Uptake and binding of radiolabelled phenylarsine oxide in 3T3-L1 adipocytes

Transcription

1 Biohem. J. (199) 269, (Printed in Great Britain) Uptake and binding of radiolabelled phenylarsine oide in 3T3-L1 adipoytes Susan C. FROST* and Marie S. SCHWALB Department of Biohemistry and Moleular Biology, University of Florida, Gainesville, FL 3261, U.S.A. 589 Phenylarsine oide (PAO), a trivalent arsenial, has been shown to inhibit insulin-stimulated gluose transport in 3T3- LI adipoytes, impliating viinal dithiols in signal transmission [Frost & Lane (1985) J. Biol. Chem. 26, ]. To assist in the diret identifiation of a PAO-binding protein whih might be involved in this proess, we have synthesized [3H]aetylaminophenylarsine oide ([3H]APAO) from the amino derivative of phenylarsine oide (NPAO). To assess the inhibitory effet of the produt, a dual-labelling eperiment was performed whih showed that [3H]APAO inhibited insulin-stimulated 2-deoy[l-14C]gluose transport in 3T3-L1 adipoytes with a K1 of 21 /um, idential with that of the parent ompound, NPAO. Further haraterization revealed that over a wide onentration range, uptake of the labelled arsine oide was linear. Although the dithiol reagent 2,3-dimeraptopropanol (DMP) reversed PAO-indued inhibition of transport, it had no effet on the uptake of [3H]APAO. In a simple frationation eperiment appro. 5% of the radioativity was assoiated with the ytosoli fration and 5% with the total membrane fration. Identifiation of radiolabelled proteins by non-reduing SDS/PAG revealed fration-speifi binding, although many proteins were observed. Covalent modifiation was time-dependent and ould be reversed by addition of DMP. These data further support a role for viinal dithiols in insulin-stimulated gluose transport. Additionally, the probe desribed may offer a new means with whih to identify the inhibitory protein or, more globally, to investigate mehanisms of ation of viinal dithiol-ontaining proteins. INTRODUCTION It has been known for many years that insulin stimulates the rate of gluose transport in speifi target tissues. Both the insulin reeptor (the f-subunit) (bina et al., 1985) and the gluose transporter (Muekler et al., 1985) are integral proteins of the membrane. Although the signalling event between the two proteins is not known, it has been proposed that stimulation of transport ativity by insulin may our by reruitment of transporters to the ell surfae (transloation) (Cushman & Wardzala, 198; Suzuki & Kono, 198) and/or by intrinsi ativation (Baly & Horuk, 1987; Kahn & Cushman, 1987; Joost et al., 1988; Blok et al., 1988; Calderhead & Lienhard, 1988). It has been further suggested that autophosphorylation of the insulin reeptor and subsequent ativation of the tyrosine kinase play a role in signal transmission (Morgan et al., 1986; llis et al., 1986). Of potential importane is the observation that phenylarsine oide (PAO) inhibits insulin-stimulated gluose transport (but not basal transport) in 3T3-L1 adipoytes (Frost & Lane, 1985). This inhibition ould be reversed by 2,3- dimeraptopropanol (DMP) but not meraptoethanol (MC), suggesting the involvement of viinal dithiols. Further eperiments indiated that PAO interfered with post-reeptor signalling, sine neither insulin binding (Frost & Lane, 1985) nor reeptor tyrosine kinase ativity (Frost et al., 1987) were affeted by the arsenial. However, PAO did inhibit ertain insulinstimulated serine-speifi phosphorylation events in intat ells (Frost et al., 1987). One interpretation of this latter observation is that the ativated insulin reeptor stimulates the ativity of a serine kinase. Other eperiments have been presented whih support the eistene of suh a asade (Czeh et al., 1988; Ray & Sturgill, 1988). However, other data also suggest that PAO may blok a tyrosine-speifi phosphatase (Bernier et al., 1987), perhaps inhibiting the turnover of a signal moleule. With PAO ovalently attahed to a resin, preliminary data suggested that PAO must enter the ell to eliit inhibition (Saylor & Frost, 1989). Thus the goals of this study were to synthesize a labelled arsenial, to haraterize its uptake into 3T3-LI