Development of a BRET2 Screening Assay Using β-Arrestin 2 Mutants

This study has focused on enhancing the signal generated from the interaction between a G-protein-coupled receptor (GPCR) and β-arrestin 2 (β-arr2), measured by the bioluminescence resonance energy transfer (BRET2) technology. Both class A (β2-adrenergic receptor [β2-AR]) and class B (neurokinin-type 1 receptor [NK1-R]) GPCRs, classified based on their internalization characteristics, have been analyzed. It was evaluated whether the BRET2 signal can be enhanced by using (1) β-arr2 phosphorylation-independent mutant (β-arr2 R169E) and (2) β-arr2 mutants deficient in their ability to interact with the components of the clathrin-coated vesicles (β-arr2 R393E, R395E and β-arr2 373 stop). For the class B receptor, there was no major difference in the agonist-promoted BRET2 signal when comparing results obtained with wild-type (wt) and mutant β-arr2. However, with the class A receptor, a more than 2-fold increase in the BRET2 signal was observed with β-arr2 mutants lacking the AP-2 or both AP-2 and clathrin binding sites. This set of data suggests that the inability of these β-arr2 mutants to interact with the components of the clathrin-coated vesicle probably prevents their rapid dissociation from the receptor, thus yielding an increased and more stable BRET2 signal. The β-arr2 R393E, R395E mutant also enhanced the signal window with other members of the GPCR family (neuropeptide Y type 2 receptor [NPY2-R] and TG1019 receptor) and was successfully applied in full-plate BRET2-based agonist and antagonist screening assays.

Use of the BRET 7TM receptor/β-arrestin assay in drug discovery and screening

To perform functional cell-based screening assays on seven-transmembrane (7TM) receptors, also known as G-protein coupled receptors, at least three distinct assays are currently needed to screen for G(alphas), G(alphai/0) or G(alphaq/11) signaling receptors. Therefore, there has long been a desire for a universal screening assay that could be used to screen all 7TM receptors independent of their signaling pathway. The receptor/beta-arrestin interaction is common to virtually all 7TM receptors. Therefore, an assay based on this interaction should achieve just that. Bioluminescence resonance energy transfer technology can be used to measure the receptor/beta-arrestin interaction in living cells but due to various technical and biological reasons, the use of the technology for compound screening has been limited. The recent development of beta-arrestin mutants that significantly improve the assay signal, in combination with new improved instrumentation, has transformed bioluminescence resonance energy transfer technology from being a highly specialized research tool in molecular pharmacology to a more drug screening-friendly technique that is useful in an industrial setting.

Beta-arrestin-based Bret2 screening assay for the "non"-beta-arrestin binding CB1 receptor

CB1 receptor (CB1R) antagonists have been demonstrated to be effective in treating obesity and related disorders. This study has been focused on establishing a beta-arrestin 2-based screening assay for the CB1R using BRET2 technology. When the existing BRET2 screening platform was applied to the CB1R, the authors discovered that the receptor interacted weakly with beta-arrestin 2, resulting in unsatisfactory assay performance. To enhance the beta-arrestin binding capacity, they replaced the C-terminal tail of the CB1R with tails from either the V2 or BRS3 receptors, both of which interact strongly with beta-arrestin 2. Using this chimeric approach, the authors screened a small compound library and identified 21 antagonist and inverse agonist hits with IC50 and EC50 values ranging from 0.3 nM to 7.5 microM. Both primary and secondary screening were performed with Z'>0.5, suggesting that the assay is a robust and cost-effective alternative to existing cell-based assays.

