed to alter cellulose synthesis in Oocystis. Perhaps the most interesting one examines the effect of various chelators. When cells are treated for 15 min. with 0.1 M EDTA, frozen. and then freeze-fractured, the typical E-fracture views show no organized linear terminal complexes (Fig. 24) . I nstead, the complexes have been dissociated by the EDTA. Occasionally subunit aggregates are seen; however, no complexes remain attached to microfibrils. If 0.1 M MgSO4 is added to the EDTA-treated cells, it is possible to reaggregate the subunits into linear complexes within 15 mm (Fig. 25) . Note, however, that the complexes are not reattached to the ends of the microfibrils from which they were dissociated. They reaggregate in the paired state. Furthermore, the pairs are not initially oriented parallel to the most recent axis of synthesis. After 90 mm, paired complexes will generate microfibrils, and no differences between treated and control samples can be discerned. If EGTA, a chelator for Ca++, is tested, no disassembly of subunits occurs. Thus, Mg++ appears to play an important role in maintaining the integrity of the macromolecular linear complex. It is also conceivable that divalent cations may help to regulate cellulose synthesis by action of subunit disassembly and reaggregation.

Turning to the question of microfibril orientation, we must examine the P-fracture face of the plasma membrane. This is the inner leaflet as viewed from the outside of the cell. Conspicuous structures known as 'granule bands" are present on this face (Fig. 26) . Granule bands are comprised of subunits which have several hierarchical levels of organization. Particles aggregate with other particles to form a row of seven to eight particles. Then two rows combine to form paired rows. Finally, paired rows aggregate (like railroad ties) to form linear arrays which we believe are largely responsible for microfibril orientation. The situation is analogous to a cog railway. The granule band "tracks" serve to keep the microfibrils oriented, either through a physical depression in the plasma membrane, or by some membrane-microfibril connection. As the linear terminal complex generates the microfibril, granule bands are seen to aggregate into the 'track" immediately behind the complex (Fig. 27) . Since the cellulose microfibril is analogous to a 'fiberglass rod," it tends to remain straight. Thus, the orientation of the granule band track will serve to orient the direction of microfibril synthesis.

How might the granule bands be involved in the change in microfibril orientation from layer to layer9 . We have seen many views of granule band tracks where they appear to interconnect into a net-like (Fig. 28) . This could be a "transition state" during the switch in microfibril orientation. Such a switch would imply self assembly and disassembly of the granule band subunits in such a manner that the new orientation could arise as a result of the recognition of the previous orientation. If this hypothesis is correct, the Initial orientation of the first-formed microfibril layer would be the "template" for all other layers to follow. This seems plausible because the initial orientation could be established by the abundant cortical microtubules immediately after cytokinesis (Fig. 12). Thereafter, microtubules would play no role in microfibril orientation.

Several experimental alterations of microfibril orientation may have some bearing on the question of the role of granule bands in microfibril orientation. If Oocystis is treated with 1% colchicine for 12 hr, microfibril synthesis continues; however, the switch to a 90 degree orientation which occurs normally in the absence of colchicine is inhibited. Instead, microfibrils continue to be made, but only in one orientation The excess synthesis only in one orientation is easily visualized by ultrathin sections of the cell wall (Fig. 29) . The process is perfectly reversible for, when the colchicine is removed, the normal 90 degree orientation shift from layer to layer is re-established. How does colchicine affect the granule bands? P-fracture views show granule band disruption but not complete disruption upon treatment with 0.5% colchicine for 30 mm (Fig. 30) . Our preliminary findings on this most interesting phenomenon suggest that colchicine (or vinblastine) has relatively little effect on previously organized granule band tracks; however, the principle effect seems to be on granule band dis-aggregation and re-aggregation 90 degrees to the track. In other words, these drugs seem to inhibit the switch by preventing subunit reorganization. There is precedent for this possibility since these drugs affect the polymerization and depolymerization of tubulin subunits of microtubules. It is possible that the granule bands may be comprised of structures not unlike tubulin subunits. Much work remains to be done before this hypothesis is verified. Nevertheless, Oocystis is a most useful experimental organism for testing these concepts which probably are fundamental to plant cell biology. The added advantage of organized wall microfibrils and the successful freeze-fracturing of Oocystis make it an ideal candidate for the study of this process (Fig. 31) . On the other hand, the cell fractionation properties of Oocystis make it less desirable for the isolation of components for biochemical investigation. This merely reinforces my earlier statement that no single experimental system is ideal for the total understanding of cellulose synthesis. Thus, we shall now examine another interesting system.


