p38 JAB1 Binds to the Intracellular Precursor of the Lutropin/Choriogonadotropin Receptor and Promotes Its Degradation*

Using the C-terminal tail of the rat lutropin/choriogo-nadotropin receptor (rLHR) as “bait” in a yeast two-hybrid screen resulted in the identification of p38 JAB1 (a protein initially identified as a co-activator of c-Jun) as a putative rLHR binding partner. More recently p38 JAB1 has been shown to promote the degradation of a cyclin-dependent kinase inhibitor and to be a component of the COP9 signalosome. Microscopic localization of an epitope-tagged p38 JAB1 expressed in 293 cells revealed a punctuated perinuclear and cytosolic localization, while cell fractionation studies showed that most of the p38 JAB1 was in a high speed supernatant. Co-transfec-tion of 293 cells revealed that p38 JAB1 binds to the immature 68-kDa precursor of the rLHR that resides in the endoplasmic reticulum and promotes its degradation. It does not appear to interact with the cell surface rLHR, however, and it does not affect its expression. When transfected into HeLa cells, p38 JAB1 potentiates the transcriptional activity of c-Jun, but co-transfection with rLHR prevents this effect. We conclude that p38 JAB1 interacts with the rLHR precursor and promotes its degradation. These results reveal a novel protein binding partner of the rLHR and are consistent with current views of the functions of p38 JAB1 . such as

Heptahelical (or G protein-coupled) receptors (GPCRs) 1 interact with several classes of intracellular proteins such as heterotrimeric G proteins, G protein-coupled receptor kinases, second messenger-dependent protein kinases, and arrestins (reviewed in Refs. [1][2][3][4]. While the functional consequences of these interactions are well characterized, recent studies have raised the possibility that some of the physiological actions of GPCRs are not mediated by the interaction of these receptors with the proteins listed above (reviewed in Ref. 2). Moreover, the use of modern assays that detect protein-protein interactions have resulted in the identification of a growing number of additional intracellular proteins that bind to GPCRs directly. These include the association of the Homer family of proteins with several subtypes of the metabotropic glutamate receptor (5), the ␣ subunit of eukaryotic initiation factor 2B with some adrenergic receptors (6,7), InaD and dynein with rhodopsin (8,9), the Na ϩ /H ϩ exchanger regulatory factor (NHERF) with the ␤ 2 -adrenergic and P2Y1 purinergic receptors (10,11), the isoform of 14 -3-3 proteins with several subtypes of the ␣ 2adrenergic receptor (7), ATRAP with the AT 1a angiotensin II receptor (12), Gc1g-R with the ␣ 1B -adrenergic receptor (13), and spinophilin with the D2 dopamine receptor (14). While the functional consequences of these interactions have not always been clarified, it seems that the formation of such GPCRprotein complexes may result in the regulation of the activity of the associated protein, the activity of the associated GPCR, and/or the targeting and expression of GPCRs. For example, the association of the ␤ 2 -adrenergic receptor with NHERF stimulates the Na ϩ /H ϩ exchanger (10), while the association of ATRAP with the AT 1a angiotensin II receptor leads to an inhibition of AT 1a -mediated activation of phospholipase C. InaD participates in the formation of large signaling complexes containing rhodopsin and other signaling molecules such as phospholipase C and protein kinase C (8), Homer can link metabotropic glutamate receptors with inositol 1,4,5-trisphosphate receptors, thus modulating Ca ϩ2 release (15), while spinophilin may be able to link D2 dopamine receptors to the actin cytoskeleton and to protein phosphatase-1 (14). Finally, the association of gC1q-R with the ␣ 1B -adrenergic receptor leads to a decrease in the expression of the receptor and its targeting to the cell surface (13).
Since truncations or mutations of the C-terminal tail of the lutropin/choriogonadotropin receptor (LHR) have effects on the targeting of the nascent receptor to the plasma membrane, and on the agonist-induced cAMP accumulation, uncoupling, internalization and down-regulation of the cell surface LHR (16 -21), we used this region to search for LHR protein binding partners. The results presented here document the interaction of the intracellular precursor of the rat (r) LHR with p38 JAB1 . This protein was initially identified as a co-activator of c-Junmediated transcription (22). It was later shown to promote the degradation of p27 Kip1 , a cyclin-dependent-kinase inhibitor (23), and to be a component of a novel protein complex that was initially called JAB1-containing signalosome but later renamed COP9 signalosome (24 -26).

MATERIALS AND METHODS
Yeast Two-hybrid Screen-The Matchmaker two-hybrid system 2 (CLONTECH Laboratories, Palo Alto, CA) was used according to the protocols provided by the manufacturer. Using polymerase chain reaction-based strategies, we subcloned the C-terminal 42 residues of the rLHR (i.e. residues 632-674) into the pAS2-1 vector to generate a fusion protein with the GAL4 DNA binding domain. This plasmid was used as "bait" to screen a human kidney 293 cells cDNA library constructed in the pACT2 vector to generate fusion products with the GAL4 activation domain. This library was also purchased from CLONTECH.