Negative cooperativity in H2 relaxin binding to a dimeric relaxin family peptide receptor 1

H2 relaxin, a member of the insulin superfamily, binds to the G-protein-coupled receptor RXFP1 (relaxin family peptide 1), a receptor that belongs to the leucine-rich repeat (LRR)-containing subgroup (LGRs) of class A GPCRs. We recently demonstrated negative cooperativity in INSL3 binding to RXFP2 and showed that this subgroup of GPCRs functions as constitutive dimers. In this work, we investigated whether the binding of H2 relaxin to RXFP1 also shows negative cooperativity, and whether this receptor functions as a dimer using BRET(2). Both binding and dissociation were temperature dependent, and the pH optimum for binding was pH 7.0. Our results showed that RXFP1 is a constitutive dimer with negative cooperativity in ligand binding, that dimerization occurs through the 7TM domain, and that the ectodomain has a stabilizing effect on this interaction. Dimerization and negative cooperativity appear to be general properties of LGRs involved in reproduction as well as other GPCRs.

Cooperative binding of insulin-like Peptide 3 to a dimeric relaxin family peptide receptor 2

Insulin-like peptide 3 (INSL3) binds to a G protein-coupled receptor (GPCR) called relaxin family peptide receptor 2 (RXFP2). RXFP2 belongs to the leucine-rich repeat-containing subgroup (LGR) of class A GPCRs. Negative cooperativity has recently been demonstrated in other members of the LGR subgroup. In this work, the kinetics of INSL3 binding to HEK293 cells stably transfected with RXFP2 (HEK293-RXFP2) have been investigated in detail to study whether negative cooperativity occurs and whether this receptor functions as a dimer. Our results show that negative cooperativity is present and that INSL3-RXFP2 binding shows both similarities and differences with insulin binding to the insulin receptor. A dose-response curve for the negative cooperativity of INSL3 binding had a reverse bell shape reminiscent of that seen for the negative cooperativity of insulin binding to its receptor. This suggests that binding of INSL3 may happen in a trans rather than in a cis way in a receptor dimer. Bioluminescence resonance energy transfer (BRET(2)) experiments confirmed that RXFP2 forms constitutive homodimers. Heterodimerization between RXFP2 and RXFP1 was also observed.

Characterization of G-protein coupled receptor kinase interaction with the neurokinin-1 receptor using bioluminescence resonance energy transfer

To analyze the interaction between the neurokinin-1 (NK-1) receptor and G-protein coupled receptor kinases (GRKs), we performed bioluminescence resonance energy transfer(2) (BRET(2)) measurements between the family A NK-1 receptor and GRK2 and GRK5 as well as their respective kinase-inactive mutants. We observed agonist induced interaction of both GRK5 and GRK2 with the activated NK-1 receptor. In saturation experiments, we observed GRK5 to interact with the activated receptor in a monophasic manner while GRK2 interacted in a biphasic manner with the low affinity phase corresponding to receptor affinity for GRK5. Agonist induced GRK5 interaction with the receptor was dependent on intact kinase-activity, whereas the high affinity phase of GRK2 interaction was independent of kinase activity. We were surprised to find that the BRET(2) saturation experiments indicated that before receptor activation, the full-length NK-1 receptor, but not a functional C-terminal tail-truncated receptor, is preassociated with GRK5 in a relatively low-affinity state. We demonstrate that GRK5 can compete for agonist induced GRK2 interaction with the NK-1 receptor, whereas GRK2 does not compete for receptor interaction with GRK5. We suggest that GRK5 is preassociated with the NK-1 receptor and that GRK5, rather than GRK2, is a key player in competitive regulation of GRK subtype specific interaction with the NK-1 receptor.

