J. theor. Biol. (1994) 171, 137-150 The Problem of Hyphal Growth in Streptomycetes and Fungi Artur L. Koca Department of Biology, Indiana University, Bloomington, IN 41405, U.S.A: (Received on 26 July 1992, Accepted in revised form on 22 June 1994) Apical growth is the growth habit of both filamentous streptomycetes and mycelia! fungi. It is also the tactic of higher plants, but only hyphal growth is considered here. The problem of apical hyphal growth is that the structure is only supported from its base giving litte support to the continuously elongating ip. Clearly, a different strategy is needed for the two classes of micro-organisms because the prokaryote hhas no eytoskeleton to perforin mechanical work, but instead fas a passive exoskeleton resisting the cell's turgor pressure; ic. it has a covalently’ cross-linked fabric covering the entire cell—the sacculus—that resists the cell's turgor pressure. Conversely, the lower eukaryote does have a eytoskeleton, but has no totally enclosing cross-linked fabric; rather, the hyphal wal of the cukaryatic fungi resembles fiberglass in that the wall is composed of fibers embedded in a plastic phase that ‘gradually sets to become more rid. Very different models are considered here for apical growth of these {Mo classes of micro-organisms. [is proposed that the prokaryotic streptomycetes carry out a rapid turnover of the tip wall, reminiscent of the well-established inside-to-oulside growth of the side wall of bacilli. This process maintains the integrity of the wall during growth because an intact covalently linked portion of the tip is always present. This model depends on the ability of the murein composing the sacculus to expand elastically. Two models are considered for the mycelial fungi, where the tip wall is enlarged by the result of fusion of vesicles. The recent model of Bartnicki-Garcia er al. (1989, Protoplasma 183, 46-51) is critically discussed. Their model is that the rate of addition to a particular clement of wall area depends on its distance from an autonomously moving Spitzenkérper (idealized as the vesicle supply ceater, VSC, in the theoretical constructs). The three-dimensional version of their” ‘model is also analyzed. The favored model is the “soft spot hypothesis"; itis hased on ideas formulated ‘more than a century ago and recently revived independently by F. M. Harold, J. G. M. Wessels, and AL. Koch for diffrent reasons. Itassumes thatthe vesicles are fused (intussuscepted) only if they teach points in the wall that are sufficiently new to be still plastic. Thus newer wall will be preferentially incorporated into fresh wall. This could provide a way for turgor pressure to facilitate the incorporation of wall materials and synthetic enzymes and couple elongation to the success of the organism in converting resources into cytoplasm. It could also provide a way for vacuole formation to force clongation. For the eficient and safe elongation of fungal hyphae there may be a role for @ combination of a “vesicle supply center” and a “soft spot” model to work together. 1, Introduction ‘An entirely different set of problems must be solved by rod-shaped cells that grow at only one end than by those micro-organisms that grow by elongation over their central cylindrical region. In the latter case the problem of maintaining shape is more easily solved because the pair of existing poles form a supporting framework for central enlargement (Koch et al., 1982; Koch, 19884, 6; 19904, b,c, 19825). Both mycelial 022-5193/94/220137 + 14 $08.00/0, prokaryotes and eukaryotes have developed different techniques for apical growth and use them as ways to burrow through soil or rotting wood. By the combi- nation of apical growth and branching these organ- isms can exploit resources effectively throughout three-dimensional solid substratum, It is relevant here to draw an analogy with the architectural construction of a dome. The apex of these growing organisms is a development beyond the more straightforward elongation of the middle of the 17 © 1994 Academic Press Limited138 ALL. KOCH cylinder, just as the Roman and Gothic domes are beyond the earlier lintel and flat roof construc- tion; indeed, arches and domes require special con- struction techniques beyond those needed for the more elementary forms. ‘The mechanisms for hyphal growth that prokary- otes use must be different from those of eukaryotes. Prokaryotes do not have cytoskeletal elements to serve as internal scaffolding, to provide stress-bearing braces, and to do the mechanical work of extension and contraction (see Koch, 1991). The bacterial wall owes its unique strength to the cross-linked peptido- alycan (or murein) of which it is composed. Instead, the surface stress theory argues that cytoplasmic growth which tends to inctease the turgor pressure concomitantly causes an increase in the tension in the sacculus network that encloses the cell, which in turn ‘causes an increase in activity of the autolysins present in the wall. The loosening of the peptidoglycan results, in the pulling into the stress-bearing part of the new wall fabric that is cross-linked to the old, but unstressed, wall, On the other hand, eukaryotes do not have the two- or threc-dimensional covalently cross-linked fabric of peptidoglycan (or murein). Still, the tip must elongate and exert considerable force while growing through media containing solid obstacles. It is probable that the cytoskeleton serves a role in this process. The wall of the fungus (Wessels, 1986) contains a one-dimensional linear polymer. This substance serves the equivalent role of the glass fibers in fiber- lass. In different fungi this function is served by chitin and cellulose fibrils. Various polymers serve as the substances equivalent to the epoxy component of fiberglass; ic. the mannans, glucans, etc. As with fiberglass, the fibers must be distributed in space and surrounded by a mastic phase, which is allowed to cross-link to itself and to the fibers as it hardens. Only then can the new wall serve a stress-resistant structural role, After curing, the fibers bequeath stress-resistance and the combination of fibers and mastic give resistance to compression and flexing. During the hyphal elongation of eukary- otes, the tip is the least cured part of the wall, but it must exert force on the environmental substra- tum, The purpose of this: paper is to explore these problems and speculate on the disparate sol- tutions employed for hyphal growth in these two kingdoms. The basic premise is that engineering tactics devised by humans to solve their problems are similar to those evolved by lower creatures (0 solve theirs. 