Isoprenylation of an inhibitory G protein alpha subunit occurs only upon mutagenesis of the carboxyl terminus

We transfected COS cells with expression vectors for the wild-type G protein alpha i1 subunit (pWT) and for mutated alpha i1 subunits, including the nonmyristylated glycine 2 to alanine mutant (pGA) and mutants in which the carboxyl termini of pWT and pGA were changed from CGLF to CVLS (pCVLS and pGA-CVLS, respectively). Immunoblot analysis of transfected COS cells with an antibody to residues 159-168 of the alpha i1 protein indicated that all four proteins were expressed. Unlike the WT and GA proteins, both CVLS mutant proteins failed to react with an antibody specific for the carboxyl terminus and failed to undergo pertussis toxin-catalyzed ADP-ribosylation. Analysis of COS cell lysates after [3H]mevalonic acid labeling indicated that specific incorporation of radioactivity occurred only in the alpha i1 subunits with the CVLS mutation. Immuno-precipitation of COS cell fractions after labeling with [35S]methionine indicated that both WT and CVLS mutant proteins were localized predominantly in the particulate fraction, whereas GA and GA-CVLS mutant proteins were found primarily in the soluble fraction. These results directly demonstrate that the carboxyl-terminal sequence, CGLF, is incapable of leading to isoprenylation but that alteration of two residues (glycine to valine, phenylalanine to serine) is sufficient to promote isoprenylation.     

The amino terminus of G protein alpha subunits is required for interaction with beta gamma

The guanine nucleotide-binding proteins (G proteins), which transduce hormonal and light signals across the plasma membrane, are heterotrimers composed of alpha, beta, and gamma subunits. Activation of G proteins by guanine nucleotides is accompanied by dissociation of the heterotrimer: G + alpha.beta.gamma in equilibrium alpha G + beta.gamma. Brain contains several G proteins of which the most abundant are alpha 39.beta.gamma and alpha 41.beta.gamma. We have used proteolysis by trypsin to study the functional domains of the alpha subunits. In the presence of guanosine 5'-(3-O-thio)triphosphate, trypsin removes a 2-kDa peptide from the amino terminus of these proteins (Hurley, J. B., Simon, M. I., Teplow, D. B., Robishaw, J. D., and Gilman, A. G. (1984) Science 226, 860-862; Winslow, J. W., Van Amsterdam, J. R., and Neer, E. J. (1986) J. Biol. Chem. 261, 7571-7579). Tryptic cleavage does not affect the GTPase activity of the truncated molecule nor the apparent Km for GTP. However, removal of the 2-kDa amino-terminal peptide prevents association of the alpha subunits with beta.gamma. Since the apparent substrate for pertussis toxin-catalyzed ADP-ribosylation is the alpha.beta.gamma heterotrimer, the trypsin-cleaved alpha subunit is not a substrate for the toxin. Digestion of the carboxyl terminus of alpha 39 with carboxypeptidase A prevents ADP-ribosylation by pertussis toxin but does not interfere with the formation of alpha 39.beta.gamma heterotrimers. We do not yet know whether the amino-terminal region of alpha 39 interacts with beta gamma directly or whether it is necessary to maintain a conformation of alpha 39 which is required for heterotrimer formation. Further studies are needed to define the nature of the contracts between alpha and beta gamma subunits since understanding the structural basis for their reversible interaction is fundamental to understanding their function.     

Selective inhibition of heterotrimeric Gs signaling. Targeting the receptor-G protein interface using a peptide minigene encoding the Galpha(s) carboxyl terminus