Expression and Subcellular Localization of HA-p38 JAB1 -The full-length p38 identified from the two-hybrid screen was modified (by polymerase chain reaction) to have a consensus Kozak sequence immediately upstream of the initiator methionine and to encode for the HA epitope (YPYDVPDYA) at the C terminus. The modified construct was subcloned into pCDNA3.1 for expression in mammalian cells. Human kidney 293T cells were plated in 60-or 100-mm plates and transfected (when ϳ80% confluent) by overnight exposure to plasmid DNA co-precipitated with calcium phosphate as described by Chen and Okayama (27). Cell lysates were prepared 24 h after transfection as described elsewhere (28). The lysates were diluted 2-fold with concentrated SDS-PAGE sample buffer, boiled for 1 min, resolved on SDS gels, and electrophoretically transferred to polyvinylidene difluoride membranes as described elsewhere (29). After blocking (29) the blots were incubated overnight with 2-5 g/ml mouse monoclonal antibody (12CA5) to the HA epitope. A horseradish peroxidase-labeled goat antimouse secondary antibody was then used during a 1-h incubation at a final dilution of 1:5,000 and the blots were finally developed using the enhanced chemiluminescence (ECL) system of detection from Amersham Pharmacia Biotech.
For the subcellular fractionation experiments, the transiently transfected 293T cells (see above) were washed in an ice-cold solution containing 0.25 M sucrose, 10 mM Hepes, 1 mM EDTA, pH 7.4, recovered by centrifugation and homogenized using 10 strokes of a motor-driven Teflon pestle at a speed of about 5,000 rpm. The homogenates were centrifuged at 800 ϫ g for 10 min, and the supernatant was saved. The pellets were re-homogenized and centrifuged again to give a final low speed pellet (called P1), and the supernatants were combined and centrifuged at 100,000 ϫ g for 60 min to give a high speed pellet (P2) and a high speed supernatant (S2). The pellets were resuspended (by hand homogenization) in the buffer described above to give a suspension with a final volume identical to that of S2. Equal aliquots of each fraction were then diluted 2-fold with concentrated SDS-PAGE sample buffer, boiled for 1 min, resolved on SDS gels, and processed as described above.
For the microscopy studies, the transfected 293T cells were trypsinized at the end of the overnight transfection and replated in polylysine-covered chambers. Twenty-four hours later, they were processed for microscopy using standard methods (30) and the transfected HA-p38 was visualized by staining the cells with the 12CA5 monoclonal antibody (at a final concentration of 2 g/ml), followed by a rhodaminelabeled goat anti-mouse second antibody. The cells were ultimately examined using a Bio-Rad MRC laser scanning confocal microscope at the Central Microscopy Research Facility of the University of Iowa.
In Vitro Binding of HA-p38 JAB1 to GST Fusion Products of the rLHR-Plasmids encoding for GST fusion proteins of the first (residues 365-373), second (residues 441-460), and third intracellular (residues 526 -548) loop as well as the extreme C-terminal tail (residues 632-674) of the rLHR were generated by subcloning the indicated regions of the rLHR into the pGEX-5X-1 vector (Amersham Pharmacia Biotech). GST fusion proteins were prepared according to the manufacturer's instructions.
The HA-tagged p38 JAB1 subcloned in pcDNA3.1 (see above) was linearized at the 3Ј end and used to prepare in vitro translated, [ 35 S]methionine-labeled HA-p38 JAB1 with an in vitro transcription-translation kit purchased from Promega (Madison, WI). For binding assays, 20 g of the different GST fusion proteins were incubated overnight at 4°C with the [ 35 S]methionine-labeled HA-p38 JAB1 in a buffer containing 20 mM Hepes, 100 mM KCl, 5 mM MgCl 2 , 0.05% Triton X-100, 10 mg/ml ovalbumin, pH 7.4 (6). At the end of this incubation, each reaction received a 20 l aliquot of a 50% suspension of glutathione-agarose (Amersham Pharmacia Biotech) in the same buffer and the incubation was continued for 2 h at room temperature. The agarose beads were recovered by centrifugation, and they were washed two times with binding buffer and two times with 10 mM sodium phosphate, 150 mM NaCl, pH 7.4, prior to elution with SDS-PAGE sample buffer. The eluted material was resolved on SDS gels, and the [ 35 S]methioninelabeled HA-p38 JAB1 was detected with a PhosphorImager (Molecular Dynamics).
Co-immunoprecipitations-Human kidney 293T cells were plated in 100-mm dishes and transfected (see above) with a constant amount of DNA (10 g) containing different combination of three different plasmids, HA-p38 JAB1 (described above), an expression vector coding for the myc-tagged rLHR in pcDNA1/Neo (described in Ref. 31), and an empty expression vector (either pcDNA 3.1 or pcDNA1/Neo). Twenty-four hours after transfection the cells were placed on ice and washed twice with ice-cold 10 mM sodium phosphate, 150 mM NaCl, pH 7.4. The cells were then lysed during a 15-min incubation with ϳ1 ml of a cold buffer containing 1% Triton X-100, 60 mM octylglucopyranoside, 100 mM KCl, 5 mM MgCl 2 , 1 mM CaCl 2 , 20 mM Hepes, pH 7.4, and protease inhibitors (Complete protease inhibitor mixture from Roche Molecular Biochemicals). The lysates were clarified by centrifugation and 1 ml aliquots (each containing the same amount of lysate protein) were rotated (overnight at 4°C) with 50 l of a 50-fold dilution of a rabbit polyclonal antibody to the HA epitope (polyclonal antibody HA.11 from BabCO) that had been prebound to 100 l of 50% slurry of protein A. The immunoprecipitates were collected by centrifugation, and the protein A beads were washed four times with 1-ml aliquots of cold lysis buffer and once with a 10-fold dilution of the cold lysis buffer. After the last wash, the beads were mixed with SDS-PAGE sample buffer and boiled for 1 min. The eluted samples were resolved by SDS-PAGE, and electrophoretic blots were prepared as described above. These blots were incubated overnight with a 1,000-fold dilution of a monoclonal antibody to the myc epitope (9E10) and ultimately developed by incubation with a horseradish peroxidase-labeled goat anti-mouse secondary antibody and the ECL detection system as described above.
