Equilibrium and kinetic analysis of microtubule assembly in the presence of guanosine diphosphate

In this paper we expand upon a previously reported observation of the effects of GDP on microtubule assembly. A ratio of GDP to GTP of ten (1 mm-GDP and 0.1 mm-GTP) is generally sufficient to completely block microtubule assembly, but only limited depolymerization is induced if GDP is added after assembly has reached a plateau in the presence of GTP. When added during polymerization, GDP arrests further elongation, and greater steady-state levels of assembly are obtained the later the time of addition of GDP. To explain this behavior we examined the rates of assembly and disassembly and the apparent critical concentration (C₀) of tubulin in the presence of GDP. GDP-tubulin polymerizes very slowly as compared to GTP-tubulin, while depolymerization rates, as determined by dilution, are nearly identical in GTP and GDP. The C₀ value calculated from the assembly and disassembly rates in GTP is within experimental error of the C₀ value at steady-state determined directly. In the presence of GDP, however, the C₀ value calculated from rate measurements is at least 60 times greater than that determined by equilibrium analysis. Our results indicate that the net assembly rate in GDP is not a valid measure of the reaction occurring at steady-state. A limited amount of depolymerization may occur upon addition of GDP to microtubules, and this appears to be due to a decrease in the fraction of protein able to participate in the polymerization reaction. The amount of tubulin “inactivated” by GDP is increased by the removal of microtubule-associated proteins. GDP-tubulin will stabilize existing microtubules, even when its polymerization cannot be demonstrated. These results are inconsistent with present models of microtubule assembly, and a new model involving co-operative interaction of microtubule-associated protein-tubulin oligomers at microtubule ends is proposed.     

Dephosphorylation of tubulin-bound guanosine triphosphate during microtubule assembly.

1. Tubulin purified from porcine brain in the presence of GTP contained 0.16 mole of GDP and 0.73 mole of GTP per 60,000 g of protein. 2. Microtubules reconstituted from the purified tubulin contained 0.43 mole of GDP and 0.41 mole of GTP per 60,000 g of protein. Guanine nucleotide bound to the exchangeable site of tubulin was converted to GDP during microtubule assembly, while GTP at the non-exchangeable site remained intact. 3. Guanine nucleotide which had been bound to the exchangeable site of tubulin before microtubule assembly was also exchangeable during disassembly.     

Nucleotide release from tubulin and nucleoside-5'-diphosphate kinase action in microtubule assembly.

ATP and UTP support microtubule assembly through the action of brain nucleoside-5'-diphosphate kinase on GDP. Penningroth and Kirschner (1977) J. Mol. Biol. 115, 643-673) have proposed that microtubule assembly may occur by either of two mechanisms: indirectly, through nucleoside-5'-diphosphate kinase-catalyzed phosphorylation of uncomplexed GDP and directly by nucleoside-5'-diphosphate kinase-mediated transphosphorylation of tubulin-bound GDP at low tubulin concentrations. We find the rates of GDP and GTP release (0.68 and 0.32 min-1, respectively) are sufficiently fast relative to assembly to permit GDP release, phosphorylation, and GTP binding as the sole mechanism of nucleoside-5'-diphosphate kinase action in microtubule assembly. Computer simulation studies accord with the conclusion that GDP release is rapid relative to microtubule assembly. The specific activity of the nucleoside-5'-diphosphate kinase is 1.7 nmol/min/mg of microtubular protein under the conditions studied. Pulse-chase experiments with tubulin . [14C]GDP complex and the rapidity of GDP phosphorylation by the kinase are in agreement with this scheme. Finally, it was observed that the extent and rate of microtubule assembly depends upon the [ATP]/[ADP] ratio.     

Control of nucleation in microtubule self-assembly

The inhibition of the rate and amplitude of assembly of microtubule protein at low GTP concentration is shown by measurement of microtubule length distributions to be due to the suppression of microtubule nucleation. This inhibitory effect is enhanced by GDP added before assembly, but can be overcome by a number of molecules such as pyrophosphate or ADP. The selective inhibition of nucleation by GDP in vitro, which occurs in addition to inhibition of elongation, could provide a mechanism for the control of spontaneous microtubule nucleation in vivo.     

Tubulin-nucleotide interactions. Effects of removal of exchangeable guanine nucleotide on protein conformation and microtubule assembly.