adipoytes, and to identify PAO-binding proteins. This information would provide the basis for future identifiation of the PAO-sensitive target in insulin-stimulated gluose transport, a step toward understanding how insulin funtions inside the ell. XPRIMNTAL Aetylation of para-aminophenylarsine oide The synthesis of aminophenylarsine oide (NPAO) has been desribed (Stevenson et al., 1978). The aetylation reation was performed in a vented hood. The reation vessel ontaining [3H]aeti anhydride (9.5Ci/mmol; 2.63,umol/25 mci) was apped with a serum stopper so that additions ould be made through the septum. The bottom of the vessel was ooled in solid CO2 to ondense the aeti anhydride. The seal was broken and 2.5,mol of NPAO, dissolved in 1 4u1 of anhydrous aetone, was injeted into the reation hamber. The vessel was brought to room temperature and allowed to sit for 6 min, with miing every 15 min. Unlabelled aeti anhydride (1,mol, in 1,ul of anhydrous aetone) was added to the reation and the entire sample was arefully dried under N2. The vessel was washed with 8 1. ml of water. The washes were frozen and dried in a Speed Va rotary evaporator (Savant). The dried preipitate was washed a seond time, dried and resuspended in 4. ml of 95 % ethanol. The solution was larified by addition of 1 M HCI and then dried under N2. The sample was finally suspended in 2 ml of a 1: 1 Abbreviations used: PAO, phenylarsine oide; NPAO, para-aminophenylarsine oide; APAO, aetylaminophenylarsine oide; DMP, 2,3- dimeraptopropanol; MC, meraptoethanol; KRP, Krebs-Ringer phosphate buffer; PBS, phosphate-buffered saline. * To whom orrespondene should be addressed.

2 59 (v/v) solution of water/95 % ethanol. The final speifi radioativity was appro. 6,Ci/,umol of aetylaminophenylarsine oide ([3H]APAO), determined as desribed below. Determination of speifi ativity To determine the onentration of reovered APAO in the above synthesis, a standard urve was generated using a modifiation of llman's proedure (llman, 1959). For eah assay, 12 nmol of DMP was mied with 3 ml of 2 mmphosphate buffer, ph 8., just before analysis. Various onentrations of PAO (-8 nmol) or [3H]APAO were then added to appropriate tubes, eah ontaining 3 ml of the DMP solution. After 5 min, 25 nmol of DTNB [5,5'-dithiobis-(2- nitro)benzoi aid] was added to eah tube and mied immediately. The absorbane of the [3H]APAO samples was read at 412 nm and ompared with the known values for PAO. Cell ulture and deoygluose transport The details of differentiation and transport have been doumented elsewhere (Frost & Lane, 1985). For transport,.1,uci of either deoy[3h]gluose or deoy[i-'4c]gluose was diluted with unlabelled deoygluose to a final onentration of.2 mm. All eperiments were onduted with ells whih were 8-12 days post-differentiation. Where neessary, the speifi details are given in the Figure legends. Uptake and binding of 13HJAPAO 3T3-L1 adipoytes were inubated in serum-free medium for 2 h in a CO2 inubator. The ells were then washed three times with 3. ml of Krebs-Ringer phosphate buffer, ph 7.4 (KRP). The ells were inubated in 1. ml of KRP with or without 1,UMinsulin for 1 min at 37 'C. Following this, the ells were inubated with various onentrations of [3H]APAO in the presene or absene of 4,tM-DMP for the indiated times. The ells were then' washed with 3 3. ml of ie-old phosphatebuffered saline (PBS; 12 mm-naci/2.7 mm-kci/1 mm-sodium phosphate, ph 7.4) and homogenized either in.5 ml of TS (Tris, 1 mm; DTA, I mm; surose, 25 mm; ph, 7.3) for subellular frationation or in.5 ml of SDS (.1 %) for nonfrationated samples. The TS homogenate was entrifuged at 212 g for 7 min at 4 'C. The supernatant (ytosoli fration) was removed and the pellet (membrane fration) was resuspended in.5 ml of TS. A portion of eah fration was assayed for radioativity or was mied with sample dilution buffer (Laemmli, 197) and run on a 12.5 % non-reduing SDS/polyarylamide gel. Before drying, the gels were inubated in 1 M-saliyli aid for 1 h. The dried gels were eposed to pre-flashed film for 7 days. When appropriate, lanes were analysed by laser densitometry. Labelling of proteins in vitro