The blockade of heptahelical receptor coupling to heterotrimeric G proteins by the expression of peptides derived from G protein Galpha subunits represents a novel means of simultaneously inhibiting signals arising from multiple receptors that share a common G protein pool. Here we examined the mechanism of action and functional consequences of expression of an 83-amino acid polypeptide derived from the carboxyl terminus of Galpha(s) (GsCT). In membranes prepared from GsCT-expressing cells, the peptide blocked high affinity agonist binding to beta(2) adrenergic receptors (AR) and inhibited beta(2)AR-induced [35S]GTPgammaS loading of Galpha(s). GsCT expression inhibited beta(2)AR- and dopamine D(1A) receptor-mediated cAMP production, without affecting the cellular response to cholera toxin or forskolin, indicating that the peptide inhibited receptor-G(s) coupling without impairing G protein or adenylyl cyclase function. [35S]GTPgammaS loading of Galpha(q/11) by alpha(1B)ARs and Galpha(i) by alpha(2A)ARs and G(q/11)- or G(i)-mediated phosphatidylinositol hydrolysis was unaffected, indicating that the inhibitory effects of GsCT were selective for G(s). We next employed the GsCT construct to examine the complex role of G(s) in regulation of the ERK mitogen-activated protein kinase cascade, where activation of the cAMP-dependent protein kinase (PKA) pathway reportedly produces both stimulatory and inhibitory effects on heptahelical receptor-mediated ERK activation. For the beta(2)AR in HEK-293 cells, where PKA activity is required for ERK activation, expression of GsCT caused a net inhibition of ERK activation. In contrast, alpha(2A)AR-mediated ERK activation in COS-7 cells was enhanced by GsCT expression, consistent with the relief of a downstream inhibitory effect of PKA. ERK activation by the G(q/11)-coupled alpha(1B)AR was unaffected by GsCT. These findings suggest that peptide G protein inhibitors can provide insights into the complex interplay between G protein pools in cellular regulation.     

RAS2 protein of Saccharomyces cerevisiae is methyl-esterified at its carboxyl terminus

Yeast and mammalian RAS gene products are GTP-binding proteins that are posttranslationally localized to the inner surface of the plasma membrane. This localization is accomplished by the addition of a lipid moiety to a conserved cysteine residue close to the carboxyl terminus. In a previous report we showed that the mammalia Ha-ras protein is also modified posttranslationally by methyl esterification. Here we show that the yeast RAS2 protein is posttranslationally modified by methyl esterification at or near the carboxyl terminus. We also present evidence indicating that the methyl ester is linked to the conserved cysteine residue, implying that RAS2 protein is cleaved to expose this cysteine as the carboxyl-terminal residue. This maturation pathway may be shared by a family of proteins that are initially synthesized as soluble proteins and must become membrane-localized to function.     

G protein activation is prerequisite for functional coupling between Galpha/Gbetagamma and tubulin/microtubules

Heterotrimeric G proteins participate in signal transduction by transferring signals from cell surface receptors to intracellular effector molecules. Interestingly, recent results suggest that G proteins also interact with microtubules and participate in cell division and differentiation. It has been shown earlier that both alpha and betagamma subunits of G proteins modulate microtubule assembly in vitro. Since G protein activation and subsequent dissociation of alpha and betagamma subunits are necessary for G proteins to participate in signaling processes, here we asked if similar activation is required for modulation of microtubule assembly by G proteins. We reconstituted Galphabetagamma heterotrimer from myristoylated-Galpha and prenylated-Gbetagamma, and found that the heterotrimer blocks Gi1alpha activation of tubulin GTPase and inhibits the ability of Gbeta1gamma2 to promote in vitro microtubule assembly. Results suggest that G protein activation is required for functional coupling between Galpha/Gbetagamma and tubulin/microtubules, and supports the notion that regulation of microtubules is an integral component of G protein mediated signaling.     

Regulator of G protein signaling Z1 (RGSZ1) interacts with Galpha i subunits and regulates Galpha i-mediated cell signaling

Regulator of G protein signaling (RGS) proteins constitute a family of over 20 proteins that negatively regulate heterotrimeric G protein-coupled receptor signaling pathways by enhancing endogenous GTPase activities of G protein alpha subunits. RGSZ1, one of the RGS proteins specifically localized to the brain, has been cloned previously and described as a selective GTPase accelerating protein for Galpha(z) subunit. Here, we employed several methods to provide new evidence that RGSZ1 interacts not only with Galpha(z,) but also with Galpha(i), as supported by in vitro binding assays and functional studies. Using glutathione S-transferase fusion protein pull-down assays, glutathione S-transferase-RGSZ1 protein was shown to bind (35)S-labeled Galpha(i1) protein in an AlF(4)(-)dependent manner. The interaction between RGSZ1 and Galpha(i) was confirmed further by co-immunoprecipitation studies and yeast two-hybrid experiments using a quantitative luciferase reporter gene. Extending these observations to functional studies, RGSZ1 accelerated endogenous GTPase activity of Galpha(i1) in single-turnover GTPase assays. Human RGSZ1 functionally regulated GPA1 (a yeast Galpha(i)-like protein)-mediated yeast pheromone response when expressed in a SST2 (yeast RGS protein) knockout strain. In PC12 cells, transfected RGSZ1 blocked mitogen-activated protein kinase activity induced by UK14304, an alpha(2)-adrenergic receptor agonist. Furthermore, RGSZ1 attenuated D2 dopamine receptor agonist-induced serum response element reporter gene activity in Chinese hamster ovary cells. In summary, these data suggest that RGSZ1 serves as a GTPase accelerating protein for Galpha(i) and regulates Galpha(i)-mediated signaling, thus expanding the potential role of RGSZ1 in G protein-mediated cellular activities.     