The removal of tightly bound GDP from the exchangeable nucleotide-binding site of tubulin has been performed with alkaline phosphatase under conditions which essentially retain the assembly properties of the protein. When microtubule protein is treated with alkaline phosphatase, nucleotide is selectively removed from tubulin dimer rather than from MAP (microtubule-associated protein)-containing oligomeric species. Tubulin devoid of E-site (the exchangeable nucleotide-binding site of the tubulin dimer) nucleotide shows enhanced proteolytic susceptibility of the beta-subunit to thermolysin and decreased protein stability, consistent with nucleotide removal causing changes in protein tertiary structure. Pyrophosphate ion (3 mM) is able to promote formation of normal microtubules in the complete absence of GTP by incubation at 37 degrees C either with nucleotide-depleted microtubule protein or with nucleotide-depleted tubulin dimer to which MAPs have been added. The resulting microtubules contain up to 80% of tubulin lacking E-site nucleotide. In addition to its effects on nucleation, pyrophosphate competes weakly with GDP bound at the E-site. It is deduced that binding of pyrophosphate at a vacant E-site can promote microtubule assembly. The minimum structural requirement for ligands to induce tubulin assembly apparently involves charge neutralization at the E-site by bidentate ligation, which stabilizes protein domains in a favourable orientation for promoting the supramolecular protein-protein interactions involved in microtubule formation.     

Effects of GDP on microtubules at steady state

We have re-examined the effect of varying GDP concentrations on the kinetics of GTP-induced assembly of microtubules from microtubule protein, and on the elongation of pre-existing microtubules subjected to a temperature jump relaxation from 21.5 to 37 degrees C. The assembly kinetics follow a simple model for assembly which involves a fast equilibrium of tubulin-GTP and tubulin-GDP coupled to the elongation process due to tubulin-GTP. The initial rate of the relaxation process is found to be dependent upon the GTP/GDP ratio, in confirmation of the results of Engelborghs and Van Houtte (Biophys. Chem. 14 (1981) 195). As an alternative to the interpretation previously advanced by them, involving modification of the reactivity of microtubule ENDs by GDP, we show that this result is consistent with the above model with one reasonable modification, namely, that the ratio of the affinities of tubulin for GTP and GDP should vary with temperature. The analysis shows that a decrease in this ratio of approx. 2-fold at 37 degrees C accounts for the observed effects. We conclude that more complex mechanisms involving consideration of modification of the reactivity of microtubule ENDs by GDP are not required to explain these results. This finding has important implications for current models of GDP-induced microtubule disassembly.     

Association of nucleosidediphosphate kinase with microtubules.

A nucleosidediphosphate kinase activity (EC 2.7.4.6) which phosphorylates GDP to GTP is present in bovine brain microtubule protein prepared by cycles of assembly-disassembly. This activity persists through 5 cycles of assembly-disassembly and sediments with microtubules in sucrose density gradients, but is not associated with the tubulin dimer. It is proposed that the kinase is an integral part of the microtubule and is therefore a microtubule associated protein (MAP). Several isozymes of nucleosidediphosphate kinase exist in our preparations with a pI 7.6 form predominant. It may be speculated that this enzyme affects tubulin assembly $\text{in vivo}$ by modulating the $\text{GTP}\text{GDP}$ ratio in the microtubule environment.     

Incorporation of GDP-tubulin during elongation of microtubules in vitro

Removal of GDP from tubulin E-site is not obligatory for the in vitro assembly of microtubule protein in 0.5 mM GMPPCP. This assembly, which is significantly enhanced by glycerol, produces microtubules of normal morphology and with normal composition of microtubule-associated proteins (MAPs). [3H]-GDP initially present at the E-site is shown to be incorporated directly into microtubules during assembly; this incorporation, maximally 60% of the assembled polymer, is dependent upon MAPs. These results are consistent with oligomeric species composed principally of GDP-tubulin plus MAPs, being incorporated directly into microtubules. The finding that stoichiometric GTP-tubulin formation is not an essential prerequisite for microtubule assembly may have important implications for the energetics of microtubule formation.     

Thermodynamic and structural analysis of microtubule assembly: the role of GTP hydrolysis.