2, The Elastic and Plastic Components of Growing Walls ‘When not growing, the cells architecture would be best served by a strong, elastic material that would not deform or creep under continuing pressure and stresses, On the other hand, in order to grow, the wall must be plastic enough to enlarge, driven solely by the internal pressure increase caused by the increase in cell volume due to the formation of new cytoplasm for the prokaryote, which may be augmented by mechano-enzymes in the cukaryote. Formulae relevant to the present problem are the expressions for the meridional, stress and the hoop stress of a pressurized vessel made of an clastic solid material which has the shape of either an ellipsoidal shell of revolution or of a cylindrical shell. The meridional normal stress resultant, N,. tends to elongate the cell, while the hoop normal stress resul- tant, Np, tends to enlarge the radius. These are engineering terms and refer to the stress per unit thickness in the tangent to the surface in either the longitudinal or circumferential direction of a pressure vessel of constant thickness of a uniform material. Equations for the ellipsoidal and cylindrical solid shells in terms of the pressure and shape are given in Fhigge (1973) and also in Koch (19886, 1992) and will be used to evaluate the stresses in the non-growing “wall. ‘Quite different mathematics applies for plastic ma terials. For the general case shown by LaPlace in 1805, T = P/(1/R, + 1/R,) where T, the surface ten- sion of the plastic material, is the work required to ‘generate a new unit of external surface. R, and R, are the principal radii of curvature, and P is the turgor pressure resulting from osmotic forces; T is the ana- logue of the normal stress resultants for the solid case and has the same units: work per unit of surface area.The important difference is that the same T applies no matter which direction in the surface one is considering, because the bonds in the plastic case can readily rearrange while they cannot in the elastic case, where N, and N, are, in general, different. The interplay during cell growth between plastic and elastic behavior of parts of the wall is important, particularly for bacterial rod extension (Koch, 1993), The covalently linked sacculus of bacteria or the wall of the fungus are clearly elastic solids when growth is not occurring: on the other hand, as they grow and enlarge, physical forces can operate in the same way as do the physical forces acting on a soap bubble. Different rod-shaped bacteria control growth in different ways, and the dual purpose of this ‘paper is to consider how it is done for apical growth of hyphalHYPHAL GROWTH IN STREPTOMYCETES AND FUNGI 139 streptomycetes and to consider the equivalent strat- egy of hyphal fungi. 3. The Relatively Simple Methods for Side Wall Elongation of Bacteria Growing by Binary Fission Bacteria show a variety of strategies of growth within the paradigm of creating and enlarging the cross-linked exoskeleton structure of peptidoglycan Probably the simplest mechanism for division is that of the Gram-positive cocci and rods. Under stress-free conditions these organisms construct a septum and then employ their autolysins to bisect it. Only after a region of the septum becomes split and exteriorized is it subject to the full stress due to the turgor pressure. ‘The details vary in comparing Bacillus subtilis (Koch & Burdett, 19860, 6; Koch, 1992) with Enterococcus hirae (Streptococcus faecium) (Koch, 1992). The key point is that some peptidoglycans, such as that of Escherichia coli are capable of a great deal of stretch- ing (Koch & Woeste, 1992) and that of Enterococcus hirae is not, According to theory, the completed poles of rod-shaped organisms should not turn over—they do appear to be inert for Escherichia coli (Koch & Woldringh, 1994)—however, the pole wall of Bacillus subtilis does actually turn over very slowly (Archibald & Coapes, 1976; Mobley et al, 1984; Clarke-Starman et al., 1989), 3.1, INSIDE-TO-OUTSIDE GROWTH OF THE CYLINDRICAL REGION OF THE GRAM-POSITIVE ROD ‘The mechanism for the safe elongation process of Gram-positive rod-shaped organisms is simply that new layers are continuously and diffusely laid down on the inside of the cell wall and the stress-stretched layers are autolyzed on the outside; this mechanism for safe elongation has been called “the inside-to-out- side” model (Koch er al, 1981) and is in accord with a large number of experimental findings. Various aspects of the process have been developed sub- sequently (Koch, 1983, 19884, 5, 1990a, c, 1993). This strategy ensures that there will always be intact layers of peptidoglycan to contain the cell's turgor pressure. Asa layer moves outwardly itis stretched, mostly in the longitudinal direction, Eventually it is stretched to its effective elastic limit; in so doing the ability to stretch longitudinally is fully used so that a given layer is not able to stretch circumferentially when it, in turn, becomes the main stress-beating layer. It is this feature (Koch, 1993) that makes the composite of unstressed inner and more highly stressed outer parts of the wall together behave in the same way that a plastic membrane would act given the same physical constraints. 3.2, THE VARIABLE STRESS MODE. Mach more complex models are needed for the growth of other types of bacteria than Gram-positive cocci. Itis the Variable-T model which serves here as the general model for both apical growth of prokary- otes and of eukaryotes. Basically, this model assumes that the organism has metabolic control of the work needed to generate a new unit of surface tension, ic. it can adjust the parameter T by modifying the biochemical situation so that T changes in the region in which pole formation is to occur. The variable-T model is general and non-mechanism orientated and can include the model for the “apical inside-to-out- side” model developed in Section 4.1 below for the streptomycetes. This is a model in which the tip is rapidly turning over so that nascent wall is able to ‘expand easily near the tip. It includes the century-old idea of intussusception in fungi (Reinhardt, 1892) in the new version called the “Soft spot” model developed in Section 5.3. By considering the tip wall to be plastic, it has been shown (Koch, 19822) that the value of the biological analogue of T could be obtained from 5, where S is the slope of the pole profile at any point of interest The exact form of the equation is S=[Pr/Q7y — 1 a This will be used to for one case examined below to calculate 1/7. This can be thought of as a measure of the turnover of the tip for the prokaryote or the kinetics of deposition and hardening of the new wall for the eukaryote. 4, The Sophisticated Growth Strategy of Streptomyces coelicolor The unsupported growth of apices is inherently difficult, particularly for the —cytoskeleton-less prokaryote. There is, however, at least one candidate solution of the problem of hyphal elongation of streptomycetes without invoking a cytoskeleton or mechano-proteins and enzymes, So far thisis the only solution that I have been able to devise. 1 hope this paper will stimulate experimental studies of its key, but unverified, assumption of rapid apical inside-to- ‘outside growth; such studies should include pulse incorporation of precursors and early measurements of release from macromolecular murein to measure any turnover. 