Ras complements the carboxyl terminus of v-Abl protein in lymphoid transformation.

Abelson murine leukemia virus (Ab-MLV) mutants expressing v-Abl proteins lacking the carboxyl terminus are compromised in the ability to transform lymphoid but not NIH 3T3 cells. This feature correlates with the presence of low levels of phosphotyrosine in lymphoid cells infected with carboxyl-terminal truncation mutants. In contrast, high levels of phosphotyrosine are observed in NIH 3T3 cells infected with wild-type and mutant Ab-MLV. Two downstream targets affected in lymphoid transformants are the GTPase-activating protein and GTPase-activating protein-associated protein p62, molecules which are heavily tyrosine phosphorylated in lymphoid cells transformed by wild-type Ab-MLV but not carboxyl-terminal truncation mutants of Ab-MLV. This difference suggested that signaling mediated via the Ras pathway may be compromised in lymphoid cells expressing the carboxyl-terminal truncation mutants. Consistent with this idea, expression of v-Ha-ras complemented these mutants in primary bone marrow transformation assays and increased transformation frequencies obtained with the Ab-MLV mutants 8- to 20-fold. These data suggest that a biologically important link exists between the carboxyl terminus of v-Abl protein and the Ras pathway. Signals transmitted via this connection may enhance those mediated via other regions of the v-Abl protein and facilitate transformation of primary, nonimmortalized cells such as pre-B lymphocytes.     

The myristoylated amino terminus of Galpha(i)(1) plays a critical role in the structure and function of Galpha(i)(1) subunits in solution

To determine the role of the myristoylated amino terminus of Galpha in G protein activation, nine individual cysteine mutations along the myristoylated amino terminus of Galpha(i) were expressed in a functionally Cys-less background. Thiol reactive EPR and fluorescent probes were attached to each site as local reporters of mobility and conformational changes upon activation of Galpha(i)GDP by AlF(4)(-), as well as binding to Gbetagamma. EPR and steady state fluorescence anisotropy are consistent with a high degree of immobility for labeled residues in solution all along the amino terminus of myristoylated Galpha(i). This is in contrast to the high mobility of this region in nonmyristoylated Galpha(i) [Medkova, M., et al. (2002) Biochemistry 41, 9962-9972]. Steady state fluorescence measurements revealed pronounced increases in fluorescence upon activation for residues 14-17 and 21 located midway through the 30-amino acid stretch comprising the amino-terminal region. Collectively, the data suggest that myristoylation is an important structural determinant of the amino terminus of Galpha(i) proteins.     

Selection of protein phosphatase 2A regulatory subunits is mediated by the C terminus of the catalytic Subunit

Protein phosphatase 2A (PP2A) is a family of multifunctional serine/threonine phosphatases all composed of a catalytic C, a structural A, and a regulatory B subunit. Assembly of the complex with the appropriate B subunit forms the key to the functional specificity and regulation of PP2A. Emerging evidence suggests a crucial role for methylation and phosphorylation of the PP2A C subunit in this process. In this study, we show that PP2A C subunit methylation was not absolutely required for binding the PR61/B' and PR72/B' subunit families, whereas binding of the PR55/B subunit family was determined by methylation and the nature of the C-terminal amino acid side chain. Moreover mutation of the phosphorylatable Tyr(307) or Thr(304) residues differentially affected binding of distinct B subunit family members. Down-regulation of the PP2A methyltransferase LCMT1 by RNA interference gradually reduced the cellular amount of methylated C subunit and induced a dynamic redistribution of the remaining methylated PP2A(C) between different PP2A trimers consistent with their methylation requirements. Persistent knockdown of LCMT1 eventually resulted in specific degradation of the PR55/B subunit and apoptotic cell death. Together these results establish a crucial foundation for understanding PP2A regulatory subunit selection.     