Glaucocystis nostochinearum is an alga, but one with taxonomic uncertainty (Fig. 32 , Fig. 33, Fig. 34, Fig. 35 ). Some believe that it represents a symbiotic association of a blue-green alga (prokaryote) with a colorless eukaryotic cell. I prefer to believe that, like Oocystis, it is a photosynthetic eukaryotic cell. Its affinities to the red algae are pronounced; however, rudimentary flagella occur during the division cycle, and flagella are not found among presently recognized red algae. I will leave the taxonomic dispute and discuss the cell biology of this organism which is fascinating.

Glaucocystis is an elliptical cell like Oocystis, but it lacks polar thickenings (Fig. 36) . The cell wall of Glaucocystis is also crossed lamellate as in Oocystis (Fig. 38). One major difference is apparent. Instead of one layer of microfibrils oriented 90 degrees to the adjacent single layer of microfibrils, there are seven to eight layers of microfibrils oriented in the same direction (Fig. 37) . The plasma membrane-wall interface is unlike that in Oocystis. There are regularly spaced invaginations subtended by cisternae called shields" ( Fig. 39 , Fig. 40 ). My collaboration with Dr. J. H. Martin Willison of Dalhousie University in Halifax, Nova Scotia, has led to some very interesting conclusions about microfibril synthesis and orientation in Glaucocystis. We performed freeze-fracturing of the plasma membrane just as in Oocystis and found linear terminal complexes; however, these complexes were not so deeply embedded in the plasma membrane (Fig. 42) . Most interestingly, we found that the direction of synthesis was unidirectional (Fig. 41). Preceding the terminal complexes are the oriented "shields" (Fig. 41). There are no paired complexes as in Oocystis, and there are no granule bands. Therefore, how are the microfibrils oriented? Examination of the polar wall fragments in Glaucocystis reveal not one rotation center as observed in Oocystis, but three overlapping rotation centers (Fig. 44). Furthermore, the microfibrils do not end abruptly in "tuning fork" arrays as In Oocystis but circle over the pole. Therefore, it becomes clear that there are no reversals of microfibril synthesis. The 90 degree orientation is achieved by an "over-the-pole" sweep (Fig. 45).

What additional independent evidence could be brought to bear on the verification of the polar traverse of microfibril synthesis? We thought of the possibility that protoplast rotation might occur as a result of unidirectional synthesis. Our rationale was that as microfibrils are anchored into the wall, the unidirectional synthesis, combined with unidirectional protoplast rotation, could account for the observed microfibril patterns. With the help of Mr. Terry Colpitts, an undergraduate student, we made a time-lapse film of Glaucocystis. To our great delight, we observed the predicted protoplast rotation. Furthermore, this rotation occurs only immediately after cytokinesis, at the time which we have confirmed to be the active period in cellulose microfibril synthesis (Fig. 43). We made a three-dimensional working model to demonstrate that unidirectional protoplast rotation can produce the crossed lamellate microfibril pattern with three rotation centers at each pole (Fig. 46a) (Fig. 46b) .

Glaucocystis and Oocystis are two examples whereby Nature has chosen different pathways to achieve similar patterns of microfibril deposition. The versatility of microfibril assembly and orientation is not limited to algal cells. I will now consider the process of cellulose synthesis in higher plant cells. Predictably, we would be most interested in the process of cellulose synthesis in this group. Unfortunately, there are many barriers to overcome before we fully understand the process in higher plants; however, I will now discuss two Systems of general interest.


One of my former graduate students, Dr. Susette Mueller, has been studying cellulose synthesis and assembly in the corn root (Fig. 47) . Using freeze-fracturing techniques, we have observed terminal complexes, but with two major differences in comparison to OoCystis: (1) the complex is more or less spherical, not linear, and (2) the complex is not embedded within the outer leaflet of the plasma membrane. F-fracture views of corn root cortical cells show typical microfibrillar impressions as in Oocystis and Glaucocystis ( Fig. 48, Fig. 49). The impressions terminate with a globular complex. The substructure of the complex is not nearly so evident because an outer phospholipid leaflet is interposed between the fracture face and the complex. Six to eight subunits appear to be organized into the globular complex. Details of subunit structure are sometimes better observed in microfibrillar tears which pull the globular complex along with the microfibril back through the plasma membrane (Fig. 48). Note that the microfibril orientations in the corn root exhibit a wavy pattern (Fig. 50). This is true for most microfibrils observed among higher plant systems. Microfibril orientation in root cells is almost invariably transverse to the longitudinal axis of the root (Fig. 50).