4.1, THE APICAL INSIDE-TO-OUTSIDE MODEL The proposed mechanism for tip growth is essen- tially the inverse of that used by the Gram-positive140 ALL. KOCH rod-shaped fission bacteria for side wall growth. The contrast between the two is shown in Fig. |. tis known that the side wall of B. subtilis elongates by the inside-to-outside mechanism that permits safe elongation; it requires new wall incorporation on the inside and gradual displacement to the outside fol- lowed by cleavage and, usually, release to the medium, while the side wall of S. coelicolor appears to turn over very slowly, Certainly, there is very little incorporation of label into the older cylindrical wall (Gray et al., 1990) and the wall retains a constant thickness, Even the small amount of label that does enter the side wall could be due to metabolism of the tracer compound, N-acetyl-D-glicosamine, into other labeled com- pounds, which might then enter the wall as other polysaccharides or into wall proteins. One laboratory hhas excluded true turnover in the side walls and thus the radioactivity would correspond to a slight wall thickening (Miguélez et al., unpublished data). Both the laboratories of Prosser (Gray et al, 1990) and Hardison (Braiia et al, 1982; Miguélez et al, 1988) have shown that the tip rapidly becomes labeled as required by the model to be developed in this section. However, neither laboratory has made measurements, timescale or otherwise, that would allow the metabolic stability of the tip to be deter- mined. Such measurements require new experiments, such as searching for the almost immediate release of labeled turnover products. In accordance with these facts, “the apical inside- to-outside” model for S. coelicolor proposed here, assumes a rapid synthesis of wall in the tip region and an obligatory turnover at the tip of the pole, and that there is very little turnover or degradation at any distance from the tip, except at sites where a branch is to form. The new aspect of the model is to provide @ mechanism to avoid both uncontrolled bulging or rupture at the growing tip. The idea of inside-to-dut- side turnover of the pole is novel; I have found no precedent in the experimental or theoretical literature, Still, even in the absence of evidence for or against turnover, it is necessary to postulate that tanover ‘occurs as an essential part of the mechanism for safe apical growth of streptomycetés. ‘The assumption that the tip wall engages in exten- sive turnover makes it possible for the tip to elongate with the continuing maintenance of intact stress-bear- wall; this is shown in Figs 1 and 2. A second assumption is that the wall, once formed, is capable of being elastically stretched a good deal; while there is no data for streptomycetes wall, there is evidence for this in other peptidoglycans (discussed below). During apical growth, the model proposes that layers are added continuously in the tip region on the inside of existing wall through the cytoplasmic membrane. At present, no attempt is made to explain how this addition is directed mainly at the tip regions. The rate of addition decreases very rapidly with distance from the tip, as evidenced most clearly from the electron microscopic autoradiographic data of Gray et al. (1990) from studies of pulse-labeled cells. Three pro- ‘cesses must be concurrent for the apical inside-to-out- side model: the addition of new layers of unstretched peptidoglycan only in the tip region, the stretching Bacillus Streptomyces. Inside to outside —-—---_____-| Very stable Turning over wall Frio. I. Difference in the tactics of non-apical and apical growth of two prokaryotes. Bacillus subtilis has been shown to elongate safely by te inside-o-outside mode; i, with the continuous addition of peptidoglycan diffusely tothe entire inner face ofthe sige wall and diffuse Aissolution on the outside. As the wall moves outward, iti stretched, main in the longitudinal direction, before it comes to bear stess in the hoop direction. Elongation without bulging occurs because of expenditure of the capability for elastic expansion dve to axial stress before the portion of wall Becomes subject (othe hoop sires. This citcumetance permite fod growth even though the hoop stress i twice the axial stess. Eventually old wal is cleaved and no longer supports the turgor pressure. The pole is formed by creating a septum that is en sp and bulges es dete spe a gor resue comes ob ape fo the enite pol, The ol wal not net ae ‘tumns over extremely slow) (On the other hand. the “apical inside-o-outside™ model proposed here assumes thet Strepromyces coelicolor does just the inverse: the side wall is very stable, and iis the tp that grows is an insideto-outside mode and turns over when it reuches the outside. This affords ‘continuous integrity 10 the tip. For elongation of the Streptomyces coelicolor hyphae to occur itis necessary that the wall, as lid down, is capable of a high degree of expansion when stressed owing to the autolysis of more peripheral outer layers. Ths frees an“underying layer which stretches and bulges forward and outward to allow th tip to grow and press through the environment. Note that neither of these strategies uilizes cytoskeletal elements, which appear 10 be lacking in prokaryotes.HYPHAL GROWTH IN STREPTOMYCETES AND FUNGI 14 ol stretched, autalyzed wall Turgor pressure Fig. 2. The “apical inside-to-outside” model for growth of strepiomycetes. The figure depicts the addition of new wall as the ‘most darkly colored line. The darkness of the line corresponds 10 the density which in turn reflects the degree of having been siretched or autolyzed. Subsequent t0 its fo a layer be ‘comes stressed. The stress supported by a unit thickness of wall, both in the meridionally and hoop direction, is least near the ‘margins ofthe pole because there are many stevs-bearing layers in this nealy cylindrical region, even though the total stress (particu larly the hoop stress) s greater On the other hand, where the wall, is effectively thinnest bacause of autolysis atthe tip, the stress is greatest Autolysis at the tip apples stress to the wall that had previously been non-stress-beating and causing i 10 stretch out- ‘ward and forwards. The streses of the stretched wall atthe tip become very high because the thinness ofthe wall bearing stress, High stress favors autolysin action, which in ten leads to further deformation of the uncleaved portion of the layer, favoring its stretching forward and outward. The sde wall is built up of these pre-stressed portions of wall that cun no longer expand inthe hoop. direction, as well as some wall that is subject to lesser siress because it was originally laid down under the existing cylindrical wall. Soe text for further details, forward and bulging outward of layers laid down earlier, and the autolysis (rupture) of still older layers that are now outmost. This third process transfers stress to underlying layers that then expand @ good deal in response. An integral part of the model is that the murein, as laid down, is unstressed and unstretched and is not affected by the cellular turgor pressure, because it is inside a stress-bearing shell. When the most apical part of the tip peptidoglycan is autolyzed sufficiently to lose its structural integrity, the underlying pre- viously unstressed peptidoglycan becomes stressed resulting in bulging and radial expansion of the side wall. However, it retains its covalent integrity as a fabric covering the tip. As the newly laid-down tip wall stretches and bulges the hypha grows forward because the elongation due to the bulging is much greater than the loss due to the autolysis of the most apical part of the tip. To visualize the apical inside-to-outside