Different models have been proposed that link the tubulin heterodimer nucleotide content and the role of GTP hydrolysis with microtubule assembly and dynamics. Here we compare the thermodynamics of microtubule assembly as a function of nucleotide content by van't Hoff analysis. The thermodynamic parameters of tubulin assembly in 30-100 mM piperazine-N,N'-bis(2-ethanesulfonic acid), 1 mM MgSO4, 2 mM EGTA, pH 6.9, in the presence of a weakly hydrolyzable analog, GMPCPP, the dinucleotide analog GMPCP plus 2 M glycerol, and GTP plus 2 M glycerol were obtained together with data for taxol-GTP/GDP tubulin assembly (GMPCPP and GMPCP are the GTP and GDP nucleotide analogs where the alpha beta oxygen has been replaced by a methylene, -CH2-). All of the processes studied are characterized by a positive enthalpy, a positive entropy, and a large, negative heat capacity change. GMPCP-induced assembly has the largest negative heat capacity change and GMPCPP has the second largest, whereas GTP/2 M glycerol- and taxol-induced assembly have more positive values, respectively. A large, negative heat capacity is most consistent with the burial of water-accessible hydrophobic surface area, which gives rise to the release of bound water. The heat capacity changes observed with GTP/2 M glycerol-induced and with taxol-induced assembly are very similar, -790 +/- 190 cal/mol/k, and correspond to the burial of 3330 +/- 820 A2 of nonpolar surface area. This value is shown to be very similar to an estimate of the buried nonpolar surface in a reconstructed microtubule lattice. Polymerization data from GMPCP- and GMPCPP-induced assembly are consistent with buried nonpolar surface areas that are 3 and 6 times larger. A linear enthalpy-entropy and enthalpy-free energy plot for tubulin polymerization reactions verifies that enthalpy-entropy compensation for this system is based upon true biochemical correlation, most likely corresponding to a dominant hydrophobic effect. Entropy analysis suggests that assembly with GTP/2 M glycerol and with taxol is consistent with conformational rearrangements in 3-6% of the total amino acids in the heterodimer. In addition, taxol binding contributes to the thermodynamics of the overall process by reducing the delta H degree and delta S degree for microtubule assembly. In the presence of GMPCPP or GMPCP, tubulin subunits associate with extensive conformational rearrangement, corresponding to 10% and 26% of the total amino acids in the heterodimer, respectively, which gives rise to a large loss of configurational entropy. An alternative, and probably preferable, interpretation of these data is that, especially with GMPCP-tubulin, additional isomerization or protonation events are induced by the presence of the methylene moiety and linked to microtubule assembly. Structural analysis shows that GTP hydrolysis is not required for sheet closure into a microtubule cylinder, but only increases the probability of this event occurring. Sheet extensions and sheet polymers appear to have a similar average length under various conditions, suggesting that the minimum cooperative unit for closure of sheets into a microtubule cylinder is approximately 400 nm long. Because of their low level of occurrence, sheets are not expected to significantly affect the thermodynamics of assembly.     

In vivo and in vitro studies on the role of HMW-MAPs in taxol-induced microtubule bundling

The involvement of high molecular weight microtubule-associated proteins (HMW-MAPs) in the process of taxol-induced microtubule bundling has been studied using immunofluorescence and electron microscopy. Immunofluorescence microscopy shows that HMW-MAPs are released from microtubules in granulosa cells which have been extracted in a Triton X-100 microtubule-stabilizing buffer (T-MTSB), unless the cells are pretreated with taxol. 1.0 microM taxol treatment for 48 h results in microtubule bundle formation and the retention of HMW-MAPs in these cells upon extraction with T-MTSB. Electron microscopy demonstrates that microtubules in control cytoskeletons are devoid of surface structures whereas the microtubules in taxol-treated cytoskeletons are decorated by globular particles of a mean diameter of 19.5 nm. The assembly of 3 X cycled whole microtubule protein (tubulin plus associated proteins) in vitro in the presence of 1.0 microM taxol, results in the formation of closely packed microtubules decorated with irregularly spaced globular particles, similar in size to those observed in cytoskeletons of taxol-treated granulosa cells. Microtubules assembled in vitro in the absence of taxol display prominent filamentous extensions from the microtubule surface and center-to-center spacings greater than that observed for microtubules assembled in the presence of taxol. Brain microtubule protein was purified into 6 s and HMW-MAP-enriched fractions, and the effects of taxol on the assembly and morphology of these fractions, separately or in combination, were examined. Microtubules assembled from 6 s tubulin alone or 6 s tubulin plus taxol (without HMW-MAPs) were short, free structures whereas those formed in the presence of taxol from 6 s tubulin and a HMW-MAP-enriched fraction were extensively crosslinked into aggregates. These data suggest that taxol induces microtubule bundling by stabilizing the association of HMW-MAPs with the microtubule surface which promotes lateral aggregation.     