Cells were treated as desribed above with or without insulin for 1 min. The ells were then homogenized in.5 ml of TS and inubated with 5SO,M-[3H]APAO at 37 'C. The membranes were olleted by entrifugation and resuspended in.5 ml of TS. Portions of the ytosoli and membrane frations were ounted for radioativity or mied with sample dilution buffer and run on a 12.5 % non-reduing SDS/polyarylamide gel as desribed above. Preipitation of radiolabelied proteins with trihloroaeti aid After inubation with [3H]APAO in the absene or presene of 2 /sm-dmp, ells were washed with 3 3. ml of PBS and etrated in.1 % SDS. To 1,u of the etrat, 1,ul of 1% trihloroaeti aid was added and inubated on ie for 3 min. After entrifugation for 5 min at 4 C, the pellet was resuspended S. C. Frost and M. S. Shwalbe in 1,l of sample dilution buffer ontaining 1,1u of 1 M- NaOH. Portions were sampled for radioativity. Materials [3H]Aeti anhydride was purhased from Amersham Corp. (Arlington Heights, IN, U.S.A.). Protein standards were purhased from Bio-Rad (Rokville Center, NY, U.S.A.). Insulin was a gift from Dr. Ronald Chane, li Lilly, Co., Indianapolis, IN. U.S.A. NPAO was a gift from Dr. Kenneth Stevenson, University of Calgary, Calgary, Canada. All other supplies were of the highest quality available. RSULTS ffet of arsenials on insulin-stimulated deoygluose transport As shown earlier (Frost & Lane, 1985) and verified here, PAO effetively bloked insulin-stimulated gluose transport (K, 9,uM) (Fig. la). Sine a derivative of this arsenial was to be used for radiolabelling and subsequent identifiation of binding protein(s), it was neessary to demonstrate its effetiveness in the transport assay. Aordingly, we show in Fig. l(a) that NPAO inhibited insulin-stimulated transport in a manner similar to PAO, although the urve was shifted to the right (K1 21 #M). Unlabelled APAO (synthesized as desribed in the perimental setion but using unlabelled aeti anhydride) gave similar results. Using a dual-labelling proedure we then monitored the effet of [3H]APAO on deoy[1-1'4c]gluose transport. With inreasing labelled arsenial onentration, inreasing inhibition of gluose transport was observed (Fig. Ib). An inhibition onstant of 21 /SM was alulated, idential to that of the unlabelled PAO derivatives. Uptake of I3HIAPAO Sine the labelled arsenial inhibited insulin-stimulated transport as did the parent ompound, we then determined elldependent uptake (assoiation) of the radiolabelled PAO. Over a wide range ofonentrations the uptake of the labelled arsenial was linear (Fig. 1 ). At eah onentration, appro. 8 % of the added radioativity was taken up during the 3 min inubation. Appro. 9 % of the ell-assoiated label ould be preipitated with trihloroaeti aid. When ells were preinubated with insulin to first ativate transport followed by NPAO addition, we were able to determine the minimum time required for arsenial ation. As shown in Fig. 2, inhibition of insulin-stimulated transport ould be seen within 2.5 min of NPAO addition, although the half-time was 1 min (inset). The assoiation of labelled arsenial also appeared to be time-dependent, ultimately reahing a steady-state plateau (Fig. 3a). The half-time for assoiation was 1 min, identifial to that for inhibition from the stimulated state. Beause the assoiation data suggested ehange between the ell and etraellular buffer, we also measured the efflu of [3H]APAO (Fig. 3b). As an be seen, ell-assoiated radioativity fell to a steady level as [3H]APAO was released from 'pre-loaded' ells into the buffer. Thus there was a ell-assoiated pool or ompartment (about 3% of the assoiated label) that allowed rapid ehange of [3H]APAO. It should be pointed out that in a omparable gluose transport eperiment, simple washing as was done in the above release eperiment was not suffiient to reverse PAOindued inhibition of insulin-stimulated transport (results not shown). Thus the ehangeable [3H]APAO did not ontribute to the inhibitory proess. As shown below, most of the radioativity was ompleed ovalently to ellular proteins. ffet of thiol reagents on [3HIAPAO uptake It has been shown previously that the dithiol reagent DMP 199

