Human Growth Hormone Enhances Pertussis Toxin-stimulated ADP-ribosylation of Gi in Nb2 Cell Membrane*

The Nb2 node lymphoma cell line has been widely used as a model for investigating lactogen cellular actions. Both pertussis (PTX) and cholera (CTX) toxins modulate lactogen-stimulated Nb2 cell mitogenesis, suggesting G protein involvement in lactogen signal transduction. The following studies were performed to further investigate this possibility. Both PTX-sensitive (41 kDa) and CTX-sensitive substrates (42 and 46 kDa) were identified in Nb2 cell membrane and rec- ognized by specific anti-Gi and anti-G. antibodies, re-spectively. Equal numbers of Nb2 cells were then in- cubated with the lactogen human growth hormone (hGH, 10 ng/ml) for 0-72 h. Membrane protein pre- pared from each time point (60 pg) was compared in toxin-stimulated ADP-ribosylation studies. CTX-stim- ulated ADP-ribosylation was unaffected by prior hGH incubation. was stained in the absence of specific antibody, so it is a nonspecific contaminant.

Jennifer L. LarsenS From the Department of Internal Medicine, University of Nebraska Medical Center, Omaha,  The Nb2 node lymphoma cell line has been widely used as a model for investigating lactogen cellular actions. Both pertussis (PTX) and cholera (CTX) toxins modulate lactogen-stimulated Nb2 cell mitogenesis, suggesting G protein involvement in lactogen signal transduction. The following studies were performed to further investigate this possibility. Both PTX-sensitive (41 kDa) and CTX-sensitive substrates (42 and 46 kDa) were identified in Nb2 cell membrane and recognized by specific anti-Gi and anti-G. antibodies, respectively. Equal numbers of Nb2 cells were then incubated with the lactogen human growth hormone (hGH, 10 ng/ml) for 0-72 h. Membrane protein prepared from each time point (60 pg) was compared in toxin-stimulated ADP-ribosylation studies. CTX-stimulated ADP-ribosylation was unaffected by prior hGH incubation. PTX-stimulated ADP-ribosylation increased 237 2 69% (X f S.E.) compared with 0-h controls (n = 11; p c 0.01) after 4-7 h of hGH incubation then decreased toward 0-h samples by 24 and 72 h. No change in Gi, concentration was observed, but fl subunit concentration increased (146 f 14% at 7 h; p < 0.01; n = 3) in a time course that paralleled the changes in PTX-stimulated ADP-ribosylation. In summary, 1) both Gi and G. were present in Nb2 cell membrane, 2) incubation of cells with a lactogen, hGH, for 4-7 h markedly enhanced PTX-stimulated ADPribosylation of Gh in vitro, whereas CTX-stimulated ADP-ribosylation of G, was unchanged, and 3) although no change in GI, concentration was observed, B subunit concentration increased in parallel with the increase in PTX-stimulated ADP-ribosylation of Gb. These results suggest that hGH may modify PTX-stimulated ADP-ribmylation of GI not by changing Gi, concentration, perhaps by increasing / 3 subunit concentration, enhancing association of Ghby subunits, which, in turn, is preferentially ADP-ribosylated. This may represent a late signal transduction event and may also have implications for other effectors dependent on GImediated events.
Prolactin was one of the first pituitary hormones to be isolated, yet the cascade of events constituting signal transduction for any lactogenic hormone is not yet well understood. The rat Nb2 node lymphoma cell line has been used exten-* This work was supported by National Institutes of Health Grant REN 5 R29 DL40752-02. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ T O whom correspondence and reprint requests should be addressed Dept. of Internal Medicine, University of Nebraska Medical Center, 42nd and Dewey, Omaha, NE 68198-3010. sively to investigate lactogenic hormone actions because it possesses high affinity, specific lactogen receptors, and, on binding of a lactogenic hormone, the cells grow and divide (1-3). Recently, we have shown that bacterial toxins, pertussis toxin (PTX)' and cholera toxin (CTX), altered the response of Nb2 cells to lactogen stimulation (4). PTX and CTX are known to alter G proteins associated with adenylate cyclase, Gi and G., respectively, as well as other signal transduction mechanisms unassociated with adenylate cyclase (see recent reviews: Refs. 5 and 6). These data suggest that lactogen actions may be coupled to one or more GTP binding proteins. The following studies investigate whether incubation of Nb2 cells with a lactogen alters G proteins in Nb2 cell membrane in order to better understand the relationship between G proteins and lactogenic hormones.