Many articles have been published on the rather controversial subject of microfibril orientation in relation to cell growth and elongation. It seems clear to me that understanding microfibril assembly and orientation in higher plants will provide the key toward understanding cell elongation, as well as differentiation and morphogenesis. The action of herbicides on cell elongation is not fully understood, nor is the action of auxin. Plant cell tumorigenesis, dedifferentiation, and redifferentiation are all in need of further investigation, especially in relation to microfibril assembly and deposition.

Let's return to the immediate question of how might microfibril orientation in the corn root cell be controlled? At present we have no direct evidence to settle this question, but we have made some Interesting observations. You will recall my earlier statement about the emerging concept of membrane fluidity. A fluid domain of the plasma membrane implies that structures theoretically could move within the membrane. The examples of terminal complexes moving m the plane of the membrane during microfibril growth tend to support the fluid mosaic model for the membrane.16 Studying the process of cell capping in animal cells, Edidin17 has provided evidence for diffusion of components within the membrane. What about a more robust and directed state of ~diffusion"? Let's consider mass flow of membranes. Only two years ago, Mark Bretscher of the MRC Laboratory of Molecular Biology at Cambridge University, England, proposed that lipid molecules of membranes are capable of a continuous, rapid, oriented flow.18 The paper was largely theoretical; nevertheless, the idea seemed extremely interesting to us when we observed microfibrillar pathways around pit fields in the corn root cells (Fig. 50). Pit fields are comprised of intercellular connections known as plasmodesmata. These connections traverse cell walls and provide direct connections between adjacent cells. The pit fields are stationary structures in relation to the plane of the plasma membrane. If one could visualize membrane flow coming in contact with the pit field, it would flow around it just as water in a rapidly moving stream flows around one's legs as one fishes for trout. In the plant cell, we may have an extraordinary marker for visualizing this process, namely the microfibril itself. What if the membrane flow could orient the microfibrils downstream" during synthesis? The case for this seems plausible in the corn root because the microfibrillar patterns around pit fields conceivably might be generated by a fluid motion past the pit field barrier. There are no granule bands in higher plant cells. Cortical microtubules are present (Fig. 52), and they always parallel the transverse microfibrils. However, cells treated with colchicine still have transverse microfibrils, even when microtubules have long since been depolymerized by the drug. Thus, some other force must orient the microfibrils.

Together with Dr. H. Stanley Bennett of the University of North Carolina School of Medicine, I have just completed a survey of actin in over forty organisms which are representative of the entire plant kingdom, including the protists. We have found actin in nearly every sample tested. What is actin? This is a protein com ponent of mammalian muscle. Actin and myosin in the presence of ATP have contractile properties and are capable of converting chemical energy into mechanical energy. That actin might be universally present in the plant kingdom might seem unreasonable at first. But consider that plant cells, like animal cells, have many organelles in common. Therefore, why not actin? Do plants have muscles"? I would state in the negative as far as a homologous structure in plant cells is concerned. But actin can modulate chemical-mechanical transduction in other ways. Consider cytoplasmic streaming. Recent studies 9 definitely have linked actin to this process, and it is highly likely that the forces necessary for the separation of chromosomes in mitosis and meiosis are generated by actin. Our immediate goals are to determine the cellular location of actin. Our present test has been based on the in vitro reaction of a specific myosin fragment (S1) with actin filaments from cell homogenates (Fig. 51). This procedure tells one that actin is present, but it does not say where.

Preliminary freeze-fracture preparations, which do not necessarily fracture the plasma membrane but reveal the cytoplasmic surface of this membrane, show microtubules as well as fine filamentous strands in close association with the cytoplasmic surface. Could these filaments serve as the force-generating System for mass membrane flow? Only future research will settle this question, but such unorthodox proposals may have some merit in explaining heretofore difficult concepts of phloem transport, auxin and plant hormone transport, not to mention the possibility for control of directed cell surface transport in recognition phenomena.