model con- sider a stack of cups, but made of elastic rubber. The operation simulating growth is the repetitive addition of a new cup at the inside of the stack, (left-hand side of Fig. 2) the removal of the bottom of the oldest cup because it has been stretched and bulged too far, and the extensive stretching of the intermediate cups as they come to support the turgor pressure. The cup analogy shows how the apical inside-to-outside model could yield stable growth because there is always intact, unstressed, and unstreiched wall and always intact, stressed, and stretched wall in the region closer to the tip surface. The cylindrical side wall does not progressively bulge or inappropriately enlarge its diameter for several reasons. Primarily, this is because at a small distance behind the tip, the wall would be made up of several conical layers, each sleeved inside another and cach already stretched from the bulging process and bearing stress. Depending on how far back from the tip new wall continues to be gradually added, the inner-most layer may have been laid down and not hhave been stretched very much. Going outward through the wall, the layers are stretched essentially to their effective elastic limit. Because many layers are sharing the stress, the stress is fairly uniform. The maximum stress on any layer would be smaller than that for the case of the inside-to-outside growth of B. subtilis where the stress is concentrated in central layers in the wall that have been stretched but not yet autolyzed. Consequently, the maximum stress per layer in the streptomycetes wall would be relatively smaller because all layers bear some stress relative with the stress exerted on the outermost intact layer in the side wall of B. subuilis. While the sharing of stress is the main factor protecting the cylindrical part of the hypha from autolysis in this model, there may be additional factors affecting autolysis of the side wall of strepto- myces (other than the obvious requirement for auto! ysis action where branches are to form). These include wall thickness or chemical structures modifying sus- ceptibility to autolysins. Yet another possible Factor that should be considered and investigated is that the autolysins may be only secreted at the tip and after functioning they may “self-destruct” oF otherwise ‘cease functioning. The latter might happen either because they become disassociated and are lost into the environment, or because they remain attached to liberated fragments of wall, or because they are unstable and denature quickly. The proposed model would only work if the pepti- doglycan were capable of a good deal of expansion, From the experimental data and the models for wall growth of various bacteria, the degree to which different nascent wall can expand when stressed canM42 ALL. KOCH, be estimated. The expansion factor of surface area measured for Escherichia coli sacculi is 4 (Koch & Woeste, 1992); the factor for the B. subtilis pole in vivo is 1-5 (Koch & Burdett, 1986a, 5); on the other hand, E. hirae wall probably expands much less once exter- nalized (although it probably can be distorted slightly) implying that the expansion factor is very close to unity (Koch, 1992). For the murein of E. coli, we showed that the area of sacculus depends critically on its net charge (Koch & Woeste, 1992). Therefore a detailed knowledge of the chemical structure of streptomycetes wall could support or falsify the model 4.2. MORPHOMETRIC ANALYSIS OF THE APICES OF STREPTOMYCES COELICOLOR Measurements of the dimensions of poles of S. coelicolor have been made by Gray et al. (1990). They fitted the electron microscopic data of the tip profile to both an ellipse and to a power series. With more fitted parameters, the power series gave a better fit than the ellipse, but will not be considered here; a power series is inappropriate because it does. not extrapolate to a constant radius at large distances behind the tip. The ellipsoidal model pieced to a cylindrical shell is more realistic, However, in the elliptical fitting there is the subjective difficulty of choosing the point dividing the pole from the side wall, so actually there is an additional parameter value that must be assigned in the fitting process. In Fig. 3, the data from figure 4 of Gray et al. (1990) have been replotted; this data is from measurements ofa single apex examined in the electron microscope. Thave achieved a better fit (solid line) to an ellipsoid of revolution than their fit (not shown) by assuming that the pole actually siarted at 0:2 um farther from the tip of the pole than the value of 0978 zm that they had assumed. The new elliptical fit is only important because then the engineering formula for the hoop and meridional stress cited above can be used to calculate the stresses. These are shown in Fig. 4 on the additional (and incorrect) assumption that the tip wall is made of material of uniform thickness with long dashes for hoop and shorter dashes for axial, or meridional, stress. However, these calculations are relevant because they show that if the wall is of uniform thickness and composition, no matter what the details of the fitting procedure, the stress resultants are smaller at the tip. In order to have the stress greater to invoke autolysin action directed by stress in the substrate murein, it must be postulated that the wall at the tip that is actually bearing the stress is much thinner owing to the autolysis of the outer-most wall, These calculations also show that os| distance gam) Asi 02 | er 0 Oi 0205 a" 95, ‘Radial distance (um) Fic. 3. Eipsoidal and hyphoidal fits to the measured shape of | pole of S. coelicolor. The éata are taken [rom Gray eal. (1990) ‘and derived from measurements from an electron micrograph of a Single median section, They fited their data to an elipsotdal pole shape, theft presented here (sold line) is closer tothe experimental ‘data and was obtained by assuming tha the elliptical portion of the tip extends farther than they had assumed. The data shows the beter fit to the hyphoid model (dashed line). The hyphoid for the prokaryote case only as the vite that it sa one-parameter single ‘equation curve that predicts a non-pointed tip anda cylinder region al large distance from the tp, the hoop stress is larger (two-fold greater) near the edge of the pole than is the meridional stress, and that both are large compared with the stresses near the tip. This is the factor that causes the radial expansion of the wall. There is a second way in which the morphological data can be used to consider growth mechanisms; this is using the framework of the Variable-T model described above to calculate the parameter 1/7, Equation (1) cited above was used for the 1/T plot in Fig. 4 and demonstrates that 1/T’ decreases dramati- cally with distance, which in this case can be inter- preted to reflect somehow to the intensity of turnover. Figure 3 shows a second fit to the hyphoid equation discussed below, where an entirely different mechan- ism is under consideration. It would seem to beHYPHAL GROWTH IN STREPTOMYCETES AND FUNGI 143, ool 02 09 04 05 Radial distance (um) Fic. 4. Axial and hoop stresses and LT for S. coelicolor. The elliptical ft to the data of Gray et a, (1980) shown