EXPERIMENTAL PROCEDURES
The Nb2 cells were graciously provided by Dr. P. W. Gout (University of Mannitoba, Vancouver, Canada). Rabbit anti-G, common antibody was generously provided by Dr. David Manning (University of Pennsylvania (7)), and rabbit anti-Gi2 (5) and anti-G. (8) antibodies were graciously provided by Dr. Suzanne Mumby (University of Texas). Rabbit anti-Gis/Go (9) and rabbit anti-common 0 subunit (10) antibodies were purchased from Du Pont-New England Nuclear.
Radioiodinated goat anti-rabbit antibody was a gift from Dr. K.
Phares (University of Nebraska Medical Center, Omaha, NE). Recombinant human GH (equipotent with National Institutes of Health hGH standard) and '=I-hGH (specific activity 40-70 pCi/pg based on a radioreceptor assay using female rat liver membrane performed in my laboratory) were gifts of Lilly. The following supplies were purchased from the vendors listed Fischer's Leukemic Cell media, antibiotics, and fetal calf serum from GIBCO horse serum, after testing to be lactogen-free, from Flow Laboratories (McLean, VA); electrophoresis-grade materials used in polyacrylamide gels and molecular weight markers from Bio-Rad; polyvinylidene difluoride (PVDF) membrane from Millipore (Bedford, MA); 13'P]NAD from New England Nuclear Research Products (24 Ci/mmol); bacterial toxins from List Biologicals Inc. (Campbell, CA); staining reagents for Western blot from Vector Laboratories, Inc. (Burlingame, CA); and 0 subunit was purchased from Du Pont-New England Nuclear.
The remaining chemicals were from Sigma.

RESULTS
Effect of Incubation of Cells with hGH (10 nglml) Prior to CTXand PTX-stimulated ADP-ribosylation of Nb2 Cell Membrane-Once both CTX and PTX substrates were identified and recognized by specific G, and Gi antisera (see Supplemental Material, Figs. 1-4); Nb2 cells were incubated with hGH for 0-24 h prior to membrane preparation. There The abbreviations used are: PTX, pertussis toxin; CTX, cholera toxin; GH, growth hormone; hGH, human growth hormone; PVDF, polyvinylidene difluoride. was no demonstrable effect of prior hGH incubation on subsequent CTX-stimulated ADP-ribosylation of the 42-and 45-kDa substrates in vitro (Fig. 5), but PTX-stimulated ADPribosylation did change over time (Fig. 6). The change in quantity of ADP-ribosylated 41-kDa protein occurred after a 4-7 h incubation with hGH, an increase of 237 & 69% over 0h controls (X f S.E.; n = 11; Fig. 7). After 24-72 h incubation with hGH, ADP-ribosylation of Gi began to decrease.