Pertussis toxin-catalyzed ADP-ribosylation of G(o) alpha with mutations at the carboxyl terminus.

The guanine nucleotide-binding protein G(o alpha) has been implicated in the regulation of Ca2+ channels in neural tissues. Covalent modification of G(o alpha) by pertussis toxin-catalyzed ADP-ribosylation of a cysteine (position 351) four amino acids from the carboxyl terminus decouples G(o alpha) from receptor. To define the structural requirements for ADP-ribosylation, preparations of recombinant G(o alpha) with mutations within the five amino acids at the carboxyl terminus were evaluated for their ability to serve as pertussis toxin substrates. As expected, the mutant in which cysteine 351 was replaced by glycine (C351G) was not a toxin substrate. Other inactive mutants were G352D and L353 delta/Y354 delta. Mutations that had no significant effect on toxin-catalyzed ADP-ribosylation included G350D, G350R, Y354 delta, and L353V/Y354 delta. Less active mutants were L353G/Y354 delta, L353A/Y354 delta, and L353G. ADP-ribosylation of the active mutants, like that of wild-type G(o alpha), was enhanced by the beta gamma subunits of bovine transducin. It appears that three of the four terminal amino acids critically influence pertussis toxin-catalyzed ADP-ribosylation of G(o alpha).     

The carboxyl terminus of the smooth muscle myosin light chain kinase is expressed as an independent protein, telokin.

It has been proposed that the carboxyl terminus of the smooth muscle myosin light chain kinase is expressed as an independent protein. This protein has been purified from tissues and named telokin (Ito, M., Dabrowska, R., Guerriero, V., Jr., and Hartshorne, D. J. (1989) J. Biol. Chem. 264, 13971-13974). In this study we have isolated and characterized cDNA and genomic clones encoding telokin. Analysis of a genomic DNA clone suggests that the mRNA encoding telokin arises from a promoter which appears to be located within an intron of the smooth muscle myosin light chain kinase (MLCK) gene. This intron interrupts exons encoding the calmodulin binding domain of the kinase. The amino acid sequence deduced from the cDNA predicts that telokin is identical to the carboxyl-terminal 155 residues of the smooth muscle MLCK. Unlike the smooth muscle MLCK which is expressed in both smooth and non-muscle tissues, telokin is expressed in some smooth muscle tissues but has not been detected in aortic smooth muscle or in any non-muscle tissues.     

Truncation of the A1 adenosine receptor reveals distinct roles of the membrane-proximal carboxyl terminus in receptor folding and G protein coupling

The carboxyl terminus (C-tail) of G protein-coupled receptors is divergent in length and structure and may represent an individualized cytoplasmic domain. By progressively truncating the A1 adenosine receptor, a Gi/o-coupled receptor with short cytoplasmic stretches, we identify two inherent functions of the C-tail, namely a role in receptor export from the endoplasmic reticulum (ER) and a role in G protein coupling. Deletion of the last 22 and 26 amino acids (of 36) reduced and completely abolished surface expression of the receptor, respectively. The severely truncated receptors were retained in the ER and failed to bind ligands. If overexpressed, even a substantial portion of the full-length receptor was retained in the ER in a form that was not functional. These data indicate that folding is rate limiting in export from the ER and that the proximal segment of the carboxyl terminus provides a docking site for the machinery involved in folding and quality control. In addition, the proximal portion is also important in G protein coupling. This latter role was unmasked when the distal portion of the C-tail (the extreme 18 amino acids, including a palmitoylated cysteine) had been removed; the resulting receptor was functional and transferred the agonist-mediated signal more efficiently than the full-length receptor. Signaling was enhanced because the coupling affinity increased (by 3-fold), which translated into a higher agonist potency. Thus, the distal portion of the carboxyl terminus provides for an autoinhibitory restraint, presumably by folding back and preventing G protein access to the proximal part of the C-tail.     

Phosphorylation of the avian retrovirus integration protein and proteolytic processing of its carboxyl terminus.