Transactivation Assays-HeLa cells (provided to us by Dr. Dawn Quelle, University of Iowa, Iowa City, IA) were plated on 35-mm wells at a density of 2 ϫ 10 5 /well and were transfected (using the calcium phosphate coprecipitation method described above) 24 h after plating. Each well was transfected with a constant amount of DNA (4.3 g) containing a constant amount (0.1 g) of the reporter plasmid (designated AP1-luc) and different amounts of four other plasmids. The AP1-luc plasmid (Stratagene, La Jolla, CA) is a reporter plasmid encoding for the firefly luciferase gene driven by several copies of an AP-1 enhancer element. The other plasmids used were an empty vector (pcDNA 3.1 or pcDNA1/Neo), pRSV c-Jun (32) (which was provided to us by Dr. Arthur Gutierrez-Hartman of the University of Colorado Health Sciences Center), an expression vector coding for the myc-tagged rLHR in pcDNA1/Neo (described in Ref. 31), and the expression vector for HA-p38 JAB1 described above. Twenty-four hours after transfection, the cells were lysed and assayed for luciferase activity using a luciferase assay kit (Promega).
Other Methods-The methods used to measure the binding of 125 I-hCG to intact cells and their detergent extracts, the incorporation of [ 35 S]cysteine/methionine into the 68-kDa precursor of the rLHR, and the recovery of 125 I-hCG binding following proteolysis of the mature cell surface rLHR have all been described (31,33). The only modification introduced here was that the incorporation of [ 35 S]cysteine/methionine into the 68-kDa precursor of the rLHR was measured using a Phosphor-Imager rather than by fluorography/densitometry.
The rate of synthesis (S) and the rate constant for degradation (k deg ) of the 68-kDa precursor were calculated by following the incorporation of [ 35 S]cysteine/methionine into the 68-kDa precursor of the rLHR as a function of time after addition the isotope. S and k deg were calculated from a nonlinear fit of these data (cf. Fig. 5) to Equation 1 (34).
[rLHR68] t ϭ amount of 35 S incorporated into the 68-kDa rLHR precursor at time ϭ t; S ϭ rate of synthesis of the 68-kDa rLHR; and k deg ϭ rate constant for degradation of the 68-kDa rLHR. If one assumes that the density of cell surface rLHR is dictated by the rate of externalization (E) and the rate constant for internalization (k int ), these parameters can be calculated from the time course of recovery of 125 I-hCG binding to the surface of proteolyzed cells as a function of time after removal of the protease (31,33). E and k int were thus calculated from a nonlinear fit of these data (cf. Fig. 6) to Equation 2 (33)(34)(35)(36).
[Binding] t ϭ 125 I-hCG bound to cells at time ϭ t; [Binding] 0 ϭ 125 I-hCG bound to cells at time ϭ 0; E ϭ rate of externalization of the rLHR; and k int ϭ rate constant for internalization of the rLHR. Nonlinear regressions were performed using the Delta Graph® software package, and they all had a regression coefficient of at least 0.95.
Statistical analysis (t test with two-sided p values) was performed using Instat (Graph Pad Software).
Hormones, Cell Lines, and Other Supplies-Human kidney 293T cells are a derivative of 293 cells that express the SV40T antigen (37) and were provided to us by Dr. Marlene Hosey (Northwestern University, Chicago, IL). HeLa cells were obtained from Dr. Dawn Quelle. The 9E10 hybridoma cell line was obtained from the American Type Culture Collection. Purified hCG (CR-127, ϳ15,000 IU/mg) was kindly provided by the National Hormone and Pituitary Agency (NIDDK, National Institutes of Health, Bethesda, MD). 125 I-hCG was prepared using the purified hCG as described elsewhere (38). Partially purified hCG (ϳ3,000 IU/mg) was purchased from Sigma, and it was used only for the determination of nonspecific binding (see above). Cell culture medium was prepared by the Media and Cell Production Core, respectively, of the Diabetes and Endocrinology Research Center of the University of Iowa. Other cell culture supplies and reagents were obtained from Corning and Life Technologies, Inc., respectively. All other chemicals were obtained from commonly used suppliers.

RESULTS
Initial Characterization of the Interaction between HA-p38 JAB1 and the rLHR-A yeast two-hybrid screen was used to identify candidate proteins that interact with the extreme Cterminal tail (residues 632-674) of the rLHR. Screening of about 10 7 transformants from a human kidney 293 cell library resulted in the isolation of about 30 clones that interacted with this portion of the rLHR. Five of these clones had a common insert with sequences identical to that of a human Jun activation domain-binding protein called JAB1 (GenBank accession no. U65928; Ref. 22). One of these clones had a ϳ1.3-kilobase pair insert that contained the entire open reading frame for JAB1.