Western Blots of Time Course Samples-Gi, concentration was determined by Western blots (for method, see Supplemental Materials, Figs. 3 and 4) using either anti-G, common or anti-Gip, antibodies (Fig. 8). No change in Gi, concentration was detected over this time period (0-24 h) using either antibody. However, p subunit concentration as determined by immunoblot did change over time (Figs. 9 and 10). Two proteins (36 and 27 kDa) were recognized by this anti-common ,8 subunit antibody. The 36-kDa protein was significantly increased by 4 h and remained elevated in the 7-h sample but decreased by 24 h (Figs. 9 and lo), whereas the 27-kDa protein steadily increased over this time period from 0 to 24 h. Recognition of both bands was markedly diminished if the first antibody (anti+ subunit) was preincubated with excess ,8 subunit peptide prior to performance of Western blotting ( Fig. 11, right l a n e ) . Thus, although /? subunit is a 36-kDa protein, the lower 27-kDa band is also specifically recognized by the anti-@ subunit antibody. FIG. 5. Effect of incubation with hGH on CTX-stimulated ADP-ribosylation. The cells were maintained as described previously (4). Nb2 cell membrane was prepared by washing the cells free of fetal calf serum containing media, and after hGH (10 ng/ml) was added for a specified incubation time (0-24 h), cells were homogenized, the membranes were pelleted a t 20,000 X g, and equal amounts of membrane protein (50 pg) as determined by the method of Lowry (28) were subjected to CTX-stimulated ADP-ribosylation. ADP-ribosylation mixture is outlined in Fig. 1 Effect of incubation with hGH on PTX-stimulated ADP-ribosylation. Membrane was prepared from Nb2 cells following different times of incubation with hGH (10 ng/ml): 0-72 h. Equal amounts of membrane protein (50 pg) from each membrane preparation was subjected to PTX-stimulated ADP-ribosylation in uitro. These samples were then applied to a sodium dodecyl sulfate-polyacrylamide gel electrophoresis before being electrophoretically transferred to PVDF membrane (31). An autoradiogram prepared from this PVDF membrane is shown. Lunes 1-8 represent ADP-ribosylation of a 41-kDa protein after 0, 2, 3, 4, 5.5, 7, 24, and 72 h, respectively. A consistent increase in ADP-ribosylation of Gia was observed in membrane prepared from cells incubated with hGH for 4-24 h (lanes 4-8). membrane. Western blot of Nb2 membrane prepared from cells following hGH (10 ng/ml) for 0,2,4,7, and 24 h was performed using an anti-p subunit antibody (1:lOOO). An increase in both a 36-kDa and a 26.6-kDa protein after a 4-h incubation with hGH was observed. The 26.6-kDa protein continued to increase from 4 to 24 h, whereas the 36-kDa protein concentration plateaued and then decreased by 24 h. The higher molecular mass band was stained in the absence of specific antibody, so it is a nonspecific contaminant.

DISCUSSION
What events direct lactogenic hormone cellular actions following receptor binding are still largely unknown. GTP binding proteins have been found to mediate many different hormones' actions. Although G proteins associated with adenylate cyclase have been best characterized, it is now clear that G proteins mediate many other mechanisms of signal transduction, including cGMP phosphodiesterase, inositol triphosphate, and ion channels (for review see Refs. 5, 6, and 11). Because bacterial toxins can alter G protein functions by stimulating the transfer of ADP-ribose to specific acceptor amino acids on the a subunits of sensitive G proteins, they are often used as probes to determine possible involvement of G proteins in effector actions. Thus, when CTX and PTX were found to modulate lactogenic actions in the Nb2 cell (4), the results suggested that one or more G proteins could be involved in lactogenic hormone signal transduction even though the second messenger for any lactogen is currently unknown. The molecular weight of the PTX-sensitive substrate, 41 kDa, was the same as previously described for Gi and was recognized by specific anti-Gi antibodies. Too and co-workers (12) have reported a similar molecular mass (41.5 kDa) for a PTX substrate in Nb2 cell membrane and demonstrated expression of Gi2 and Gi3 mRNA in Nb2 cell (12). However, Barkey and co-investigators (13) identified a much smaller PTX substrate (38 kDa).
The CTX-sensitive substrates, 42 and 45 kDa, were consistent with the molecular masses of G, observed by other investigators due to differential splicing (6), as well as those reported by Too and colleagues in Nb2 membrane (42 and 45 kDa (12)). The molecular mass of the CTX substrate described by Barkey and co-investigators (13) was again lower than either of those seen in these studies (41 kDa). The 42and 45-kDa proteins were not identified with horseradish peroxidase staining, but the 42-kDa protein was recognized by an anti-G, antibody using a more sensitive detection method (data not shown). This suggests G, was present at a much lower concentration in the Nb2 cell membrane as has been described in other cell membranes.