3 Uptake of phenylarsine oide 591 (A) -a J._S <O ) -a I (V a C - C._ ~ a l-, D m : U 4- : L I m 4 1 [Arsenial] (pm) [3H]APAO (#M) prevented and/or reversed the inhibition of insulin-stimulated gluose transport by PAO (Frost & Lane, 1985). This was also true for NPAO-indued inhibition, as shown in Fig. 4(a). Despite its ability to reverse inhibition, DMP had no effet on the assoiation of radiolabelled arsenial to whole ells (Fig. 4b and Table 1). This suggests that the hydrophobi harater of the arsenial rather than its interation with dithiols ditates physial loation. Unlike DMP, MC was ineffetive in reversing inhibition (Fig. 4a). However, MC enhaned assoiation by 4- fold (Fig. 4b) presumably beause of non-speifi interation of the labelled arsenial with free thiols whih are eposed upon redution of oligomeri proteins. Compartmentation of 13HIAPAO To identify the subellular ompartment to whih the labelled arsenial assoiates, a simple frationation was performed. After inubation with [3H]APAO for 3 min in the absene or presene of insulin, the ells were washed, homogenized and separated by 1 entrifugation into membrane and ytosoli frations. Assay of these frations revealed that 52 % of the radioativity was assoiated with the membranes, while a surprising 48 % was assoiated with the ytosol (Table 1). The frations were analysed by non-reduing SDS/PAG. The ytosoli fration ontained at least ten labelled proteins, ranging in moleular mass from 13 kda to 95 kda, several of whih were unique to this fration (inluding 13.9 kda, 15.4 kda, 36 kda) (Fig. 5). The membrane fration also showed unique proteins (3 kda and 64 kda). For omparison, a seond eperiment is shown in whih intat ells were treated with insulin or not, homogenized and seondarily inubated with [3H]APAO. The inubation in vitro labelled proteins of idential migration as in the intat ells but labelling was more intense, presumably beause of greater aessibility to [3H]APAO. Insulin had no signifiant effet on uptake (see Table 1) or binding of [3H]APAO in any fration or under any eperimental ondition. Time-dependent binding of 13HIAPAO and its reversal When ells were preinubated for 1 min with NPAO and then assayed in the presene of insulin and DMP, inhibition of transport was quikly reversed (Fig. 6a). A orresponding eperiment was performed with [3H]APAO to follow proteinspeifi binding and its reversal by DMP. Figure 6(b) shows the time-dependent binding of [3H]APAO to proteins etrated from 71 ~ intat ells at various time points during a 1 min preinubation, 8) [3H]APAO (pm) Fig. 1. Charaterization of effets of APAO in 3T3-L1 adipoytes (a) 3T3-Ll adipoytes were preinubated for 1 min at 37 C in the presene or absene of various onentrations of arsenial (PAO, A; NPAO, *; unlabelled APAO, *). This was followed by inubation with 1 /LM-insulin for 1 min and subsequently with 2 /M-2-deoy[3H]gluose for 1 min. The ells were then washed with 3 3. ml of ie-old PBS and air-dried. The ells were etrated with 1. ml of.1 % SDS and a portion was ounted for radioativity. The data are epressed as uptake of deoygluose in nmol/min per 16 ells, representing the averages of eperiments performed on four different days (n = 8) with a standard deviation of less than 8 %. (b) 3T3-L1 adipoytes were preinubated for 1 min at 37 C in the presene of various onentrations of [3H]APAO, determined from the speifi radioativity of the arsenial, for 1 min. This was followed by inubation with insulin for 1 min and 2 /um-2- deoy[l-14c]gluose for an additional 1 min. Uptake was terminated by washing, as desribed above, and etration to determine transport of deoygluose and uptake of labelled PAO (see ). The data represent the averages of dupliate points from four different eperiments (n = 8), with standard deviations of less than 1%`. The values for insulin-stimulated and basal gluose transport were and nmol/min per 16 ells respetively. () Uptake of [3H]APAO from the eperiment desribed in (b). The data are shown as the average uptake of arsenial in nmol/16 ells, with a standard deviation of less than 1% (n = 8). The duration of the uptake eperiment was 3 min.