The insert of the chosen clone was excised, and it was modified at the 5Ј end to encode for a consensus Kozak sequence (the sequence immediately upstream of the initiator ATG was changed from TCGGCC to GCCACC) and at the 3Ј end and to encode for the HA epitope (YPYDVPDYA) immediately upstream of the stop codon. The modified cDNA was then subcloned into pcDNA3.1 for further studies. A protein (henceforth called HA-p38 JAB1 ) of the appropriate size (ϳ38 kDa) could be readily detected in Western blots of human kidney 293 cells transfected with the expression vector (Fig. 1A). Confocal microscopy of transfected cells revealed that HA-p38 JAB1 is distributed perinuclearly and throughout the cytosol in what ap-pears to be small clusters (Fig. 1B). This cytosolic localization was confirmed by subcellular fractionation experiments, which showed the presence of immunoreactive HA-p38 JAB1 almost exclusively in the high speed supernatant of homogenates of transfected cells (Fig. 1C).
An interaction between HA-p38 JAB1 and the rLHR could be demonstrated in vitro and in vivo. Fig. 2 shows that 35 S-labeled HA-p38 JAB1 prepared by in vitro transcription/translation can bind to GST fusion proteins of the C-terminal tail (residues 632-674) or the third intracellular loop (residues 526 -548) of the rLHR. It does not bind to GST fusion proteins of the first (residues 441-460) or the second (residues 526 -548) intracellular loops of the rLHR or to GST, however. Thus, the bait used in the yeast two-hybrid screen represents only one of the two regions of the rLHR that interact with HA-p38 JAB1 .
More importantly, co-immunoprecipitation experiments done on lysates of 293 cells co-transfected with the myc-tagged rLHR and HA-p38 JAB1 also revealed an association of these two proteins in vivo (Fig. 3). There are two species of rLHR present in transfected cells, a ϳ68-kDa intracellular precursor that is localized in the endoplasmic reticulum and a ϳ85-kDa mature receptor that is localized at the plasma membrane (31,39). While both of these species are readily detectable in stably transfected cells (31,39), immunoblots of cell lysates prepared from transiently transfected cells reveal only the 68-kDa precursor as a discrete band (identified by an arrow on the left panel of Fig. 3). In transiently transfected cells the mature receptor is often identified as a smear or an aggregate at the top of the gels (Fig. 3, left panel). Anti HA-immunoprecipitates of lysates of cells co-transfected with HA-p38 JAB1 and myc-rLHR followed by blotting with an antibody to the myc epitope Human kidney 293T cells were transfected with an empty expression vector or HA-p38 JAB1 as indicated and used 24 h after transfection. Panel A, lysates of transfected cells were resolved on SDS-polyacrylamide gels; the gels were electrophoretically blotted, probed with the 12CA5 antibody, and ultimately visualized using the ECL system (see "Materials and Methods"). The numbers on the right of the blot indicate the position of molecular weight standards. Panel B, the subcellular localization of HA-p38 JAB1 was ascertained by laser confocal microscopy of fixed and permeabilized 293T cells transiently transfected with HA-p38 JAB1 . The expressed HA-p38 JAB1 was visualized using the 12CA5 monoclonal antibody and a fluorescein-labeled secondary antibody as described under "Materials and Methods." No signal was detected in cells transfected with an empty vector (data not shown). Panel C, the subcellular localization of HA-p38 JAB1 was ascertained by subcellular fractionation of 293T transiently transfected with HA-p38 JAB1 . Cell homogenates were separated into three fractions: a 800 ϫ g for 10 min pellet (P1), a 100,000 ϫ g for 60 min pellet (P2), and a 100,000 g for 60 min supernatant (S), as described under "Materials and Methods." The volumes of all the fractions were equalized, and duplicate aliquots of equal volume were resolved on SDS-polyacrylamide gels. The gels were electrophoretically blotted, probed with the 12CA5 antibody, and ultimately visualized using the ECL system (see "Materials and Methods"). The numbers on the right of the blot indicate the position of molecular weight standards.
(9E10) revealed the presence of the 68-kDa precursor of the rLHR in the immunoprecipitates (identified by an arrow on the right panel of Fig. 3) but not that of the higher molecular weight forms of the rLHR. The presence of the mature form of the rLHR in these immunoprecipitates cannot be completely ruled out, however, due to its relative levels compared with that of the immature form (see left panel of Fig. 3

and Ref. 31).
Finally it should also be noted that the formation of an HA-p38 JAB1 /myc-rLHR complex could only be demonstrated by immunoprecipitating the lysates with an anti-HA antibody and probing the immunoprecipitates with an antibody to the myc epitope. Reversing the antibodies used for immunoprecipitation and blotting could not be used to reveal such an association. An explanation for this finding was not sought.
Does the rLHR Affect the Function of HA-p38 JAB1 ?-As mentioned above, HA-p38 JAB1 was initially identified by Claret and co-workers (22) in a yeast two-hybrid screen using the activation domain of c-Jun as bait. These authors also showed that co-transfection of p38 JAB1 with c-Jun enhanced the transcriptional activation of c-Jun as measured in AP-1-driven transactivation assays. Since this is the only known functional assay for p38 JAB1 , we used it in an attempt to further document an interaction between the rLHR and HA-p38 JAB1 .