The cotton fiber is a single cell which originates from the epidermis of the ovule' (Fig. 53, Fig. 54, Fig.55, Fig.56, Fig. 57, Fig.58 ). At maturity, the ovaries contain ovules with fully developed fibers (Fig. 57). The work I am about to describe was in collaboration with a former undergraduate student, Dr. John Westafer, and my colleague, Dr. J. H. Martin Willison. Freeze-fracturing the cotton fiber was extremely difficult at first because the moment the fibers were torn from the ovule they would collapse. Dr. Willison developed a technique whereby the entire ovule with attached fibers could be frozen in Freon 22.20 After freezing, the fibers could be broken from the ovule in liquid nitrogen. The use of butyl benzene as a mounting medium turned out to be very beneficial. E-fracture views revealed the presence of spherical complexes, much like those observed in the corn root system (Fig. 60); however, evidence for the association of these complexes with microfibrillar ends has been lacking, despite numerous freeze-fracture attempts. At present, we can only speculate on the pathway of cellulose synthesis in the cotton fiber. Ultrathin sections of fixed material have suggested certain definable changes in the ultramorphology of the cotton fiber during primary and secondary wall formation.21 During primary wall formation, typical Golgi activity is observed (Fig. 61). During this phase the fiber elongates, and the Golgi apparatus probably is involved in the addition of membranes. A schematic diagram of the cytological events during primary wall formation is shown in Fig. 62. Since the secondary wall comprises the bulk of the cotton cellulose, I will focus on this phase of differentiation (Fig. 63). One interesting observation is the presence of membrane-like fragments at the plasma membrane surface during secondary wall formation (Fig. 64). Also, we have noted a lack of Golgi activity during this phase and have observed microvesicles associated with the endoplasmic reticulum (ER) as well as micro-invaginations in the plasma membrane. Although we have no direct evidence, we have speculated that the microvesicles may contain ER-derived glucan synthetases, and the micro-invaginations may represent fusion sites for the microvesicles to release the synthetases to the cell surface, whereupon they are activated for cellulose synthesis. This is plausible since the freeze-fracture evidence substantiates the presence of a spherical complex on the exterior of the plasma membrane.

One of the enigmas of cellulose synthesis in the cotton fiber has been the failure until most recently to achieve in vitro cellulose synthesis from detached fibers. Every time the fibers were disturbed, (l~3)-fl-glucans (callose) were synthesized instead of (l~4)- -glucans (cellulose). I believe this points to the extreme lability of the cellulose synthesizing system. Any disturbance of the plasma membrane will destroy the capability of generating (l~4)- glucans which can crystallize to form microfibrils. Recently Dr. Nicholas C. Carpita, a postdoctoral student in Dr. Deborah Delmer's laboratory at the ERDA/MSU Plant Research Laboratory in East Lansing, Michigan, has succeeded in making (l~4)- -glucans from detached cotton fibers.22 Dr. Carpita used polyethylene glycol (mol wt 4000) prior to detachment. This presumably protected the plasma membrane from severe osmotic shock and therefore maintained the system sufficiently intact to generate (l~4)- -glucans. No evidence presently is available to prove that microfibrils are generated by this procedure. If microfibrils, in fact, are produced by detached cotton fibers, this will be the first successful in vitro synthesis of cellulose.

The economic implications of test tube cellulose are virtually limitless. Consider, for example, the possibility of producing highly purified cellulose on a mass scale. Only the present economics would dictate the process to be economically unreasonable. But in the long run, where highly purified celluloses of specific molecular weight are desired, it might be more economically feasible to produce them in vitro. It is at once clear that producing purified celluloses from wood pulp is an energy-consuming process. Obtaining cellulose from the cotton plant is also energy- and labor-intensive if one considers all of the parameters in producing cotton such as land management, fertilizer, pesticides, climate, harvesting, and ginning. I am not advocating at this symposium that we immediately convert to test tube or in vitro cellulose production. I do hope to make you aware of a new set of possibilities as we continue to progress into our energy-deprived future.

While cotton cellulose is a widely used product, I wish to consider cellulose synthesis by a prokaryotic organism, Acetobacter xy1inum. I believe that future research in cellulose synthesis (in vivo and in vitro) is brightest with this organism.