Evidence for a role of caveolin-1 in neurokinin-1 receptor plasma-membrane localization, efficient signaling, and interaction with beta-arrestin 2

This study was focused on the relationship between the plasma-membrane localization of neurokinin-1 receptor (NK1-R) and its endocytic and signaling properties. First, we employed electron paramagnetic resonance (EPR) to study the domain structure of HEK-293 cells and NK1-R microlocalization. EPR spectra and the GHOST condensation routine demonstrated that NK1-R was distributed in a well-ordered domain of HEK-293 cells possibly representing lipid raft/caveolae microdomains, whereas the impairment of caveolae changed the NK1-R plasma-membrane distribution. Internalization and second messenger assays combined with bioluminescence resonance energy transfer were employed subsequently to evaluate the functional importance of the NK1-R microlocalization in lipid raft/caveolae microdomains. The internalization pattern was delineated through the use of dominant-negative mutants (DNM) of caveolin-1 S80E (Cav1 S80E), dynamin-1 K44A (Dyn K44A), and beta-arrestin (beta-arr 319-418) and by means of cell lines that expressed various endogenous levels of beta-arrestins. NK1-R displayed rapid internalization that was substantially reduced by DNMs of dynamin-1 and beta-arrestin and even more profoundly in cells lacking both beta-arrestin1 and beta-arrestin2. These internalization data were highly suggestive of the predominant use of the clathrin-mediated pathway by NK1-R, even though NK1-R tended to reside constitutively in lipid raft/caveolae microdomains. Evidence was also obtained that the proper clustering of the receptor in these microdomains was important for effective agonist-induced NK1-R signaling and for its interaction with beta-arrestin2.

Opsin oligomerization in a heterologous cell system

Using bioluminescence resonance energy transfer (BRET) we studied opsin oligomerization in heterologous expression systems and quantitatively assessed its oligomerization state. BRET2 saturation and competition experiments were performed with live COS-7 cells expressing Rluc-and GFP2-tagged receptor constructs. BRET2 saturation curves obtained were hyperbolic, and the calculated oligomerization state (N = 1 for dimers) suggested that opsin (N = 1.34 +/- 0.25) forms higher oligomers. Very high BRET2 values obtained for the opsin homo-dimer pair indicated a large energy transfer efficiency (E) and for cases where E >> 0.1 a modified saturation curve was proposed. The existence of homo-dimer complexes was additionally supported by competition assay results and was also observed in HEK-293 cells. Furthermore, evidence was provided for homo-and hetero-dimerization of family A (beta2-adrenergic) and B (gastric inhibitory polypeptide, GIP) receptors. In summary, these experiments demonstrate homo-and hetero-dimerization for opsin, beta 2-adrenergic, and GIP receptors.

Bioluminescence resonance energy transfer as a screening assay: Focus on partial and inverse agonism

The reported data for compound screening with the bioluminescence resonance energy transfer (BRET2) assay is very limited, and several questions remain unaddressed, such as the behavior of agonists. Eleven beta2 adrenergic receptor (beta2-AR) agonists were tested for full or partial agonism in an improved version of the receptor/beta-arrestin2 BRET2 assay and in 2 cyclic adenosine monophosphate (cAMP) assays (column cAMP assay and ALPHAscreen cAMP assay). Tested in the highly sensitive ALPHAscreen cAMP assay, all selected agonists behaved as full agonists, using isoproterenol as a reference compound. In the less sensitive column cAMP assay, ephedrine and dopamine had a clear partial response. For the BRET2 assay, a highly graded picture was obtained. Moreover, beta2-AR antagonists were tested for inverse agonism. Pronounced inverse agonism was detected in the ALPHAscreen cAMP assay. Only marginal inverse agonistic responses were seen for alprenolol and pindolol in the column cAMP assay, and no inverse agonism was seen in the BRET2 assay. For the beta2-AR, the BRET2 assay may be superior for secondary screening of agonists where a separation of full and partial agonists is needed and the ALPHAscreen cAMP assay may be preferred for primary screening of agonists where all receptor activating compounds are desired.