The experiments presented on the left portion of Fig. 4 reproduce some of the results initially published by Claret and co-workers (22). These experiments readily demonstrate that 1) co-transfection of HeLa cells with c-Jun transactivates AP1luc, a luciferase construct driven by the AP-1 promoter, 2) co-transfection of HeLa cells with HA-p38 JAB1 does not transactivate AP1-luc, and 3) co-transfection of Hela cells with c-Jun and HA-p38 JAB1 results in an enhancement of transactivation above and beyond that detected with c-Jun alone. Thus, whereas the relative luciferase activity of cells co-transfected with 0.1 g of AP1-luc cDNA and 0.2 g of c-Jun cDNA is ϳ9, the addition of 2 or 5 g of HA-p38 JAB1 cDNA to these transfection mixtures increase relative luciferase activity to ϳ17 and ϳ28, respectively. We reasoned that if HA-p38 JAB1 can interact with c-Jun and with the rLHR, overexpression of the rLHR may result in a decrease in the ability of HA-p38 JAB1 to potentiate the c-Jun-mediated transactivation of the AP1-luciferase reporter. As shown in the right portion of Fig. 4, this was indeed found to be the case. The relative luciferase activity of cells co-transfected with 0.1 g of AP1-luc cDNA, 0.2 g of c-Jun cDNA, and 2 g of HA-p38 JAB1 cDNA was ϳ17 and the addition of 0.1, 0.5, or 0.8 g of rLHR cDNA to these transfection mixtures had no effect on luciferase activity. The addition of 2 g of rLHR cDNA, however, resulted in a return of the transactivation signal elicited by c-Jun ϩ HA-p38 JAB1 (i.e. ϳ17) to that elicited by c-Jun alone (i.e. ϳ9). These results also imply that, despite our inability to detect immunoreactive HA-p38 JAB1 in the nucleus (cf. Fig. 1, B and C), a portion of the transfected HA-p38 JAB1 must localize to this structure.
Does HA-p38 JAB1 Affect the Expression or Functional Properties of the rLHR?-While our experiments were in progress, two additional papers on p38 JAB1 were published. One of these (24) identified p38 JAB1 as a subunit of a novel and very large (ϳ450 kDa) protein complex related to the plant COP9 signaling complex and the 26 S proteasome (26,40). The other (23) showed that p38 JAB1 binds to p27 Kip1 , a cyclin-dependent kinase inhibitor, and promotes its degradation.
These results, together with those presented above, prompted us to examine the possibility that p38 JAB1 affected the turnover of the 68-kDa rLHR precursor. Whereas this is best done by pulse-chase experiments, such an approach is difficult in this case because some of the 68-kDa rLHR precursor is degraded, while some of it is also converted to the 85-kDa mature cell surface rLHR (31,33,39). Thus, the turnover of FIG. 2. In vitro interaction of HA-p38 JAB1 with portions of the rLHR expressed as GST fusion proteins. 35 S-Labeled HA-p38 JAB1 was prepared by in vitro transcription/translation, and 15-l aliquots of the labeled product were incubated (overnight at 4°C) with 20-g aliquots of the different portions of the rLHR expressed as GST fusion proteins. The GST fusion proteins were then were allowed to bind to 20-l aliquots of a glutathione-agarose resin (50% suspension) for 2 h at room temperature, and the agarose beads were collected by centrifugation. After washing the bound material was eluted and resolved on SDS gels, and the 35 S-labeled HA-p38 JAB1 was detected with a PhosphorImager. The lanes marked 1L, 2L, 3L, and CT refer to GST fusion proteins of the first (residues 365-373), second (residues 441-460), and third (residues 526 -548) intracellular loops and the extreme C-terminal tail (residues 632-674) of the rLHR, respectively. B refers to GST alone, and the lane labeled input contains an aliquot of the 35  this precursor is best measured by following the time course of incorporation of 35 S-labeled amino acids (a mixture of cysteine and methionine) into the precursor using a time course when its conversion into the 85-kDa mature receptor is not yet detectable (33,39). By fitting the incorporation data to non-linear equations that describe protein turnover (see "Materials and Methods"), one can readily calculate the rate constant for degradation (k deg ) and the rate of synthesis (S) of the 68-kDa rLHR precursor. Fig. 5 summarizes the results of several experiments in which the time course of incorporation of 35 S-labeled amino acids into the 68-kDa rLHR precursor was followed and used to calculate these two parameters. The k deg shown correspond to degradation half-lives of ϳ60 and ϳ22 min for the 68-kDa rLHR precursor in cells co-transfected with an empty vector or HA-p38 JAB1 , respectively. 2 Since at steady state the levels of the 68-kDa rLHR are dictated by the S/k deg ratio, one can readily calculate that the steady state levels of this precursor in cells co-transfected with an empty vector or HA-p38 JAB1 are 114 (i.e. 1.32/1.2 ϫ 10 Ϫ2 ) and 48 units/cell (i.e. 1.52/3.2 ϫ 10 Ϫ2 ), respectively. These results clearly demonstrate that the expression of HA-p38 JAB1 reduces the steady state levels of the 68-kDa rLHR precursor to ϳ40% of control. They also show that this reduction is more likely due to a ϳ3-fold increase in the rate constant of degradation (k deg ) of the 68-kDa rLHR, rather than to a reduction in the rate of synthesis (S).