Acetobacter xylinurn is a Gram-negative bacterium. It is rod shaped (Fig. 65). Some of the first electron microscope studies of cellulose were made in the 1940's using Acetobacter,23 and most of the biochemical studies which claim in vitro cellulose synthesis have been made with this organism.24 I became interested in Acetobacter cellulose synthesis because the published reports indicating that cellulose synthesis occurs extracellularly at a distance from the bacterium were incompatible with the accumulating evidence in our laboratory with Oocystis, Glaucocystis, and higher plants. Furthermore, this organism seemed to be an excellent candidate for visualizing the in vivo assembly of cellulose since the product occurs, not in the form of a cell wall, but as an extracellular "pellicle". The pellicle is the mass of intertwining cellulosic ribbons seen in Fig. 69.

Acetobacter utilizes glucose or succinate for conversion into cellulose. The organism is an obligate aerobe. Therefore, it grows at the air-liquid interface of a standing liquid culture. A pellicle of cellulose is generated at this interface (Fig. 66). Visible pellicles are formed within 20-30 mm, and this makes Acetobacter an extremely attractive system in which to follow the synthesis of cellulose microscopically. To obtain cells active in cellulose synthesis, we usually transfer a 24 hr pellicle to a new growth medium or deionized water. Vigorous shaking of the pellicle will free the active bacteria into the liquid. The turbid bacterial suspension is rapidly transferred to a watch glass. Samples for observation are placed on a microscope slide and examined with dark-field microscopy. Cells immediately separated from the pellicle show no extending structures; however, cells 5 mm into cellulose synthesis have short ribbons attached to and projecting from the bacteria (Fig. 67). After 20 mm, the ribbons have elongated and intertwined with adjacent ribbons to produce the familiar pellicle (Fig. 68). Preparations incubated for periods longer than 60 mm are not useful for microscopic examination.

The time-course of ribbon extension also has been followed by electron microscopy. Bacterial preparations can be dried down on electron microscope grids and negatively stained with an electron dense material such as uranyl acetate. A 5 mm preparation is shown in Fig. 70. A 20 mm preparation is depicted in Fig. 69 ,and a 30 mm preparation in Fig. 71. Closer examination of the ribbon with negative staining reveals it to consist of microfibrils arranged in two tiers (Fig. 72). Approximately twenty-three microfibrils are in each tier. Thus, the ribbon consists of approximately forty-six microfibrils. The dimensions of each microfibril suggest that twenty-eight glucan chains can be packed into this area. Therefore, 1288 glucan chains can be packed into the ribbon.

A fortuitous circumstance occurred whereby the relationship of the microfibrils to the synthesizing regions could be analyzed. During the process of vigorously shaking the bacteria from the pellicle, part of the ribbon remained attached to the bacterial surface; however, it was torn from the surface with enough force to pull the nascent cellulose microfibrils from their respective synthesizing centers. The result is shown in Fig. 74. Note the lateral projections. There are approximately forty-six such projections, and the ribbon tapers from its full width on one end to only several microfibrils on the other. The indirect conclusions to be drawn are obvious, namely, that the cellulosic ribbon is composed of microfibrils, and the synthesizing centers are arranged in a row on the cell surface, paralleling the longitudinal axis of the bacterium. Are the glucose residues being incorporated into the cellulosic ribbon at the bacterial surface or at the opposite end of the ribbon? Electron microscope autoradiography was the tool for testing this problem. In collaboration with Susette Mueller and Kay Cooper, we were able to conclude from 3H-glucose that the label, indeed, is incorporated into the ribbon at the cell surface ( Fig. 75, Fig. 76 ).

Furthermore, from measurement of the labeled contour length, an approximation of the kinetics of cellulose synthesis could be calculated (Fig. 77). On this basis the maximum incorporation rate was equivalent to an elongation rate of 2.5 microns/ min~1. This is equivalent to 1.5-3.5 x lO8 glucose residues per cell per hr incorporated into cellulose. Independent results for the kinetics of cellulose synthesis have come from a detailed analysis of time-lapse films of the in vivo process (Fig. 78). From these data, the rate of linear elongation of ribbons was calculated to be 2.59 microns/ min-1 + or - 0.49microns/ min-1 at one standard deviation.