in Fig. 3 is reproduced here by the solid line, From the cited formulae the ‘tresses were computed from this ft by assuming the wall to bea ‘on-growing elastic solid of uniform thickness. The hoop (circurn- ferential stress resultaats, symbolized by Ne isshown by (———), and the axial (eneridional) stress resultans, symbolized by Ny, is shown by (~--). These Would reflect the streses in the actuat ‘Bpidifed portions ofthe wall, fhe stess-bearing part had constant thickness. Two conclusions fllow: the ip must have a thinner wall ‘supporting the stress for the apical inside-to-outside mode! to Tunction; and the stresses near the edges of the poles will force orward and outward movement. The plot of L/T (---) is appropei- ate tothe degree thatthe wall ean be treated 28 plastic thi is the Paramoter of the surface stress theory that is 2 measure of the wal's ability to grow. appropriate simply because it meets the general re- quirement for the basic mathematical form; viz., an tunpointed tip and a constant radius at large distances from the tip. A third reason for its choice is that only ‘one parameter is involved in the in fitting; it fits quite well, 5. The Eukaryote Strategy Superficially, the hyphae of actinomyces and those of fungi are similar. Obvious differences are their citizenship in different kingdoms and the fact that the radius of eukaryotic hypha is generally an order of magnitude larger than that of prokaryotic hypha. Since the force acting to thrust the hypha forward is in proportion to the cross-sectional area multiplied by the turgor pressure, the growing eukaryotic hypha can do more work. However, from the formulae given above it follows from the difference in size that work to enlarge the tip of the wall area would be larger in the larger eukaryote if the turgor pressure and wall thick- ness were the same. It appears unlikely that the turgor pressure in the cukaryotic hypha is lower than that in prokaryotes but likely that the stress is borne by a greater thickness of wall. The fact that the larger cross- section surface would sulfce to exert a higher amount of hydrostatic force on the environment may be basic to the fact that eukaryote hyphae typically elongate an order of magnitude faster than those of mycelial prokaryotes and that the wider ones efongate faster. A less obvious difference between the two types of hyphae is the existence of a cytoskeleton in the eukaryote but not in the prokaryote. This difference is exceedingly relevant because the eytoskeleton of the eukaryote is probably involved in the movement of vesicles to the tip of the cell. Although not involved in the discussion below, the cytoskeleton also could be involved in various stress-bearing roles. It could force extension and prevent bulging, 5.1. THE MOVING VESICLE SUPPLY CENTER (THE “HYPHOID” AND “*3-D" MODELS) Earlier models of apical growth have been carefully reviewed by Prosser (1990), Wessels (1986, 1990a, 5), and by Bartnicki-Garcia et al. (1989) and will not be covered here. The latter paper also presented a new model, the “Hyphoid model” that appeared quite attractive and is considered in this section. Tt is concluded here that even in a revised form it is inadequate to account for the apical growth of hyphae. ‘The Bartnicki-Garcia et al. (1989) “moving VSC” model depends on the existence of an analogue to the Spitzenkorper—a vesicle supply center (VSC)—in all eukaryotic hyphae, In many fungi the VSC can bbe associated with the — ytologically observed Spitzenkérper. This body is a collection of vesicles located near the tip of the hypha. In many cases, it consists of the aggregation of a large number of vesicles. New vesicles are continuously recruited from the more distal parts of the cell. Vesicles, perhaps after being altered or matured, continuously leave the Spitzenkérper and move toward and fuse with the wall For the fungal tip to grow the precursors of chitin ‘must be attached to the cell membrane, transported144 ALL. KocH to the outside, and polymerized there. The glucan is partially polymerized in the cytoplasm and then passed through the membrane. There the glucan is gradually crosslinked to itself and to the chitin. Thus, enzymes are needed in the cytoplasm, the membrane, and the growing wall. Mostly, the vesicles arise from the Golgi apparatus, and contain enzymes and sub- strates that have a role in wall formation (see Harold, 1990), ‘The VSC, whether visible as a Spitzenkérper or not, besides serving as a depot for the temporary storage of vesicles with some of the components for wall formation, somehow moves to maintain a fixed distance behind the tip of the hyphae. The Hyphoid model postulates: (i) that it is the controlled move ment of the VSC that sets the rate of fungal tip extension; (ji) that the vesicles move out from the Spitzenkérper in all directions; (ii) they proceed until they encounter the wall; (iv) they fuse with the wall; and (v) this results in the local enlargement of the wall. For simple geometric reasons, the tip receives vesicles at a greater rate because it is close by and therefore enlarges more rapidly. If the Spitzenkérper were a point source; iit emitted vesicles at a constant velocity, NV (see); if the vesicles radiated uniformly in all directions in a plane; if they proceeded in straight lines; if they fused and increased the wall arca locally; and if the VSC moved with a velocity V while remaining at a distance, d behind the tip, then the authors showed that the shape of hyphae would be ‘mathematically described by: 2 =reot(r/d), where d= N/V. In this formulation the center of the cylindrical co-ordinate system is the center of the Spitzenkérper, isthe radial distance and z is the height in the axial; ive. the cylindrical, direction. (Note: I have changed the nomenclature of their co-ordinate system to agree with the mathematical development for other cases of cylindrical growth discussed below and elsewhere.) This equation describes a mathematical curve called the hyphoid. Therefore this term will also be used for this model of tip growth; it has two desirable features mentioned above: one is that a single equation de- scribes the shape of both the pole and the side wall; the other is that it has only a single parameter, d. Many previous experimental studies and the theoreti- cal models for prokaryotes have made a distinction between pole formation and the longation of side walls, and therefore depend on a number of par- ameters simply because there are two sections to the structure. Thus, some carlier models presumed a discrete break between the pole and the side wall. For some bacteria, justification for this division has been given: for example, the sidewall of the Gram-positive rod-shaped bacteria are made by an entirely different process than the pole. For the growing mycelium of either eukaryotes or prokaryotes, such arguments now seem less appropriate. Unfortunately there are three incorrect assump- tions in the derivation of the hyphoid model for fungal growth. First, the published mathematical treatment actually dealt with the two-dimensional analogue of a hypha. The authors felt and gave qualitative arguments that the formula would still apply to growth of a three-dimensional shell. Also their Monte Carlo simulation by computer and the mathematical equation given above were derived from a model which actually calculated the increase Of the cell substance instead of calculating the increase in the surface area surrounding the new cytoplasm. ‘These two problems were later overcome by a more refined analysis by Gierz (unpublished personal com- munication). His results, however, led to:a similar, but by no means identically shaped curve. He derived an ordinary differential equation: dp/dp =(Ci(1 + cos B)* — p?)'