4 592 S. C. Frost and M. S. Shwalbe Je 2~~~~~~ 4 OL~~~~~~~~~~3 n, -\ < s O o a 5.5_ Time (min) a) a) Inubation period (min) Total time of eposure to NPAO (min) Fig. 2. Time-dependent inhibition of insulin-stimulated gluose transport Cells were preinubated for 1 min with 1 /M-insulin. NPAO (5 sm) was added and transport was measured for 2.5 min intervals at various times thereafter. The inset shows the same data plotted on a log sale for half-time determination. A, Basal; *, insulinstimulated; *, insulin plus NPAO. analogous to the eperiment with NPAO desribed in Fig. 6(a). If DMP was added at the same time as [3H]APAO, no labelled proteins were observed (results not shown). If DMP was added after O min (again analogous to Fig. 6a), DMP immediately aused a derease in the amount of label bound to proteins (Fig. 6). Densitometri analysis of the autoradiogram was performed to ompare labelled proteins and quantify binding. The atual profiles for lane 4 in Fig. 6(b) (1 min uptake of [3H]APAO) and lane 2 in Fig. 6() (1 min uptake of [3H]APAO followed by 5 min inubation with DMP) are presented in Fig. 7(a). The upper trae reveals that ten major proteins were labelled by the arsenial after 1 min of preinubation (see legend for moleular masses). At 5 min after addition of DMP (lower trae), label remained only in proteins 1, 6, 9 and 1, whih may be resistant to DMP beause of inaessibility, by virtue of either struture or ompartmentation. Fig. 7 also onfirms the time-dependent binding (b) and further shows that within 2.5 min of DMP - C.-. a)x I - ) :- 8 Time of release (min) I--. I-,.XDO * ' O f.5 X CD- Fig. 3. Time-dependent uptake of [3HIAPAO (a) 3T3-L1 adipoytes were inubated with 4 /m (A), 1/M()m 25 /SM (@) or 5/M (U) [3H]APAO for various times at 37 'C. As indiated, the ells were washed and etrated with.1 % SDS and analysed for radioativity. The data are epressed in nmol/16 ells. The standard error in dupliate determinations was less than 1 %. (b) Cells were pre-loaded with 5 /tm-[3h]apao for 3 min at 37 'C. After si washes with PBS over 3 s, the ells were inubated for up to 12 min in 1. ml of KRP ontaining 5 mm-gluose. Radioativity was measured in both the buffer (*) and in SDS-solubilized ells (). The total radioativity (v) (ell-assoiated plus released) did not hange over the ourse of the eperiment. The data represent the average of dupliate determinations, with an error of less than 1 %. 51) Co - Ct = a) en o r- C I * o a)- :> (a) -A A <n 6 n a o: W. -l 4 - -~~~~~~~~ T l D ( D Thiol reagent (jam) Fig. 4. ffet of thiol reagents on transport inhibition and I3HIAPAO uptake (a) Cells were inubated in KRP with 5,tM-NPAO for 1 min. This was followed by the onurrent addition of thiol reagent (A, DMP;, MC) and 1 gum-insulin. After 1 min, transport was measured during a 1 min interval. The data represent the averages of three separate eperiments. The insulin-stimulated and basal values for transport were and nmol/min per 16 ells respetively. (b) Cells were inubated for 1 min in the presene of 5 #mm-[3h]apao. At that point, dithiol reagent (A, DMP;, MC) was added and the inubation was ontinued for an additional 2 min. Cells were washed and etrated in.1 % SDS for ounting of radioativity. As in (a), total eposure to the arsenial was 3 min. The data represent the averages of three separate eperiments. ww - 2r 199