While Acetobacter has given us new insight into the kinetics of cellulose synthesis, we know even less about the ultrastructural cytology of the process. For example, ultrathin sections show the vague outlines of the ribbon (Fig. 65). We do know that the ribbon is in close contact with the bacterial cell envelope. Thus, unlike eukaryotes, the ribbon is at some distance from the plasma membrane. Freeze-fracture of the cell envelope reveals images of the ribbon in contact with the envelope, but we cannot see the synthesizing sites ( Fig. 79, Fig. 80 ). I t is quite evident, however, that the ribbon is in intimate contact with the envelope, and that a true ribbon is formed in close contact with the envelope. F-fracture faces through the outer lipopolysaccharide (LPS) layer of the cell envelope reveal a linear row of particles in the expected region of synthesis.25 Several recent biochemical studies have suggested that enzymes for the synthesis of cellulose and particularly of its intermediates are located in the bacterial protoplasm, the plasma membrane, and the periplasmic space.26 We have recently suggested that these enzymes are arranged as a multienzyme complex, which not only spans the plasma membrane and periplasmic space but also the outer LPS layer27 (Fig. 81). We have suggested that the terminal glycosyl transferase is located in the outer LPS layer. It is therefore possible, but not confirmed, that the linear array of particles revealed by freeze-etching (Fig. 79) may be such complexes. We have proposed that the outer membrane may also contain a catalyst involved in the crystallization of the nascent glucan chains. Such a catalyst may serve to regulate the degree of crystallinity or the observed capacity for change to an alternate band-like state (Fig. 82). One major point to be derived from Fig. 81 is the concept that a coordinated enzyme complex is necessary to produce cellulose microfibrils. Disturbance of any part of the complex could result in cessation of the process.

Another point of considerable importance relates to the location of the complexes on the bacterial surface. They are fixed with respect to the cell surface. They are not free to move in the plane of the cell surface, as is the case for Oocystis, Glaucocystis, and corn root. I have postulated the fixed vs. mobile site hypothesis to explain a fundamental dichotomy of cellulose synthesis in relation to the structures generated by the process. This hypothesis states that organisms with fixed site cellulose synthesis produce microfibrils which can form tufts, ribbons, stalks, but not cell walls, whereas organisms with mobile site cellulose synthesis produce microfibrils which can become incorporated into cell walls.

The next obvious question relates to the occurrence of fixed site cellulose synthesis among eukaryotic cells. With a former student, Harold Kyriazi, I have examined cellulose stalk formation in the cellular slime mold Dictyostelium discoideum, and we have concluded that pre stalk cells could have fixed sites on the plasma membrane (Fig. 6). These sites could be held in fixed position by a subtending actin system Furthermore, these sites are polarized inasmuch as they are capable of generating a unidirectional segment of the total stalk. Together, many prestalk cells produce the familiar stalk or rising steliogen. Once inside the stalk, the prestalk cells switch over to mobile site synthesis and encyst, forming typical cell walls (Fig. 6).

Returning to Acetobacter, I would like to conclude with the presentation of some aberrant forms of polymer and by discussing events related to the duplication and growth of the cellulose-synthesizing complex during the cell cycle. Recalling the forty-six lateral projections which presumably arose from a physical shearing during separation from the pellicle, let's examine several other examples of lateral projections. One is the presence of projections at the terminus of a ribbon (Fig. 73). This is interpreted diagramatically in Fig. 83. During pellicle separation, part of the nascent microfibrils are torn from their sites of synthesis at the cell envelope. As new ribbon formation proceeds, the remnants of the nascent cellulose torn from the surface remain attached.

The relation between the ribbon and the lateral projections is also seen in the production of a wavy 'bandlike" material (Fig. 84). This material usually is formed in older cultures, and we have not yet succeeded in getting the cultures to produce only this material. Furthermore this material is not produced under defined conditions, therefore it is difficult to obtain in quantity. However, we have obtained sufficient band material to provide a preliminary analysis. It is composed of glucose, is (l~4)- linked, and X-ray analysis reveals cellotetrose and/or cellopentose as the major polymers.28 The band material can be hydrolyzed to glucose by cellulase. Mild acid hydrolysis yields cellobiose. From these preliminary data, it is suggested that the band material is very similar to cellulose. Its close association with individual ribbons of microfibril cellulose strongly suggest that the same sites of cellulose synthesis are involved in the synthesis of the band material. Two possibilities come to mind for explaining the production of the band material: (1) cellulases might hydrolyze the cellulose as it is being made, or (2) the synthesizing site for each microfibril might start to rotate during synthesis, producing so much strain on the glucan chains that they are unable to crystallize into a microfibril.