?, for the three-dimensional (3-D) model, which can be solved numerically for cylindrically symmetrical hy- hae. Cis an arbitrary constant and in the cylindrical co-ordinate system p is the distance dimension from the origin (placed in the center of the VSC) and f is the angle from the axis of the cylinder. Even though the hyphoid model is unrealistic as presented in Bartnicki-Garcia er al. (1989, 1990) and Bartnicki-Garcia (1990), itis still useful because it is the simplest mathematical framework for describing a tube of constant diameter that ends in a non-pointed tip. In the hyphoid model, the diameter of the hypha at a large distance from the tip is predicted to be 2nd, the distance from the center of the Spitzenkérper to the tip. It also predicts that the diameter, when ‘measured at the level of the SpitzenkOrper, should be ‘one-half the maximal cell diameter. These properties of the curve lead to an easy test for its applicability. measure the diameter of the cylinder at a large distance from the tip, divide by 2x to obtain an estimate of d, and measure back from the tip this distance and then measure the diameter at this point. For the hyphoid equation to accurately apply, this distance should be 0-5 of the diameter of the cylindri- cal region. For the biology that led to the equation to apply, the distance d behind the apex should be at the microscopically observed center of the Spitzenkorper. For the more relevant 3-D model, onty two changes are needed. First, that instead of dividing the diameterHYPHAL GROWTH IN STREPTOMYCETES AND FUNGI 145 by 2x =6283, the maximum diameter should be divided by 8:0 to find d, the distance to be measured back ftom the tip to the center of the Spitzenkérper. ‘Thus for the 3-D model the VSC is proportionately closer to the tip. Second, the diameter at this point should be 1/4 = 0:7854 (instead of 0-5) of the maxi- mum diameter. Thus the 3-D curve proportions are ‘quite different from that of the hyphoid, Tn inspecting the curve fittings of either the hyphoid or the 3-D model to electron micrographs of various fungi, two questions must be considered: is the degree of fit of the equation to the cell shape satisfactory, and is the point at a distance, d, from the tip, in fact, at the center of the SpitzenkOrper? Because there is only one adjustable parameter, the fitting is easily and unambiguously done when the photographic evidence extends far enough into the cylindrical region. Then, the maximum diameter can be measured and set equal to 2nd or 8d for the two variant models. For Polystictus versicolor (figure 5 of Bartnicki-Garcia ‘et al, 1989) the fit for the hyphoid model appears very good, but what T would judge as the center of the SpitzenkSrper is 1-5 times d behind the tip instead of 10 times d; the fit is poorer for the 3-D model. For Armillaria meltea (their figure 7) the hyphoid model fit is good, but the Spitzenkdrper is not quite centered on the origin of the co-ordinate system. In addition to these, I considered three organisms whose electron micrographs or cell dimension, from the work of others, were presented in Harold’s review (1990). Sclerotium rolfsi is blunter than predicted and the center of the Spitzenkérper is a factor of two closer to the tip than predicted by the hyphoid but the fit is alittle bit better by the 3-D model; while Saprolegnia ‘ferax gives a very poor fit indeed. The measurements on Neurospora crassa show it to be blunter than predicted by cither model. Its radius at the presumed center of the co-ordinate system is 30% greater than predicted by the hyphoid model. Simitar statements can be made from other photographs presented in the literature. Thus there does not appear to be exper- imental evidence in favor of one or the other and it may be that the shapes are so variable that they cannot be used to choose between these or other models. Some of the variation could be due also to slightly off-axis electron micrographs 52, THE INDETERMINICITY OF THE MOVING VSC MODELS Both the hyphoid and the -D models were based fon the idea that a vesicle would leave the VSC and travel outward in a random direction, but proceed in straight line. Seemingly, with only slight elaboration of the 3-D model this assumption could be replaced with the assumption that the vesicles leave the Spitzenkérper and travel in an irregular random diffusion path until they collide with the wall. This would further increase the chances of a vesicle en- countering the tip instead of the cylindrical part of the cylinder (see Berg, 1983). It is easy to see that random motion would lead to a narrower cylindrical region for a fixed value of d than predicted by a straight-line model. Thave not succeeded yet in elaborating the theory in detail for diffusive movement for the 3-D case because conceptual difficulties arose with the basic approach. On trying to simulate models that included diffusion, I found I could not ereate a definitive shape, in the sense that the models only generated a specific cell shape when unreasonable constraints were added. In the simulations initially, the shape was indetermi- nant because in my formulation I permitted outward ‘movement of some of the yet unhardened wall, The model used by Bartnicki-Garcia et al, (1989) and its 3-D version is soluble and gives a steady-state stable shape because both are constrained by the implicit assumption that the new material must remain a fixed distance from the cell axis; whereas I had modified the model shown in Fig. 5 to approach that shown in Fig. 6, under the tacit assumption that Wessels “steady state hypothesis”, Harold “soft spot hypoth- esis”; or my own “Variable-T model” were function- ing. I seems reasonable to expect that new wall added near the tip sooner or later moves more peripherally and eventually comes to be no longer part of the tip, but instead part of the cylindrical region. Given this extra flexibility of peripheral movement, a new incre- ‘ment of area in an annulus around the tip may be used, either to change the slope of the tip in that immediate vicinity or to be pushed outward, Eventu- ally it would be moved into the hyphal cylinder. Consequently, other factors can be important in determining the shape of the tip and the velocity of the VSC is not a sufficient condition to define the shape. These factors would probably include: how fast the wall hardens; the relative amounts of move- ‘ment in the radial and in the axial direction as a Funetion of distance from the tip; how the viscoelastic- ity changes with time after formation; and the effect of turgor pressure on growth. Without additional constraints, the system is “underconstrained” (Kuznetsov, 1991) 5.3, THE “SOFT SPOT” MODEL AND THE ROLE OF SURFACE STRESS In the moving VSC models there is