Incubation of Nb2 cells with hGH for 0-24 h prior to preparing membrane altered PTX-stimulated ADP-ribosylation in vitro, whereas CTX-stimulated ADP-ribosylation was unchanged. The change in PTX-stimulated ADP-ribosylation was not explained by a change in Gi, subunit concentration. Other hormones have been reported to alter the ability of a toxin to stimulate ADP-ribosylation of a G protein (14)(15)(16)(17)(18). In all cases, it was assumed that if a hormone can alter ADPribosylation of a G protein, it must in some way alter its conformation or concentration, although the exact mechanism in any of these cases has not yet been identified.
Although there was no change in Gi,, p subunit concentration (36 kDa) did change over time, parallelling the changes observed in PTX-stimulated ADP-ribosylation. Intact G protein heterotrimer is required for PTX-catalyzed ADP-ribosylation of the a subunit (19). Thus, even though Gi, concentration does not change, the fact that more 0 subunit is available, in addition to the described increase in PTXstimulated ADP-ribosylation, suggests that a greater proportion of Gi, may be bound to /3r in the inactive trimeric form.
Regulation of G protein p subunit has been proposed as a mechanism of thyroid hormone action (20,21), but the significance of a change in B subunit to signal transduction in intact cells is not yet well understood. In addition, a second protein, 27 kDa, was also specifically recognized by the antip subunit antibody. The concentration of the 27-kDa protein increased steadily over the period of hGH incubation (0-24 h), so it could represent a degraded fragment of B subunit. A tryptic digest of / 3 subunit with a molecular mass 27 kDa has been identified by other investigators (22), but the significance of this second band is otherwise not known.
As the genes for the prolactin and hGH receptor have now been cloned (23,24), it is known that these receptors do not have the seven trans-membrane receptor spanning regions commonly associated with effectors that interact with GTP binding proteins. Yet there is growing evidence to suggest that lactogenic hormones may interact with G proteins. As already discussed, bacterial toxins inhibit both lactogenic hormone-stimulated DNA synthesis and mitogenesis (4, 12), guanine nucleotides inhibit the binding of '251-human growth hormone to the lactogen receptor (12), and the lactogenic receptor has been cro!s-linked to G proteins using a long molecular length (16.1 A), suggesting some proximity between the two, although shorter lengths were ineffective (25). Classically, signal transduction events are thought to occur within minutes of receptor binding. Yet a prolonged signal may be required for lactogen action in the Nb2 cell as suggested by the fact that stimulation of ornithine decarboxylase is not maximal until 8 h (26), and mitogenesis is prevented if hGH is removed after 4 h, even though maximal expression of the proto-oncogene c-my is achieved by 3 h (27). Even if this late event is not important to signal transduction, it could still represent a change in Gi function that could alter the actions of other effectors dependent on this G protein.
In summary, 1) Gi2, Gi3, and G, were present in Nb2 cell The ADP-ribosylation reaction vas found to be maximal after 60 mln at 3OoC, 90 this tire and temperature were used for all subsequent nN Trls, follaed by centrifuqatlon at 12,OOOq x 10 mi", washed once with 50

membrane. 2) PTX-stimulated ADP-ribosylation of Gi, prepared from cells incubated with hGH for 4-7 h was increased by over 200%. 3) This change in PTX-stimulated ADPribosylation of Gin was not explained by a change in Gi
Incubations. The reactron was stopped by the addltion Of 8OOul Ice-cold 50 W Trla and then solublllzed rn lOOul electrophoresis sample buffer (In1 qlycerol, 2nl 10% SDS, 0.5ml 2-WE1 a t 100°C. x 5 mln. S m p l e e were applied to a discontinuous 120 polyacrylmide qel. Follainq electrophoresis the gel was drled before autoradioqraphy. Ldne 1. Three labeled proteins 0; 42, 4 5 , and 1OOkD are observed. Lane ?. When mmbrane was prepared from Nb2 cells Incubated first wlth CTX (IOnqlmll for 24h. the h q h e r molecular welqht specles (100kDl wa9 not darn-regulated so was secondary to non-speciflc membrane ribosyl-traneferasa activity.  antibody dkd not label the 41kD protein but drd ldentlfy the zein when a radiolabeled second antlbody was used (data not shoun).