Two independent approaches were next used to confirm the observed decrease in the levels of the 68-kDa rLHR precursor induced by the expression of HA-p38 JAB1 . In one approach, we measured the levels of the 68-kDa rLHR precursor by 125 I-hCG binding to detergent extracts prepared from intact cells that had been subjected to surface proteolysis (31). This approach is useful to determine the relative amount of the 68-kDa precursor because limited proteolysis of intact cells results in the exclusive destruction of the 85-kDa mature cell surface receptor while maintaining the levels of the 68-kDa precursor (31). Since the 68-kDa precursor binds hCG with the same affinity as the 85-kDa mature receptor, the binding of 125 I-hCG to this precursor can be readily measured following detergent extraction of the proteolyzed cells (31). The results of these experiments are presented in Table I and again show that expression of HA-p38 JAB1 reduces the steady state levels of the 68-kDa rLHR precursor. In contrast to the data obtained by 35 S incorporation that revealed a ϳ60% reduction in the levels of this precursor (see above), the binding assays revealed only a 24% reduction, however (Table I).
For the second approach, we measured the recovery of 125 I-hCG binding to the cell surface after proteolysis of intact cells (31,33). We have previously shown that following proteolytic destruction of the 85-kDa cell surface rLHR, the complement of cell surface rLHR (which can be conveniently measured by 125 I-hCG binding to intact cells) can be readily replenished by the externalization of the 68-kDa rLHR precursor and by the synthesis of new receptors (31). We reasoned then that, since the expression of HA-p38 JAB1 enhances the degradation of the 68-kDa rLHR precursor (thus decreasing its steady state levels, cf. Fig. 5), this should result in a reduction in the amount of mature rLHR that is replenished at the cell surface after proteolysis of intact cells. The results of these experiments are summarized in Fig. 6 and show that, as predicted, the recovery of mature rLHR following proteolysis of intact cells is lower in cells co-transfected with HA-p38 JAB1 than in cells co-transfected with an empty vector. From these data one can calculate (see "Materials and Methods") that the rate of externalization  (E) of the rLHR is 14.9 and 10.3 units/h in cells co-transfected with an empty vector or HA-p38 JAB1 , respectively, while the rate constant for internalization (k int ) of the cell surface rLHR is 0.133/h and 0.150/h in cells co-transfected with an empty vector or HA-p38 JAB1 , respectively. These rate constants for internalization correspond to internalization half-lives of 313 and 277 min for cell surface rLHR in cells co-transfected with an empty vector or HA-p38 JAB1 , respectively. 3 Since at steady state the levels of the cell surface rLHR are dictated by the E/k int ratio, one can readily calculate that the steady state levels of the cell surface rLHR in cells co-transfected with an empty vector or HA-p38 JAB1 are 112 (i.e. 14.9/0.133) and 69 units/cell (i.e. 10.3/0.150), respectively. These results clearly show that the expression of HA-p38 JAB1 reduces the recovery of the mature rLHR following proteolysis. They also show that this reduction is more likely due to a reduction in the rate of externalization rather than by increasing the rate constant for internalization.
Finally, since transfected cells express an excess of the 68-kDa precursor over the 85-kDa mature receptor (31), we predicted that co-transfection of 293 cells with HA-p38 JAB1 would have little or no effect on the expression of the 85-kDa mature rLHR in cells that had not been proteolyzed and allowed to recover. This prediction was tested by measuring 125 I-hCG to intact cells that had not been proteolyzed, and it turned out to be correct, as shown in Table II. The data presented in Table II also show that expression of HA-p38 JAB1 had no effect on the ability of transfected cells to respond to hCG or cholera toxin with an increase in cAMP accumulation. DISCUSSION A yeast two-hybrid screen of a human kidney 293 cell library with the C-terminal tail of the rLHR identified p38 JAB1 as an rLHR binding partner (Figs. 2 and 3). While detailed studies of the regions of the rLHR that interact with p38 JAB1 were not pursued, in vitro binding assays show that p38 JAB1 can interact not only with the C-terminal tail of the rLHR, but also with its third intracellular loop (Fig. 2). This is an interesting finding since some proteins that participate in GPCR signaling, such as G proteins and arrestins, are also known to interact with these two broad regions of other GPCRs (1,2,41). Transfection of 293 cells with HA-p38 JAB1 showed that this protein is localized mostly in the cytosol (Fig. 1), that it interacts mostly with the 68-kDa precursor of the rLHR (Fig. 3), and that it promotes its degradation (Fig. 5). The enhanced degradation of the rLHR precursor induced by expression of HA-p38 JAB1 leads to a 30 -60% reduction in the steady state levels of this precursor (Figs. 5 and 6 and Table I) but has no effect on the levels of the mature cell surface rLHR or the ability of cells to respond to an LHR agonist with increased cAMP accumulation (Table II).