Guanosine-5′-triphosphate hydrolysis and tubulin polymerization

Summary GTP hydrolysis associated with polymerization is a distinctive feature of microtubule assembly. This reaction may be fundamentally linked to the dynamic properties of microtubules in vivo. Kinetic analysis of the connection between microtubule assembly and associated GTP hydrolysis indicates that these two events are kinetically uncoupled, GTP hydrolysis occurring after tubulin incorporation in the microtubule. As a consequence, the combination of the diffusionnal incorporation of GTP in microtubules at steady-state and of subsequent GTP hydrolysis results in the formation of a steady-state GTP cap at microtubule ends. The interplay between GTP and GDP at microtubule ends is examined. Inhibition by GDP of steady-state GTP hydrolysis at microtubule ends and of microtubule elongation is understood within a tight reversible binding of GDP at microtubule ends generating inactive elongation sites. Nucleotides are freely exchangeable at microtubule ends. This result indicates that the nature of the nucleotide present at microtubule ends must be considered in a model for microtubule assembly.These data are pooled in order to define the general features of a model describing microtubule assembly and treadmilling in terms somewhat different from previously proposed models.     

Effect of S-100 protein on assembly of brain microtubule proteins in vitro

S-100 protein inhibits the assembly of brain microtubule proteins in vitro in the presence of 10 microM free Ca2+. The S-100 effect is generally greater on the rate than on the extent of assembly, and even greater as the microtubule protein concentration decreases and the time of preincubation between S-100 and microtubule proteins before GTP addition increases, at a given S-100/tubulin dimer molar ratio. The S-100 effect is greatly enhanced in the presence of physiological concentrations of K+ and is completely reversed by EGTA.     

Nek2B stimulates zygotic centrosome assembly in Xenopus laevis in a kinase-independent manner

Pronuclear migration and formation of the first mitotic spindle depend upon assembly of a functional zygotic centrosome. For most animals, this involves both paternal and maternal contributions as sperm basal bodies are converted into centrosomes competent for microtubule nucleation through recruitment of egg proteins. Nek2B is a vertebrate NIMA-related protein kinase required for centrosome assembly, as its depletion from egg extracts delays microtubule aster formation from sperm basal bodies. Using Xenopus as a model system, we now show that protein expression of Nek2B begins during mid-oogenesis and increases further upon oocyte maturation. This is regulated, at least in part, at the level of protein translation. Nek2B protein is weakly phosphorylated in mitotic egg extracts but its recruitment to the sperm basal body, which occurs independently of its kinase activity, stimulates its phosphorylation, possibly through sequestration from a phosphatase present in mitotic egg cytoplasm. Importantly, although Nek2B is not required to organize acentrosomal microtubule asters, we show that addition of either active or kinase-dead recombinant Nek2B can restore centrosome assembly in a dose-dependent manner to a depleted extract. These results support a model in which maternal Nek2B acts to promote assembly of a functional zygotic centrosome in a kinase-independent manner.     

Regulation of the microtubule steady state in vitro by ATP.

ATP increases microtubule steady state assembly and disassembly rates in vitro in a concentration-dependent manner. Bovine brain microtubules, composed of 75% tubulin and 25% high molecular weight microtubule-associated proteins (MAPs), were purified by three cycles of assembly and disassembly in the absence of ATP. When assembled to steady state, these microtubules add dimers at one end and lose them at the other in a unidirectional assembly-disassembly process. In the presence of 1.0 mM ATP the unidirectional flow of tubulin from one end of the microtubules to the other increases as much as 20 fold, as revealed by loss of 3H-GTP from uniformly labeled microtubules under GTP chase conditions and by the rate of disassembly following addition of 50 microM podophyllotoxin. UTP, CTP and 5' adenylylimidodiphosphate (AMP-PNP) cannot substitute for ATP in producing this effect. Furthermore, the increase in steady state flow rate persists afer ATP is removed. Thus microtubules assembled in ATP and centrifuged through sucrose cushions to separate them from nucleotides continue to exhibit increased rates in the next assembly cycle in the absence of ATP. It is possible that an ATP-dependent microtubule protein kinase is responsible for the observed increase in tubulin flow rate. A kinase activity associated with brain MAPs has been reported to be cAMP-dependent (Sloboda et al., 1975). We have found an adenylate cyclase activity associated with these microtubules. Whether the adenylate cyclase is a contaminant or due to a specific microtubules-associated protein, and whether its activity is functionally linked to the increased rate of assembly and disassembly in the presence of ATP, remain to be determined.     