5 Uptake of phenylarsine oide Table 1. Uptake of 13HIAPAO Cells or homogenate was inubated with 25 1tM-[3H]APAO for 3 min followed by frationation as desribed in the perimental setion. The values represent dupliate eperiments with standard error of less than 7/. The orresponding values for gluose transport in similar ells were as follows (in nmol/min per 16 ells): basal, ; 1 4uM-insulin, ; 25 1uM-NPAO+insulin,.72 ±.3; NPAO + insulin + 2,tM-DMP, l-3 Uptake (.p.m./16 ells) -DMP +DMP + Insulin - Insulin + Insulin - Insulin Whole ells Pellet Supernatant Homogenate Pellet Supernatant addition, more than 9 % of the protein-assoiated radioativity was lost (Fig. 7), whih orrelates with the immediate reovery of transport stimulation in the presene of DMP (Fig. 6a). From these data, we onlude that binding of [3H]APAO (and thus inhibition) was dependent on the interation of the arsenial with (a) (b) 1-3M, Mr (b) *:... * ::i :a56 : li ::: -:,.B. :..:..!:::. :.... ::.i!.::... i.i: 8.:-.. :: :'': '."::..... ::: : o 1.5 =._ a ) > Time (min) ().:::..:.::.:..: j.: ::.:.:.:::::. ::.:::.:!: }!.::::: *.... ;.. ::.:.:. ::.: :: :.. :. :.....:.. :.... * :..:::::::. :::: :. : Insulin t Fig. 5. Identifiation of PAO-binding proteins Cells were inubated with or without 1 /zm-insulin for 1 min. Following this, the ells were either inubated for 3 min in 1. ml of KRP with 5 4aM-[3H]APAO (a) or homogenized in.5 ml of TS and then inubated for 3 min with 5 1zM-[3H]APAO (b). In this eperiment the speifi radio-ativity of the [3H]APAO was lowered to 97.p.m./pmol by adding unlabelled NPAO. The ells were washed and homogenized in TS and all samples were separated into ytosoli (lanes 1 and 2) and total membrane (lanes 3 and 4) frations by entrifugation. A sample (representing appro. 3 ells) of eah fration was dissolved in sample dilution buffer and run on a non-reduing SDS/polyarylamide gel. The migration of labelled proteins was ompared with that of known proteins (phosphorylase, 97.4 kda; bovine serum albumin, 66.2 kda; ovalbumin, 42.7 kda; arboni anhydrase, 31 kda; soybean trypsin inhibitor, 21.5 kda; lysozyme, 14.4 kda). t3hiapao ran at the bottom of the gel Fig. 6. ffet of DMP on arsenial binding (a) Cells were preinubated with (*, A) or without () 5 S,M- NPAO for 1 min. At zero time, I,uM-insulin with (*) or without (,A) 4,uM-DMP was added to the ells and transport was eamined at various times. (b) Cells were inubated for various times with 28 gum-[3h]apao (62.p.m./pmol). Cells were washed with PBS and lysed with.1 % SDS. An aliquot was mied with sample dilution buffer and run on a non-reduing SDS gel. Lanes 1-4 represent preinubation times with [3H]APAO of 2.5, 5., 7.5 and 1 min respetively, and show an inrease in labelling intensity with time. () Cells were preinubated with [3H]APAO for 1 min as in (b), at whih point 4,uM-DMP was added. At the appropriate time thereafter ells were homogenized in.1 % SDS and an aliquot was mied with sample dilution buffer for SDS/PAG. Lanes 1-7 represent inubation times with DMP of 2.5, 5., 7.5,1., 12.5, 15. and 3. min respetively, and show a derease in labelling intensity with time. dithiol-ontaining proteins. This was supported by trihloroaeti aid preipitation eperiments, whih demonstrated that the amount of preipitated radiolabelled proteins was dereased from 91 % in the absene of DMP to 19% in the presene of DMP. DISCUSSION The data presented here demonstrate the use of a new labelled derivative of PAO to identify dithiol-ontaining proteins. We