The product of the human JAB1 gene, a 37.5-kDa protein designated here as p38 JAB1 , was discovered by Claret and co-workers (22) using the activation domain of c-Jun as bait in a yeast two-hybrid screen of a B-lymphocyte cDNA library. The JAB1 transcript is widely expressed, however (22). In this initial study, p38 JAB1 was reported to be a nuclear protein and it was found to interact with the ␦ domain of c-Jun, to stabilize the complexes formed by c-Jun and AP-1 sites, and to potentiate the c-Jun-mediated transactivation of an AP-1 reporter gene (22). Two years later, Seeger and co-workers (24) showed that JAB-1 is a component of a large (ϳ450 kDa) protein complex that they isolated from outdated human blood and initially called JAB1-containing signalosome but later renamed COP9 signalosome (25). This new signaling complex was localized to the cytosol of a human mesothelioma cell line, and it was reported to phosphorylate IB␣, the p105 precursor of NF-B and c-Jun (24,25). More recently the mouse homologue of the human JAB1 gene was identified by Tomoda and co-workers (23) using the C-terminal domain of the cyclin-dependent-kinase inhibitory protein (p27 Kip1 ) as bait in a yeast two-hybrid screen of a mouse T cell lymphoma library. These authors also reported that transfection of a FLAG epitopetagged p38 JAB1 into NIH3T3 cells resulted in a preferential (ϳ80%) nuclear localization of FLAG-p38 JAB1 , whereas cotransfection of the cells with FLAG-p38 JAB1 and HA-p27 Kip1 resulted in the translocation of HA-p27 Kip1 and FLAG-p38 JAB1 from the nucleus to the cytoplasm and accelerated the degradation of HA-p27 Kip1 . Our experiments show that HA-p38 JAB1 transfected into 293 cells is localized in the cytosol (Fig. 1) and that it reduces the levels of the 68-kDa intracellular precursor of the rLHR (Figs. 5 and 6 and Table I) by promoting its 293T cells were transfected with myc-rLHR ϩ an empty vector or myc-rLHR ϩ HA-p38 JAB1 as indicated. Twenty-four hours after transfection the cells washed and treated with protease XIV to destroy the cell surface rLHR (31). After removal of the residual protease, the cells were solubilized with a non-ionic detergent and used to measure 125 I-hCG binding as described elsewhere (31). Each value shown represents the mean Ϯ S.E. of four independent transfections. *, significantly different (p Ͻ 0.05) from cells co-transfected with pcDNA3.1.  6. Effect of HA-p38 JAB1 on the time course of replenishment of the cell surface rLHR following proteolysis of the intact, transfected cells. 293T cells were transfected with myc-rLHR ϩ an empty vector or myc-rLHR ϩ HA-p38 JAB1 as indicated. Twenty-four hours after transfection, the cells washed and treated with protease XIV to destroy the cell surface rLHR as described (31). After removal of the residual protease, the cells were placed back in protease-free medium and the recovery of the cells surface rLHR was monitored by measuring 125 I-hCG binding to the intact cells as described under "Materials and Methods" and Ref. 31. Each point represents the mean Ϯ S.E. of four independent transfections, and the data are expressed as percentages of amount of 125 I-hCG bound to the cells transfected with myc-rLHR ϩ an empty vector at the 24-h time point. The lines shown through the points represent the nonlinear least-square fit to the data points using Equation 2, as described under "Materials and Methods." The rate of externalization (E) and the rate constant for internalization (k int ) of the cell surface rLHR were also calculated using Equation 2, as described under "Materials and Methods." The units for E are arbitrary units because all incorporation data are expressed as percentages of the maximum amount of radioactivity incorporated (i.e. that detected in the 24-h time point in cell co-transfected with myc-rLHR and an empty vector). *, significantly different (p Ͻ 0.05) from cells not expressing HA-p38 JAB1 . degradation (Fig. 5). A nuclear localization for some of the transfected HA-p38 JAB1 , however, can also be inferred from the finding that it potentiates the c-Jun-mediated transactivation of an AP-1 reporter gene in HeLa cells (Fig. 4). It should also be mentioned that most of these experiments have been done by localizing the product of a transfected plasmid, rather than the endogenous product, and that the subcellular localization of these two products may not be the same (25).
Since there are now three proteins (c-Jun, p27 Kip1 , and the rLHR) that have been shown to interact with p38 JAB1 and (at least portions of) their p38 JAB1 binding domains have been identified, we aligned the amino sequences of these regions and found some interesting homologies that may form a p38 JAB1binding motif (Fig. 7). It should be noted, however, that while the regions of c-Jun and p27 Kip1 shown in Fig. 7 are likely to represent their respective binding sites for p38 JAB1 , the region shown in Fig. 7 for the rLHR is only one of the two rLHR regions (the other being the third intracellular loop; cf. Fig. 2) that interact with p38 JAB1 . As other p38 JAB1 binding partners are discovered and their p38 JAB1 binding domains mapped, it will be interesting to determine if they also align well to the sequences shown in Fig. 7.
It should be clear from the foregoing discussion that there is no consensus on the subcellular localization (i.e. nuclear versus cytosolic) or the role(s) that p38 JAB1 plays in cellular functions. p38 JAB1 is a member of a family of proteins known as the Mov-34 family (42,43). This family of proteins shares a conserved domain (the MPN domain) that spans ϳ140 residues located toward their N terminus and is predicted to have an ␣/␤ structure (43). Members of the Mov-34 family include, in addition to p38 JAB1 , the S12 subunit of the 26 S human proteasome, Mov-34, which is the mouse homologue of human S12, two different subunits (p40 and p47) of the mammalian translation initiation factor 3 (eIF3), several transcription factors such as the product of the pad1 ϩ gene in Schizosaccharomyces pombe, and a few other mammalian proteins of unknown function (see alignments in Fig. 6 of Ref. 42). Since may of these proteins are members of multisubunit complexes (i.e. the 26 S proteasome, eIF3, and the COP9 signalosome) or are known to form homoor heterodimers (i.e. transcription factors), it has been proposed that the MPN domain is involved in protein-protein interactions either in a homotypic (i.e. with other proteins that have MPN domains) or in a heterotypic fashion (i.e. with proteins that do not have an MPN domain). In fact, several other subunits of the 26 S proteasome, eIF3, and the COP9 signalosome are characterized by a different homology domain (the PCI domain) that is purely ␣-helical, spans ϳ200 residues and is generally located at their C terminus (43). Thus, heterotypic interactions between proteins with MPN and PCI domains have been postulated as being involved in for the formation of the 26 S proteasome, eIF3, and the COP9 signalosome (24,26,40,43). If this is the case, then it is possible that p38 JAB1 localizes to the nucleus (as reported in Refs. 22 and 23), where it serves as a co-activator of transcription, as well as to the cytosol ( Fig. 1 and Refs. 23 and 24), where it may be a component of multiprotein complexes such eIF3, the COP9 signalosome, and the 26 S proteasome. The relative distribution of p38 JAB1 between the nucleus and the cytosol may be a function of the cells used, the metabolic state of the cells, or it may also be a reflection of the different functions of p38 JAB1 (see below).