a logical deficit because in the absence of vacuoles, surely elongation is not ordinarily autonomous, but is consequent to the success of the organism in accumulating resources146 ALL. KOCH Vesicle transport ns Fio, 5, Eukaryotic strategy for apical grovith according to the ‘moving VSC model of Burinicki-Gateta eta. (1989). Vesicles are made in the body of the organism and move, via eytoskeletal lements to the Spitzenkérper. This organelle, called the vesicle supply center (VSC), is an aggregation of many vesicles and it is ‘assumed that even in those fungi where it cannot be visualized that there sill is 2 VSC. Apparently the vesicles are activated there and then dissociate from the VSC and move toward the wall, Reaching, the wal, they fuse und enable the process of wall expansion. A key point of the Bartnicki-Garcia eta. (1989) made! is that the VSC moves forward at a constant velocity. The model further assures that simply because the Spitzenkdrper is nearer the tip wall than the side wall that more vesicles wil fuse in the polar region. A ‘mathematical theory Teading to the “hyphoid” equation was de- ‘ved by these authors using these assumptions and used to predict the shape of some eukaryotic hyphae quite accurately. In spite of this succes, there are a number of biological and mathematical problems with both the hyphoid and the °3-D" variant of the ‘moving VSC model, although the 3-D model corrects for two ofthe problems of the calie model. A further revision in which the still plastic wall material can move laterally is needed. and processing them to make cytoplasm. Bartnicki- Garcia et al. (1989) pointed out that their theory would operate as well if the Spitzenkdrper was either pushed by the cellular mass of pulled by the tip, but they did not address the question of what could make either movement possible. This connection may be supplied because it takes less work to put new ma- terial into a region that has unpolymerized wall than in other regions where the wall has partially, or completely, set and become “rigidified™. In this sec- tion I will develop the “soft spot” model alone and combine it with the ideas of the Bartnicki-Garcia group in Section 5.5. The idea of Reinhardt (1892; see review by Wessels, 1986) that new wall takes time to harden so that still ‘newer wall can readily be fused, whereas new wall ‘enters less or does not incorporate at all in regions that have had time to fully harden (cross-link) has fay L Droplets of eye newly mixed Epoxy (10 min old) oy See ea ©) Droplets of Epoxy (5 min old) newly mixed epoxy NW eee tutter ster —7 © Desist epoyonina) Senna or AoE Stretched aber sheet —7 Fic. 6. An analogue demonstration ofthe “sot spot” model, The basi soft spot model sat least 100 years old and has been modified and discussed by many workers; ii simply tha the newest portion Of the cell walls soft and ean take up wall precursors in form ‘of material that gradually hardens. Depending on the kinetics and ‘spacial location the several processes, the shape ofthe apical tip is termined. Here is presented an experimental analogue system to ‘show the feasibility ofthis type of model to respond tothe growth needs of the cell. Freshly mixed epoxy is applied to a rubber sheet fF tubing. After an appropriate time, droplets of freshly mixed epoxy are added and itis observed whether they remain on the ‘surface or fuse into i. () is & control in which enough time has been allowed so thatthe band of epoxy has hardened too much and droplets do not fuse. (b) shows that at an earlier time when it was smote plastic, it would accept a droplet of fresh epoxy. (€) shows ‘that inrease tension inthe viscoplastic material favors fusion. This experiment was repeated twice once with "20-minute” epoxy and the socond time with “two-ton™ epoxy with the same results, but for the latter material the times mere differnt and are given in the bbeen revived in modem form. Schreurs et al. (1989) called this model the “soft-spot” hypothesis (F. M. Harold, personal communication, 1992): this is the name that will be used here, It is more descriptive than the Variable-T model that can be used to describe and quantitate such a process. Wessels et al. (1983) and Wessels (19904, 6) developed the “steady state model” which was predicated on three kinds of “observations: (i) that the newly synthesized wall at the tipis chitin in a non-crystalline and non-firillar form; (i) that the new glucan is not alkali-insoluble and has yet formed no branches or cross-links; and (iii) thatHYPHAL GROWTH IN STREPTOMYCETES AND FUNGI 147 stoppage of elongation allowed the wall over the apex to assume a structure indistinguishable from that of the side walls. Those observations are exactly the expectation for a plastic composite material as it is formed, ages and hardens. So from four different approaches the same concept arises. To put these concepts together, the growth of cytoplasm in the body of the cell tends to increase cellular pressure and this in turn could force both tip expansion (and drag the Spitzenkdrper) and lead to the elongation of the cylindrical region. Since the cell's turgor pressure is only experienced at a wall separating the cell from the outside, then the con- trolling force (or at least its intermediate control point) has to be the tension developed in the wall at the tip. Then tension, at feast momentarily, must increase if cellular wall enlargement lags behind vol- tume increase and, vice versa, tension must fall when the pressure is lowered as the result of growth. If tension in the wall facilitates the fusion of vesicles or their action then elongation of the hypha will be linked to cellular growth per se. Thus we must imagine that the partially hardened wall is a vis- coplastic material that not only is deformable by stress but is also rendered more amenable to incorpor- ation of new wall material from recently arrived vesicles, It must further be considered that the very tip is most plastic because it was most recently formed by accretion of new vesicles. Figure 6 shows dramatically a demonstration of the key elements of this “soft spot” model. This exercise ‘makes two points relative to apical growth of fungi: first, that fusion takes place more readily, the younger the age of the wall; and second, that fusion takes place more readily with wall that is under tension. These assumptions can be readily demonstrated with com- mercial “20-minute” or with “two-ton” epoxy prep- arations, Films of about 1 mm thickness and 3 mm in width were made upon a rubber sheet (actually I used a piece of rubber tubing with the second epoxy). The figure shows that the rapidly polymerizing epoxy after only Smin would no longer admit freshly mixed epoxy, but when stretched would even at 10 min allow the fusion of the droplet of newly prepared epoxy. For the second epoxy the same results were obtained but the timescale was prolonged. Such a film of 35-min old material was found to accept a newly mixed droplet of epoxy less well than did a film of a 30-min old material. When the rubber support was stretched at 40 min, it was found that freshly mixed epoxy was more easily accepted and integrated itself into the layer. Consequently, the current version of the “‘soft spot” model is that successful fusions of vesicles occur near the tip; this creates a gradient of the average age of the wall away from the axis, This favors further vesicle fusion (of competent vesicles) to the tip par- ticularly under conditions where the turgor pressure is high, leading to tension in the substratum, This mechanism can account for Reinhardt’s 1892 obscr- vation and the more modem versions. From these considerations, it can be anticipated that a much more elaborate model that includes the mechanical engin- ering of viscoplastic materials will be needed to account for the growth of fungal apices; it will have many parameters that will have to do with the rate of curing of wall, the rate of vesicle fusion, and the action of stresses in favoring fusion and function of the vesicles. 