In the first case, we have no present evidence for an endogenous cellulase system in Acetobacter. One possibility is that such a system could be activated when all of the glucose in the medium is consumed. The bacterium could "feed" upon its own cellulose. Since the function of the pellicle is unknown, this is one possibility. In the absence of definitive data, this hypothesis remains untested.

The other concept, namely, the rotation of individual synthesizing centers is depicted in Fig. 85. Some evidence tends to substantiate the possibility of rotation. Note in Fig. 84 that during the conversion from ribbon to band material there are cables of microfibrils rather than conventional tiers of twenty-three microfibrils. This suggests that during normal synthesis of the ribbon, the extrusion regions of the synthesizing centers are directed toward one pole of the bacterium (Fig. 85, Fig. 86, Fig.87 ). The result is the synthesis of a ribbon only from one pole. The time-lapse films substantiate unidirectional ribbon synthesis. However, at cell division just before the cleavage furrow is completed, ribbons often are extruded from both ends of the same bacterium. It is concluded that half of the row of synthesizing centers has reversed its polarity (Fig. 87). Since the cleavage furrow will bisect the row of synthesizing sites, the conservation of the synthesizing apparatus in both daughter cells seems reasonable. What is extremely important to note is that one of the daughter cells will never stop producing the ribbon. This implies that it may be possible to produce cellulose with molecular weights into the millions so long as synthesis continues unabated. However, during division, one of the daughter cells must start synthesis anew. Another interesting point is that ribbons of daughter cells are smaller than those of mature cells. This implies that as the daughter cells grow they acquire new synthesizing sites (Fig. 88, Fig. 89 ). Another case in point as demonstrated by the films and with negative staining (Fig. 90, Fig. 91) is that the linear row of synthesizing centers becomes bisected by the cleavage furrow, but the direction of synthesis is not reversed.

Returning to the concept of continual rotation of the synthesized sites, it is obvious that we are considering a hypothetical case; however, if the rotations are slow enough relative to the rate of glucan chain synthesis, microfibrils will still be produced. They will be in the form of twisted cables and will not be unidirectionally oriented parallel to the longitudinal axis of the bacterium. On the contrary, they would tend to diverge perpendicular to the longitudinal axis. If the rotation rates increase relative to synthesis, or if the polymerization rates slow down relative to a constant rotation rate, the result is an unacceptable strain on the glucan chains. The strain could cause glucan collapse. Are the cellotetrose and cellopentose oligosaccharides a result of this rotational torque? I cannot answer that question at present, but it is an intriguing idea which may be testable as we learn more about cellulose synthesis in Acetobacter.

One final point concerns the microfibrillar ribbon. Note that it twists at a periodic locus (Fig. 73, Fig. 91 ). At first, one might believe that this is a drying artifact of the negative stain; however, the time lapse films demonstrate unequivocally that the bacteria rotate as the ribbons are made. This demonstrates that the twisting of the microfibrillar ribbon is not an artifact of drying but could represent the structural rearrangement of glucan chains as they are undergoing crystallization into microfibrils. A slight dislocation might inevitably lead to the twisting. The time-lapse films also show that under conditions of synthesis in very thin films of liquid, the bacteria do not present what macromolecular forces are involved in these transitions. We do know that we have tapped the basic principles for understanding how cellulose microfibrils are made. I hope that you have acquired an appreciation for the tremendous diversity and complexity in the biosynthesis of Nature's most abundant macro-molecule.