Whereas all groups that have examined the potential role of p38 JAB1 on AP-1-mediated transcription agreed that this molecule potentiates the transcriptional activation induced by c-Jun ( Fig. 4 and Refs. 22 and 23), it seems increasingly clear that p38 JAB1 plays other roles in cellular functions. It may serve as a docking site for the phosphorylation of c-Jun catalyzed by the COP9 signalosome (24,25), a process that leads to the activation and or stabilization of c-Jun. The studies of Tomoda and co-workers (23) and those reported here (Fig. 5) emphasize the involvement of p38 JAB1 in the degradation of p27 Kip1 and the 68-kDa rLHR precursor, respectively. The mechanisms by which p38 JAB1 enhances the degradation of these two protein binding partners remain to be investigated, however. In this respect, it is worth noting that in a recent publication Naumann and co-workers (25) also reported decreased levels of endogenous c-Jun in HeLa cells transfected with p38 JAB1 . In this cell line most of the exogenously expressed p38 JAB1 is not incorporated into the COP9 signalosome (25), however, and the reduced levels of c-Jun detected upon transfection with p38 JAB1 seem to be a result of the binding of the endogenous c-Jun to the transfected and unincorporated p38 JAB1 rather than to the endogenous p38 JAB1 associated with the COP9 signalosome. This competition prevents the binding of the endogenous c-Jun to the COP9 signalosome, a process FIG. 7. Amino acid sequence alignment of the regions of the rLHR, p27 Kip1 and c-Jun that interact with p38 JAB1 . The 632-674 region of the rLHR (i.e. the region used as bait in the screen employed here to identify p38 JAB1 ) was aligned to the 97-151 region of human p27 Kip1 and the ␦ domain of human c-Jun (residues 31-57). These latter two domains were also identified as the domains of p27 Kip1 and c-Jun, respectively, that interact with p38 JAB1 (22,23). The sequences for the rLHR (accession no. P16235), human c-Jun (accession no. P05412), and human p27 Kip1 (accession no. P46527) were obtained from the SwissProt data bank and were aligned using ClustalW using an identity matrix. Identical residues are shaded in dark gray, and homologous residues are shaded in light gray. Gaps introduced for optimal alignment are shown by dashes. A "consensus sequence" comprising identical or homologous residues in at least two of the three sequences is shown at the bottom. Homologous residues in this consensus sequence are indicated by dots.

TABLE II
Effect of HA-p38 JAB1 on hCG binding and agonist-stimulated cAMP accumulation in intact cells 293T cells were transiently co-transfected with the myc-rLHR-wt and the indicated constructs. 125 I-hCG binding was measured during a 1-h incubation (room temperature) of the transfected cells with 100 ng/ml 125 I-hCG. Total cAMP accumulation was measured during a 2-h incubation with buffer only, or the indicated concentrations of hCG or cholera toxin as described under "Materials and Methods." A response ratio was calculated for each experiment by dividing the hCG-induced response by the cholera toxin-induced response. Each value represents the mean Ϯ S.E. of three independent transfections. that promotes the phosphorylation (and stabilization) of c-Jun (24,25). The 68-kDa rLHR precursor represents the full-length rLHR that is core-glycosylated and as such is thought to reside in the endoplasmic reticulum (39,44). Some of this precursor matures into the 85-kDa cell surface rLHR (39,44), but the majority seems to be degraded with a relatively short half-life of 60 -100 min ( Fig. 5 and Ref. 33). Since the other two proteins (c-Jun and p27 Kip1 ) that interact with p38 JAB1 are soluble proteins, a question arises as to how an integral membrane protein may be able to interact with p38 JAB1 and why this interaction occurs mostly with the 68-kDa precursor present in the endoplasmic reticulum as opposed to the 85-kDa mature receptor present in the plasma membrane. This issue will also have to be addressed experimentally, but the involvement of soluble multiprotein complexes, such as the proteasome, in the degradation of integral membrane proteins present in the endoplasmic reticulum and the plasma membrane is already well documented (reviewed in Ref. 45).
Since some loss of function mutations of the hLHR associated with human reproductive disorders appear to prevent the conversion of the hLHR precursor into the mature cell surface hLHR (reviewed in Ref. 46), further studies on the functional impact of the interaction of p38 JAB1 with the LHR precursor may also help improve our understanding of the pathophysiology of these disorders.