34, TWOLESS LIKELY MECHANISMS, DEPENDENTON AKEY ROLE FOR VACUOLE FORMATION: A DIFFERENT MODEL FOR FUNGAL ELONGATION ‘Vacuoles are quite prominent in some fungi. It can be imagined that various fungi in different stages of their colony growth act in a regulated way to form vacuoles actively and thereby increase the turgor pressure of the cell. This would enlarge the volume of the mycelium beyond that of the cytoplasm alone and may maintain the turgor pressure and force the tip growth process, just as discussed in the previous ‘subsection, but for a different purpose. So it could be that vacuoles are developed for the purpose of allow- ing the hyphal tips to cxplore the soil or wood substratum more effectively Perhaps the surface stress theory is not appropriate for a mycelial organism for which wall growth is not the way to enlarge the cytoplasmic space, but mainly is as a way to move towards a nutrient-rich environ- ment. Pethaps wall growth is autonomous and the cellular contents move toward the tips on cytoskeletal elements, leaving vacuoles as the discarded spaces unfilled with eytoplasm. In such a case, it could be imagined that the force favoring localized tip growth was generated by tension created by the cytoskeletal system, Such an expansive force would be somewhat exceptional: mechano-enzymes and proteins usually function in a contractile way. Recently Money & Harold (1993) have shown that Saprolegnia and ‘oomycetes form a normal hypha in the absence of turgor pressure. Itean also be imagined that the motion is a periodic ‘one, analogous to the way that earthworms, cross country skiers, and possibly myxobacteria, move by sliding (Koch, 19904). Having stated the caveat that mechanical force may push the tip outward, 1 will return to the more likely hypothesis that the turgor pressure is usually the driving force for tip growth,148 ALL. KocH although it would still be interesting to study the relationship between the rate of hyphal elongation of an entire organism, the proportion of vacuolar length, and the tip shape. Also, studies should be made for germinating spores, germlings and very young colonies and the results compare with different por- tions of an older colony that is very necrotic internally and where it is known that the protoplasm is moving peripherally. 5.5. DIFFERENTIATION BETWEEN THE MOVING VSC AND. ‘THE “SOFT SPOT” MODELS FOR FUNGAL ELONGATION, A critical experiment to test for the “soft spot” model is to halt protein synthesis with an agent such as cycloheximide, While cytoplasm synthesis would ‘be quickly halted, wall synthesis should continue for some time because the polysaccharide components could still be formed and transtocated as there would still be adequate reserves of energy and active en- zymes, Hopefully, preliminary experiments would ¢s- tablish just how fast and how completely protein synthesis can be stopped and how quickly it will resume after the inhibitor of protein synthesis is removed. If the moving VSC models alone function, after a temporary blockade there should be a delayed inhibition of elongation, but no other effect on mor- phology is to be expected. For the “soft spot” model functioning alone, wall growth would temporarily stop, but the curing of the polar wall would continue. When normal growth conditions resumed some, or all, of the wall would now be too hard to accept vesicles containing wall precursors or the polymeriz~ ing enzymes. Thus previous blockade, depending how extensive it has been, might lead to: (@) branching elsewhere; (i) new branch formation at the old tip; Gii) blow outs of the very terminus of the tip, or (iv) growth emanating from the terminus of the tip leading to a permanent, deep constriction observable at later times. 1 recognize that similar experiments have been done a century ago: originally, apical growth was stopped by lowering the external osmotic pressure, These results could be taken to be consistent with the “soft spot” model, but they do not clearly differentiate between the moving VSC and the “soft spot” model. It could be that a critical and clear experiment would result when the selective biochemi- cal inhibitors now available arc used. It is expected that such experiments would not eliminate one or the other, but show that they both had a role. 5:6, SYNTHESIS OF THE MOVING VSC AND THE SOFT SPOT MoDEL Coupling of movement of the VSC to the move- ‘ment of the tip and the rate of generation of vesicles and biomass generally are matters needed for a full theory. The VCS is a reality in those fungi that have evident Spitzenkérper and clearly in these cases have a role in hyphal growth; moreover, itis reasonable to assume that there are “virtual Spitzenkdrperen” gen erally. The component vesicles are largely formed in appreciable amounts in the bulk of the cytoplasm behind the Spitzenkérper and they are moved upon cytoskeletal elements to the Spitzenkérper: in this motion they themselves or the process of microtubule elongation could force the. movement of the Spitzenkérper. Under such a scheme it would also be, necessary to couple the motion in some sort of a feedback loop to the elongation of the hypha. Both kinds of models have their virtues. Figure 7 attempts to combine them, but it does not include all the obvious elements. Additional factors that must be incorporated into a realistic model concern whether vesicles follow random or straight paths from the SC, the kinetics of deposition of vesicles, the kinetics of the change of the degree of viscoplasticity, the influence of turgor pressure in causing the vesicles to fuse with the wall, and the degree with which one packet of viscoplastic material can move relative to ‘another. In fact, the entire science and engineering of viscoplastic material must be considered. Stretched, cured wall Stretching Spitzonkirper Fig. 7. A possible synthesis of the moving VSC and the “sont- spot" model: Figure § has been adumbrated to add some of the factors that may determine apical Gp shape. These include the plasticity of the new wall and its ability to yield to the turgor pressure. I is not x model, but an outine of a model. Additional actors that must be measured in in rio and in in wiro studies concern whether vesicles follow random or straight paths from the VSC, the contro! of movement of the YSC, the kinetics of deposition of vesicles, the kinetics of the change of degree of visopastcity, and the influence of turgor pressure in eausing the Vices to fuse with the wall