I would like to express my appreciation to Richard Santos for excellent technical assistance, to Ana Birkner for providing drawings, to Roy Coomans for Fig. 35, to Alan White for Figs. 8 and 10. I would like to thank Dr. David Montezinos especially for his beautiful electron micrographs of Oocystis, Dr. Martin Willison for his collaboration and electron micrographs on Glaucocystis, cotton, and Acetobacter, to Dr. Susette Mueller for her collaboration and electron micrographs of corn seedlings, to Dr. H. Stanley Bennett for his friendship and collaboration on the actin project, to Harold Kyriazi for his collaboration and electron micrographs of Dictyostehum, to Mrs. Kay Cooper for her technical assistance in a variety of projects, to Mr. Terry Colpitts for collaboration with time-lapse cinematography projects and model of Glaucocystis, and to Dr. John Westafer for collaboration on cotton fiber ultrastructure. The research described in this symposium was supported, in part, by a grant from the National Science Foundation (PCM 77-05628).


1. K. Hess, Die Chemie der Zellulose and Jhrer Begleiter, Leipzig: Akademischen Verlagsgesallschaft, M.B.H., 1928.

2. F. Hoyle and N. C. Wickramasinghe, Nature, 268, 610 (1977).

3. A. B. Wardrop, Protoplasma, 70, 73 (1970).

4. D. A. Hall, F. Happey, P. F. Lloyd, and H. Saxl, Proc. R. Soc. London, Ser. B,

151, 497 (1%0).

5. L. S. Olive, The Mycetozoans, New York, NY: Academic Press, 1975.

6. R. D. Preston, The Physical Biology of Plant Cell Walls, London: Chapman and Hall, 1974.

7. K. Keegstra, K. w. Talmadge, W. D. Bauer, and P. Albersheim, Plant Physiol.,

51, 188 (1973).

8. R. Hooke, Micrographia, or, Some Physiological Descriptions of Minute Bodies Made by Magnifting Glasses; with Observations and Inquiries Thereupon, Reprint, New York, NY: Dover Publications, inc., 1961.

9. (a) R. M. Brown, Jr. and D. K. Romanovicz, Appl. Polym. Symp., 28, 537 (1976). (b) D. K. Romanovicz and R. M. Brown, Jr., Appl. Polym. Symp., 28, 587 (1976).

10. D. P. Delmer in Recent Advances in Phytochemistry, Vol.11; The Structure, Biosynthesis and Degradation of Wood (F. A. Loewus and V. C. Runeckles, editors), New York, NY: Plenum Publishing Corp., 1977.

11. D. G. Robinson and R. D. Preston, Planta, 104, 234 (1972).

12. R. M. Brown, Jr. and D. Montezinos, Proc. Natl. Acad. Sci. U.S.A., 73, 143 (1976).

13. D. Montezinos and R. M. Brown, Jr., J. Supramol. Struct., 5, 277 (1976).

14. D. Montezinos and R. M. Brown, Jr., Cytobios, in press.

15. D. Branton, et al., Science, 190, 54 (1975).

16. S. J. Singer and G. L. Nicolson, Science, 175, 720 (1972).

17. M. Edidin and T. Wei, J. Cell Biol., 75, 475 (1977).

18. M. S. Bretscher, Nature, 260, 21(1976).

19. Y. M. Kersey, P. K. Hepler, B. A. Palevitz, and N. K. Wessells, Proc. NatI. Acad. Sci. U.S.A., 73, 165 (1976).

20. J. H. M. Willison and R. M. Brown, Jr., Protoplasma, 92, 21(1977).

21. J. M. Westafer and R. M. Brown, Jr., Cytobios, 15, 111(1976).

22. N. C. Carpita, Demonstration of Cellulose Synthesis in Detached Cotton Fibers," presented at 176th ACS National Meeting (Cellulose, Paper and Textile Chemistry Division), Miami, Florida, 1978.

23. K. Muhlethaler, Biochim. Biophys. Acta, 3, 527 (1949).

24. A. Forge, Ann. Bot., 41, 447 (1977).

25. R. M. Brown, Jr., J. H. M. Willison, and C. L. Richardson, Proc. NatI. Acad. Sci. U.S.A., 73, 4565 (1976).

26. D. Cooper and R. St. J. Manley, Biochim. Biophys. Acta, 381, 109 (1975).

27. R. M. Brown, Jr. and J. H. M. Willison in International Cell Biology, 1976-1977 (B. R. Brinkley and K. R. Porter, editors), New York, NY: The Rockefeller University Press, 1977.

28. R. J. Chandross, personal communication.

Thanks to Phlip Morris for permission to reproduce in its entirety, this article on the Internet. For questions, please contact:R. Malcolm Brown, Jr.