Dimerization of ADAR2 is mediated by the double-stranded RNA binding domain

Members of the family of adenosine deaminases acting on RNA (ADARs) can catalyze the hydrolytic deamination of adenosine to inosine and thereby change the sequence of specific mRNAs with highly double-stranded structures. The ADARs all contain one or more repeats of the double-stranded RNA binding motif (DRBM). By both in vitro and in vivo assays, we show that the DRBMs of rat ADAR2 are necessary and sufficient for dimerization of the enzyme. Bioluminescence resonance energy transfer (BRET) demonstrates that ADAR2 also exists as dimers in living mammalian cells and that mutation of DRBM1 lowers the dimerization affinity while mutation of DRBM2 does not. Nonetheless, the editing efficiency of the GluR2 Q/R site depends on a functional DRBM2. The ADAR2 DRBMs thus serve differential roles in RNA dimerization and GluR2 Q/R editing, and we propose a model for RNA editing that incorporates the new findings.

Nonpeptide and Peptide Growth Hormone Secretagogues Act Both as Ghrelin Receptor Agonist and as Positive or Negative Allosteric Modulators of Ghrelin Signaling

Two nonpeptide (L692,429 and MK-677) and two peptide [GH-releasing peptide (GHRP)-6 and ghrelin] agonists were compared in binding and in signal transduction assays: calcium mobilization, inositol phosphate turnover, cAMP-responsive element (CRE), and serum-responsive element (SRE) controlled transcription, as well as arrestin mobilization. MK-677 acted as a simple agonist having an affinity of 6.5 nm and activated all signal transduction systems with similar high potency (0.2–1.4 nm). L-692,429 also displayed a very similar potency in all signaling assays (25–60 nm) but competed with a 1000-fold lower apparent affinity for ghrelin binding and surprisingly acted as a positive allosteric receptor modulator by increasing ghrelin’s potency 4- to 10-fold. In contrast, the potency of GHRP-6 varied 600-fold (0.1–61 nm) depending on the signal transduction assay, and it acted as a negative allosteric modulator of ghrelin signaling. Unexpectedly, the maximal signaling efficacy for ghrelin was increased above what was observed with the hormone itself during coadministration with the nonendogenous agonists. It is concluded that agonists for the ghrelin receptor vary both in respect of their intrinsic agonist properties and in their ability to modulate ghrelin signaling. A receptor model is presented wherein ghrelin normally only activates one receptor subunit in a dimer and where the smaller nonendogenous agonists bind in the other subunit to act both as coagonists and as either neutral (MK-677), positive (L-692,429), or negative (GHRP-6) modulators of ghrelin function. It is suggested that an optimal drug candidate could be an agonist that also is a positive modulator of ghrelin signaling.

Novel method for the study of receptor Ca2+ signalling exemplified by the NK1 receptor

We have used a novel technology (NovoStar from BMG Labtechnologies) for the study of the Ca2+ signalling of the human tackykinin NK1 (hNK-I receptor). The NovoStar is a microplate reader based on fluorescence and luminescence. The instrument implements a robotic pipettor arm and two microplate carriers, typically one for samples and one for cells. The robotic pipettor arm can transfer sample (agonist or antagonist) from the sample plate or other liquid containers to the cell plate, facilitating the study of Ca2+ signalling to such a degree that the instrument can be used for Medium Throughput Screening (MTS). Using the NovoStar we have found the molecular pharmacology of the NK1 receptor to be comparable to that observed in classical signal transduction assays. Thus, we have observed an EC50 value of 3 nM for substance P induced Ca2+ response. This value corresponds well with previously published values for substance P induced IP and cAMP turnover. [1] Using the NovoStar technology we have studied the pharmacological profile of the well known non-peptide NKI receptor antagonists CP96,345 and SR140,333 [2,3] in respect of inhibition of the Ca2+ response induced by substance P. Interestingly, the antagonistic potency of the antagonists depended greatly on the experimental design, e.g., a dependency of timing in the addition of antagonists vs. agonist was noted. Also, metal-ion site engineered NK1 receptors [2] were tested for the ability of metal-ions to inhibit signalling. It is concluded that the NovoStar is a reliable tool for the study of receptor Ca2+ signalling, both as a research tool and as a MTS system.