Nucleoside diphosphate kinase from brain. Purification and effect on microtubule assembly in vitro

Tubulin strictly requires GTP for its polymerization. Nevertheless, microtubule assembly can be observed in the presence of ATP as the only nucleotide triphosphate, due to the nucleoside diphosphate kinase (NDP kinase) present in microtubule preparations, and which phosphorylates the GDP into GTP. We have purified this enzyme from pig brain to homogeneity, and shown that its relative mass is close to 100 000 in its native state, and 17 000 under denaturing conditions. Therefore it is probably a hexamer, as previously shown for the enzyme from other sources, and also presents a microheterogeneity, with the major isoforms between pI 5.0 and 6.0. The enzyme is transiently phosphorylated during catalysis, as expected within a ping-pong bi-bi mechanism. The effect of the NDP kinase on pure tubulin polymerization was studied: in the presence of NDP kinase, the lag time observed in the kinetics of microtubule assembly was shorter and the final extent of assembly was unchanged. The effect of the enzyme was observed at enzyme concentrations 900-fold lower than tubulin concentration, which shows that the NDP kinase acts catalytically. Kinetic data show that the catalytic effect of the NDP kinase is faster than the rate of nucleotide exchange on tubulin under the same conditions. This result demonstrates that the tubulin-GDP complex itself is a substrate for the enzyme, which may indicate that the GDP bound to tubulin at the E site is exposed on the surface of dimeric tubulin.     

Constitutive dynein activity in She1 mutants reveals differences in microtubule attachment at the yeast spindle pole body

The organization of microtubules is determined in most cells by a microtubule-organizing center, which nucleates microtubule assembly and anchors their minus ends. In Saccharomyces cerevisiae cells lacking She1, cytoplasmic microtubules detach from the spindle pole body at high rates. Increased rates of detachment depend on dynein activity, supporting previous evidence that She1 inhibits dynein. Detachment rates are higher in G1 than in metaphase cells, and we show that this is primarily due to differences in the strengths of microtubule attachment to the spindle pole body during these stages of the cell cycle. The minus ends of detached microtubules are stabilized by the presence of γ-tubulin and Spc72, a protein that tethers the γ-tubulin complex to the spindle pole body. A Spc72-Kar1 fusion protein suppresses detachment in G1 cells, indicating that the interaction between these two proteins is critical to microtubule anchoring. Overexpression of She1 inhibits the loading of dynactin components, but not dynein, onto microtubule plus ends. In addition, She1 binds directly to microtubules in vitro, so it may compete with dynactin for access to microtubules. Overall, these results indicate that inhibition of dynein activity by She1 is important to prevent excessive detachment of cytoplasmic microtubules, particularly in G1 cells.     

Microtubules and nucleoside diphosphate kinase. Comparison of kinetics of GTP- and CTP-induced assembly.

The kinetics of assembly of MAP2-tubulin microtubule protein were examined as a function of the GTP concentration in order to test the hypothesis that CTP-induced assembly results from the generation of GTP by nucleoside diphosphate kinase. These studies show that (a) there is no assembly below a minimum GTP concentration and that this represents a nucleation requirement, (b) the rate of elongation is inconsistent with a single assembly-species, and (c) the elongation rate increases markedly as the GTP concentration is raised, although GTP is not absolutely required for elongation. These assembly kinetics have been compared with those with increasing CTP concentrations, by using microtubule protein containing a very low nucleoside diphosphate kinase activity of known substrate specificity. Neither nucleation nor the observed rates of elongation can be attributed to the formation of GTP, either (a) in terms of the generation of free GTP and subsequent binding to tubulin or (b) by the direct charging of GDP bound to the tubulin exchangeable site. The results show that nucleoside diphosphate kinase is not required for CTP-induced microtubule assembly, and suggest that CTP directly effects microtubule assembly.     