The integration protein (IN) of the Prague A strain of Rous sarcoma virus (RSV) was analyzed by high-resolution sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Three polypeptides of similar proportions and molecular mass (32 kDa) were immunoprecipitated by an antiserum directed against the first 10 amino acids of the amino terminus of IN. However, the faster-migrating nonphosphorylated polypeptide was not immunoprecipitated by two different polyclonal antisera directed against the last 11 amino acids of the carboxyl terminus of IN. These results suggest that the faster-migrating species was proteolytically processed at its carboxyl terminus. RSV IN is phosphorylated on an S residue located five amino acids from its carboxyl terminus. Two different missense mutations at this S residue resulted in the isolation of slow-growing viable mutants whose phenotypes were stable. Each mutation at residue 282 eliminated both major phosphorylated-Ser-containing tryptic peptides observed with wild-type IN. An S----F mutation resulted in the conversion of all IN polypeptides to one species that was not precipitable by carboxyl-terminal antisera, suggesting that this amino acid transition promoted proteolysis at the carboxyl terminus. An S----D mutation resulted in the recovery of one major (greater than 95%) slower-migrating polypeptide that was immunoprecipitated by carboxyl-terminal antisera, suggesting that this negatively charged D residue (similar to phosphorylated Ser) inhibited proteolysis. Modification of the S residue at amino acid 262 to R had no apparent effect on the proteolytic processing or phosphorylation of IN.     

Mutational analysis of the carboxyl terminus of the Moloney murine leukemia virus integration protein.

The integration protein (IN) is required for retrovirus DNA integration into the host DNA. The function of the C terminus of the Moloney murine leukemia virus IN protein was examined. The terminal 28 amino acids were found to be nonessential. A linker insertion at position 6025, within a 36-amino-acid insertion not found in any other retrovirus, also produced viable virus.     

Modulation of type-1 protein phosphatase by synthetic peptides corresponding to the carboxyl terminus.

Protein phosphatase type-1 (PP-1) has a protease resistant catalytic core Mr = 35,000 (PP-35K) and carboxyl terminal segment which affects activity with various substrates. We found that micromolar concentration of a synthetic peptide, corresponding to residues 312-326 of the PP-1 carboxyl terminus (P312-326) that is missing from PP-35K, increased the phosphatase activity of PP-35K with phosphorylase and myosin light chains as substrates by decreasing the apparent Km without a change in Vm. Purified PP-1 and PP-35K were inhibited identically by okadaic acid, but peptide P312-326 only stimulated the activity of PP-35K, not full-length PP-1. Other peptides corresponding to the carboxyl terminus of phosphatase-2A or to the amino terminus of PP-1 did not affect the activity of PP-35K. A sequence conserved in PP-1 from different species, Pro-Ile-Thr-Pro-Pro was implicated as the active region because a derivative peptide, Ala-Pro-Ile-Thr-Pro-Pro-Ala, stimulated the activity of PP-35K to the same extent as peptide P312-326 although at higher concentrations. These results indicate that the carboxyl terminus of PP-1 interacts with the catalytic core to modulate its activity, and suggest that the physiological regulation of PP-1 may involve this segment.     

Regulation of heat shock protein 90 ATPase activity by sequences in the carboxyl terminus.

Hsp90, in addition to being an abundant and pivotal cytoplasmic chaperone protein, has been shown to be a weak ATPase. In an effort to characterize the ATPase activity of hsp90, we have observed marked differences in activities among various species of hsp90. Chicken or human hsp90 hydrolyzed ATP with a k(cat) of 0.02 min(-1) and a K(m) greater than 300 microm. In contrast, yeast hsp90 and TRAP1, an hsp90 homologue found in mitochondria, were 10-100-fold more active as ATPases. Sedimentation studies confirmed that all are dimeric proteins. Chicken hsp90 mutants were then analyzed to identify regions within the protein that influence ATPase activity. A truncation mutant of chicken hsp90, N1-573, was found to be monomeric, and yet the catalytic efficiency (k(cat)/K(m)) was greater than 100 times that of the full-length protein (k(cat) of 0.24 min(-1) and K(m) of 60 microm). In contrast, an internal deletion mutant, Delta661-677, was also monomeric but failed to hydrolyze ATP. Finally, deletion of the last 30 amino acids resulted in a dimeric protein with an ATPase activity very similar to full-length hsp90. These data indicate that sequences within the last one-fourth of hsp90 regulate ATP hydrolysis.     