Internalization kinetics of the gonadotropin-releasing hormone (GnRH) receptor

This study quantified the agonist-induced endocytotic and recycling events of the mammalian gonadotropin releasing hormone receptor (GnRH-R) and investigated the role of the intracellular carboxyl (C)-terminal tail in regulating agonist-induced receptor internalization kinetics. The rate of internalization for the rat GnRH-R was found to be exceptionally low when compared with G-protein coupled receptors (GPCRs) which possess a cytoplasmic C-terminal tail (thyrotropin-releasing hormone receptor (TRH-R), catfish GnRH-R (cfGnRH-R) and GnRH/TRH-R chimeric receptor). These data provide evidence that the presence of a functional intracellular cytoplasmic C-terminal tail is essential for rapid internalization of the studied GPCRs.

The rat gonadotropin-releasing hormone receptor internalizes via a beta-arrestin-independent, but dynamin-dependent, pathway: addition of a carboxyl-terminal tail confers beta-arrestin dependency

This study examined the mechanism underlying the rat GnRH receptor (GnRH-R) internalization pathway by investigating the role of added/extended C-terminal tails and the effect of beta-arrestins and dynamin. The internalization of the wild-type (WT) rat GnRH-R, stop codon mutants, GnRH-R/TRH receptor (TRH-R) chimera, rat TRH-R, and catfish GnRH-R was examined using radioligand binding assay. Overexpression of beta-arrestin in COS-7 cells expressing each of the receptor constructs substantially increased endocytosis rate constants (k(e)) of the TRH-R, catfish GnRH-R, and GnRH-R/TRH-R chimera, but not of the WT rat GnRH-R and stop codon mutants. The beta-arrestin-promoted increase in the k(e) value was diminished by cotransfecting cells with the dominant negative beta-arrestin-(319-418) mutant, whereas WT GnRH-R and stop codon mutant internalization were unaffected. Additionally, confocal microscopy showed that activated GnRH-Rs failed to induce time-dependent redistribution of either beta-arrestin-1- or beta-arrestin-2-green fluorescent protein conjugate to the plasma membrane. However, the dominant negative dynamin (DynK44A) mutant impaired internalization of all of the receptors regardless of their beta-arrestin dependency, indicating that they internalize via a clathrin-mediated pathway. We conclude that the mammalian GnRH-R uses a beta-arrestin-independent, dynamin-dependent internalization mechanism distinct from that employed by the other receptors studied.

Pivotal role for the cytoplasmic carboxyl-terminal tail of a nonmammalian gonadotropin-releasing hormone receptor in cell surface expression, ligand binding, and receptor phosphorylation and internalization

The gonadotropin-releasing hormone receptor (GnRH-R) of the African catfish couples to phospholipase C and belongs to the large family of G protein-coupled receptors. We recently demonstrated that removal of the carboxyl-terminal tail (S331-Q379) from the catfish GnRH-R results in a loss of agonist binding; the current study sought to define more precisely the role of this region in receptor function. Progressive truncations of the carboxyl-terminal tail decreased cell surface expression detected by either enzyme-linked immunosorbent assay or agonist-binding. The two most truncated receptors (stop331 and stop337) showed no binding but were detected at the cell surface by enzyme-linked immunosorbent assay. All receptors able to bind agonist were also able to activate phospholipase C. The catfish GnRH-R was phosphorylated after agonist-occupation and use of truncated mutants showed this phosphorylation to be within the carboxyl-terminal tail. Furthermore, studies with S356A, S363A and SS356,363AA mutant receptors demonstrated that Ser363 is a major site of agonist-induced phosphorylation. The absence of this phospho-acceptor site markedly impaired agonist-mediated receptor internalization. In addition, both, Ser363 and the last 12 residues of the tail (not containing Ser363) were shown to be important for beta-arrestin-dependent internalization. These observations are relevant to the regulatory function of the carboxyl-terminal tail of G protein-coupled receptors in general and are particularly intriguing given the absence of this region in mammalian GnRH-Rs.