Effects of pH on tubulin-nucleotide interactions

Significant GTP-independent, temperature-dependent turbidity development occurs with purified tubulin stored in the absence of unbound nucleotide, and this can be minimized with a higher reaction pH. Since microtubule assembly is optimal at lower pH values, we examined pH effects on tubulin-nucleotide interactions. While the lowest concentration of GTP required for assembly changed little, GDP was more inhibitory at higher pH values. The amounts of exogenous GTP bound to tubulin at all pH values were similar, but the amounts of exogenous GDP bound and endogenous GDP (i.e., GDP originally bound in the exchangeable site) retained by tubulin rose as reaction pH increased. Endogenous GDP was more efficiently displaced by exogenous GTP than GDP at all pH values, but displacement by GTP was 10-15% greater at pH 6 than at pH 7. Dissociation constants for GDP and GTP were about 1.0 microM at pH 6 and 0.02 microM at pH 7. A small increase in the affinity of GDP relative to that of GTP occurs at pH 7 as compared to pH 6, together with a 50-fold absolute increase in the affinity of both nucleotides for tubulin at pH 7. The time courses of microtubule assembly and GTP hydrolysis were compared at pH 6 and pH 7. At pH 6, the two reactions were simultaneous in onset and initially stoichiometric. At pH 7, although the reactions began simultaneously, hydrolysis seemed to lag substantially behind assembly. Unhydrolyzed radiolabeled GTP was not incorporated into microtubules, however, indicating that GTP hydrolysis is actually closely coupled to assembly. The apparent lag in hydrolysis probably results from a methodological artifact rather than incorporation of GTP into the microtubule with delayed hydrolysis.     

Interrelationships of tubulin-GDP and tubulin-GTP in microtubule assembly

We previously reported that direct incorporation of GDP (i.e., without an initial hydrolysis of GTP) into microtubules occurs throughout an assembly cycle in a constant proportion. The exact proportion varied with reaction conditions, becoming greater under all conditions in which tubulin-GDP increased relative to tubulin-GTP (low Mg2+ and GTP concentrations, high tubulin concentrations, and in the presence of exogenous GDP). These findings led us to explore further interrelationships of tubulin-GDP and tubulin-GTP in microtubule assembly. We have now determined the minimum amount of tubulin-GTP required for the initiation of microtubule assembly and the relative efficiency with which tubulin-GDP participates in microtubule elongation. When GTP, GDP, and tubulin concentrations were varied at a constant Mg2+ concentration (0.2 mM), initiation of assembly required that 35% of the nucleotide-bearing tubulin be in the form of tubulin-GTP, and incorporation of tubulin-GDP into microtubules during elongation was only 60% as efficient as would be predicted on the basis of its proportional concentration in the reaction mixtures. Very different results were obtained when the Mg2+ concentration was varied. Even though Mg2+ enhances the binding of GTP to tubulin (the equilibrium constant for the exchange of GTP for GDP was 0.2 in the absence of exogenous Mg2+, 3 with 0.2 mM Mg2+, 5 with 0.5 mM Mg2+, and 11 with 2 and 4 mM Mg2+), as Mg2+ was increased the proportion of tubulin-GTP required for the initiation of microtubule assembly rose greatly, and the direct incorporation of tubulin-GDP into microtubules during elongation became progressively more efficient. In the absence of exogenous Mg2+, only 20% tubulin-GTP was required for initiation, and tubulin-GDP was directly incorporated into microtubules half as efficiently as would be predicted on the basis of its concentration in the reaction mixture. At the highest Mg2+ concentration examined (4 mM), 80% tubulin-GTP was required for initiation of assembly, and tubulin-GDP was incorporated into microtubules as efficiently as tubulin-GTP.     

Free subunits of ribulose-1,5-bisphosphate carboxylase in pea leaves

Pea leaves, supplied with [³⁵S]methionine, were homogenized and a crude hypotonic soluble fraction was centrifuged on sucrose gradients to separate fully assembled ribulose-1,5-biphosphate (RuBP) carboxylase from any free or partially assembled carboxylase subunits. Slowly sedimenting subunits of the enzyme were identified in upper fractions of the sucrose gradient, using polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate (SDS), isoelectric focussing, and immune precipitation. The presence of these subunits in low molecular weight form was shown not to be due to artefactual dissociation of the enzyme. It is suggested that these subunits are related to the assembly of RuBP carboxylase.