Heterotrimeric G protein-independent signaling of a G protein-coupled receptor. Direct binding of ARNO/cytohesin-2 to the carboxyl terminus of the A2A adenosine receptor is necessary for sustained activation of the ERK/MAP kinase pathway

The A2A adenosine receptor is a prototypical G(s)-coupled receptor, but it also signals, e.g. to mitogen-activated protein (MAP) kinase, via a pathway that is independent of heterotrimeric G proteins. Truncation of the carboxyl terminus affects the strength of the signal through these alternative pathways. In a yeast two-hybrid interaction hunt, we screened a human brain library for proteins that bound to the juxtamembrane portion of the carboxyl terminus of the A2A receptor. This approach identified ARNO/cytohesin-2, a nucleotide exchange factor for the small (monomeric) G proteins of the Arf (ADP-ribosylation factor) family, as a potential interaction partner. We confirmed a direct interaction by mutual pull down (of fusion proteins expressed in bacteria) and by immunoprecipitation of the proteins expressed in mammalian cells. To circumvent the long term toxicity associated with overexpression of ARNO/cytohesin-2, we created stable cell lines that stably expressed the A2A receptor and where ARNO/cytohesin-2 or the dominant negative version E156K-ARNO/cytohesin-2 was inducible by mifepristone. Cyclic AMP accumulation induced by an A2A-specific agonist was neither altered by ARNO/cytohesin-2 nor by the dominant negative version. This was also true for agonist-induced desensitization. In contrast, expression of dominant negative E156K-ARNO/cytohesin-2 and of dominant negative T27N-Arf6 abrogated the sustained phase of MAP kinase stimulation induced by the A2A receptor. We therefore conclude that ARNO/cytohesin-2 is required to support the alternative, heterotrimeric G protein-independent, signaling pathway of A2A receptor, which is stimulation of MAP kinase.     

Baculovirus expression of mammalian G protein alpha subunits.

Complementary DNAs encoding three subtypes of the alpha subunit (alpha i-1, alpha o and alpha s) of rat guanyl nucleotide regulatory proteins were used to construct recombinant baculoviruses which direct high-level expression of the corresponding proteins in cultured Sf9 insect cells. The expressed proteins were recognized by polyclonal antisera specific for the different alpha chains, and co-migrated with the native proteins from rat brain membranes in immunoblotting analyses. Soluble and particulate forms of all three immunoreactive alpha chains were observed following ultracentrifugation of cell lysates. Biosynthetic radiolabelling of infected cells with [35S]methionine or [3H]myristate showed that both soluble and particulate forms of alpha i-1 and alpha o were myristoylated; in contrast, alpha s did not incorporate myristate. The soluble fractions from cells expressing alpha chains showed high levels of GTP-binding activity over that observed in uninfected cells, or in cells infected with wild-type virus. The peak expression levels observed at 72 h post-infection were highest for alpha o at ca. 400 pmol of GTP-gamma-35S/mg protein, or roughly 2% of the total soluble protein. The results of this work show that the baculovirus system can be employed for high-level production of mammalian G protein alpha chains which retain GTP-binding activity and are appropriately modified by myristoylation.     

Quaternary structure of chloroplast F1-ATPase in solution. Conformational changes in spatial arrangement of subunits upon activation

The hydrodynamic properties of isolated ATPases were studied via their rotational diffusion in buffer solution. Chloroplast F1-ATPase (CF1) and Escherichia coli F1-ATPase (EF1) were covalently labeled with eosinisothiocyanate and then investigated by polarized laser spectroscopy. The rotational correlation time in aqueous buffer of latent (five-subunit) CF1 was 390 ns. Four-subunit (delta-deficient) CF1 showed the same correlation time, however, for three-subunit (delta, epsilon-deficient) CF1 the rotational correlation time was more than eight times larger (3200 ns). The rotational correlation time of activated CF1 was three times larger than the one of latent CF1. These large changes in the rotational correlation times are directly related to changes in the quaternary structure of CF1 upon activation. EF1 was found to behave essentially as activated CF1. Based on the observed rotational correlation times we concluded that the mass distributions of latent CF1 and of delta-deficient CF1 resemble a dimeric arrangement. The structure of delta, epsilon-deficient CF1 more likely resembles a hexagon, the mass centers of the six main subunits lie in one plane. The structure of the activated forms of CF1 can be described best as an intermediate between the dimeric arrangement of latent CF1 and an octahedron. The large changes in the quaternary structure of isolated CF1 are reversed when the activation of the enzyme is reversed.