Lack of a C-terminal tail in the mammalian gonadotropin-releasing hormone receptor confers resistance to agonist-dependent phosphorylation and rapid desensitization

The mammalian gonadotropin-releasing hormone receptor (GnRH-R) is, at present, the only G-protein-coupled receptor that activates phospholipase C and lacks a C-terminal tail. We have previously demonstrated that this unique structural feature is associated with resistance to rapid desensitization of phosphoinositide signaling in COS-7 and HEK-293 cells (Heding, A., Vrecl, M., Bogerd, J., McGregor, A., Sellar, R., Taylor, P. L., and Eidne, K. A. (1998) J. Biol. Chem. 273, 11472-11477). Using receptors tagged with a nonapeptide of the influenza hemagglutinin protein to enable immunoprecipitation, we now demonstrate that the mammalian GnRH-R is not phosphorylated in an agonist-dependent manner. In contrast, the mammalian thyrotropin-releasing hormone receptor and the African catfish GnRH-R, both of which have a C-terminal tail, are phosphorylated in response to agonist challenge. Furthermore, chimeras of the mammalian GnRH-R with the C-terminal tail of either the mammalian thyrotropin-releasing hormone receptor or the catfish GnRH-R are also phosphorylated in an agonist-dependent manner. Only those receptors having C-terminal tails showed desensitization of phosphoinositide responses within 5-10 min of agonist challenge. We also show that the internalization of all these receptors when expressed transiently in COS-7 cells is similar. This dissociates receptor internalization from rapid desensitization and demonstrates that the lack of a C-terminal tail in the mammalian GnRH-R results in an inability of the receptor to undergo agonist-dependent phosphorylation and that this results directly in a resistance to rapid desensitization.

Gonadotropin-releasing hormone receptors with intracellular carboxyl-terminal tails undergo acute desensitization of total inositol phosphate production and exhibit accelerated internalization kinetics

The mammalian gonadotropin-releasing hormone receptor (GnRH-R) is the only G-protein-coupled receptor (GPCR) in which the intracellular C-terminal tail is completely absent. In contrast to other GPCRs, the GnRH-R does not show rapid desensitization of total inositol (IP) production, and the rates of internalization are exceptionally slow. We investigated whether the incorporation of a cytoplasmic tail into the C terminus of the GnRH-R affects desensitization events and receptor internalization rates. A GnRH-R/TRH-R chimera was created where the intracellular tail of the rat thyrotropin-releasing hormone receptor (TRH-R) was engineered into the C terminus of the rat GnRH-R. Three different rat GnRH-R cDNA stop codon mutations (one for each reading frame) were also made. The GnRH-stimulated IP production of the wild-type rat GnRH-R expressed in either COS-7 or HEK 293 cells did not desensitize even after prolonged stimulation with GnRH. In contrast, the catfish GnRH-R (which does possess an intracellular tail) and the TRH-R rapidly (<10 min) desensitized following agonist stimulation. The GnRH-R/TRH-R chimera also desensitized following treatment with GnRH, resembling the pattern shown by the TRH-R and the catfish GnRH-R. Two of the stop codon mutants did not show desensitization of IP production, and the third mutant with the longest tail was not functional. Internalization experiments showed that the rat GnRH-R had the slowest endocytosis and recycling rates compared with the TRH-R, the catfish GnRH-R, and the chimeric GnRH/TRH-R. This study demonstrates that the addition of a functional intracellular C-terminal tail to the GnRH-R produces rapid desensitization of IP production and significantly increases internalization rates.

Fate of insulin analogs in intact and nephrectomized rats determined by their receptor binding constants

After intravenous injection of 125I-labeled human insulin and analogs in normal and nephrectomized rats, we examined their kinetic fate by Q-Sepharose separation into intact ligand, "fragments" (genuine fragments and protein-bound radioactivity), and iodide. Receptor binding association and dissociation constants (kass and kdis, respectively) of the analogs were estimated dynamically in vitro by BIAcore. The very fast disappearance of intact ligand from serum was found to be determined by 1) both kass and kdis of receptor-bearing tissue, thus substantiating our primary hypothesis; 2) elimination by kidneys, and 3) fast extravascularization. The rate of appearance of degradation products from receptor-mediated intracellular processing seems determined by kdis. With the possible exception of a truncated analog, ligand appears protected against degradation while the intracellular receptor-ligand complex remains intact. Non-receptor-mediated processing in kidneys is slow, compared with the receptor-mediated uptake and degradation of ligands with rate constants comparable to those of insulin. We observed binding of insulin and analogs putatively to serum proteins; binding capacity and affinity appeared insignificant for insulin but considerable for some analogs.

Biosensor measurement of the binding of insulin-like growth factors I and II and their analogues to the insulin-like growth factor-binding protein-3

Most insulin-like growth factor (IGF) molecules in the circulation are found in a 150-kDa complex containing IGF-binding protein-3 (IGFBP-3) and an acid-labile subunit, which does not itself bind IGF. Affinities (Kd values) between 0.03 and 0.5 nM have been reported for IGF-I/IGFBP-3 binding, but no kinetic data are available. In this study we measured the high affinity binding of unlabeled IGFs and IGF analogues to recombinant unglycosylated IGFBP-3, using a BIAcoretrade mark instrument (Pharmacia Biosensor AB). IGF-I binding showed fast association and slow non-first-order dissociation kinetics, and an equilibrium Kd of 0.23 nM. IGF-II had similar kinetics with slightly higher affinity. Analogues with mutations in the first 3 amino acids of the B-region (des(1-3) IGF-I and long IGF-I) showed 25 and 50 times lower affinity than IGF-I. Replacement of residues 28-37 by Gly-Gly-Gly-Gly or deletion of residues 29-41 in the C-region had little effect on the kinetic parameters, contrasting with the markedly impaired binding of these analogues to the IGF-I receptor. Swapping of the disulfide bridges in IGF-I and the C-region mutants decreased the affinity dramatically for IGFBP-3, primarily by decreasing the association rate. Insulin had approximately 1000 times lower affinity than IGF-I.

Amino acid residues 721-729 are required for full factor VIII activity

Recombinant two-chain factor VIII, from which the B domain had been deleted, was expressed in Chinese hamster ovary cells. In addition to the major product, three minor factor VIII forms were isolated. The A2 domains generated by thrombin cleavage showed different electrophoretic mobilities. Peptide mapping of the A2 domains showed that two of the factor VIII forms had the expected C-terminus of the heavy chain at Arg740 [FVIII-(1-740)] and that the other factor VIII forms had C-termini at Tyr729 [FVIII-(1-729)] or Glu720 [FVIII-(1-720)]. The major FVIII-(1-740) form, FVIII-(1-729), and FVIII-(1-720) contained sulfated tyrosine residues at Tyr718, Tyr719 and Tyr723. The minor FVIII-(1-740) form was shown to lack sulfation at these positions. The specific clotting activity was approximately 1 x 10(4) U/mg for FVIII-(1-740) (both forms) and FVIII-(1-729), but twofold lower for FVIII-(1-720). A time study of thrombin activation showed that FVIII-(1-720) was activated slower than FVIII-(1-740), FVIII-(1-729) and plasma-derived factor VIII. Partially sulfated FVIII-(1-740) was activated at the same rate as the fully sulfated FVIII-(1-740). The equilibrium dissociation constant for binding of factor VIII to inactivated immobilized thrombin was the same for all factor VIII forms, showing that the slower activation of FVIII-(1-720) was not due to a lower affinity for the anion-binding exosite in thrombin.