6 594 S. C. Frost and M. S. Shwalbe n (A C 4) to a, (U) a) -r [ 4 [ 2 (b) "-%' 4. 2 / 4 Lane no Time (min) Binding of [3H]APAO O fl- N 'q 1-3 The monothiol redues disulphide bonds whih are distant from ao) M, one another and through this ation an dissoiate subunits of 6 oligomeri proteins. The labelled arsenial an then interat non- A (a) speifially with the eposed sulphydryl groups. It had been our initial hope that we ould utilize MC to redue non-speifi l interations, but the results proved the opposite to be true. On the other hand, the dithiol reagent does not funtion to dissoiate 4 5 proteins but redues redo viinal sulphydryls (Webb, 1966). Sine DMP reverses PAO inhibition of transport and [3H]APAO binding, this strongly suggests that PAO binds to proteins whih ontain viinal dithiols. It is apparent that insulin ation requires the ativation of a protein(s) whih ontains a redo dithiol. R, One ould envision a variety of proesses with whih the arsenial might interfere. Perhaps the most simple would be interalation into membranes, reduing fluidity and thus affeting () membrane funtion. However, uptake of the arsenial was not suffiient for inhibition: ovalent modifiation of neighbouring sulphydryls was required. Although unlikely, based on the hemial speifiity of the arsenial, sulfhydryl modifiation ould ause etensive ross-linking of proteins in either the membrane or other ompartment. However, we have found no evidene that suh ross-linking ours, as no residual protein was observed at the top of the non-reduing gels. Sine neither hormone binding nor kinase ativity of the insulin reeptor are *-. \* affeted by PAO (Frost et al., 1987), one ould suggest that the arsenial interferes with a seondary signal. This ould be aomplished by either diret interation with that signal leading to redution of an endogenous ativity (a kinase or phosphatase), Reversal with DMP or indiretly by limiting aess to the signalling pathway (in- Fig. 7. Densitometri analysis of time-dependent and reversible binding terfering with reeptor internalization). It has been shown previously that internalization of several hormone-reeptor The autoradiograph presented in Fig. 6 was sanned by laser omplees is bloked by PAO (Wiley & Cunningham, 1982; densitometry for quantifiation of binding and identifiation of Hertel et al., 1985), inluding the insulin reeptor in 3T3-C2 major proteins. Atual traing frrom the sans of lane 4 from Fig.' 6(b) and lane 2 from Fig. 6(). TFen major proteins were identified fibroblasts (Knutson et al., 1983) whih serve as the ontrol ell from lane 4 (upper trae) orresp)onding to 13.9 kda (1), 15.4 kda line for 3T3-LI adipoytes. However, it is unlear whether (2), 22.4 kda (3), 3 kda (4), 33ikDa (5), 36 kda (6), 4 kda (7), insulin reeptor internalization is required for ativation of 42 kda (8), 58 kda (9) and 64 kdla (1). Only four proteins ould be intraellular events. resolved from lane 2 (lower trae) orresponding to 13.9 kda (1), Of potential interest is the ytosoli protein whih migrates at 36 kda (6), 58 kda (9) and 64 k] Da (1). The san from eah lane 15.4 kda and an be reversibly labelled with [3H]APAO. During was integrated and the relative a orresponding diretly to (b) and () of Fig. 6. Note that zero time differentiation of the 3T3-LI ell line, translation of adipoyte- 6b) for both panels (b) and (). speifi proteins ours (Makall et al., 1976), one of whih is a is the same point (lane 4 from Fiig. fatty-aid-binding protein whih migrates as a 15 kda speies (Bernlohr et al., 1984). Although there are two ysteine residues in the amino aid sequene of this protein, they are distant from have shown that, in all aspets )f inhibition of insulin-stimulated eah other (Bernlohr et al., 1984), whih would not provide the gluose transport, the labelled arsenial behaves similarly to most stable arsenial-dithiol onfiguration (whih is a five- or PAO, making it a reasonable hoie for identifying potential si-membered ring as ours with PAO-lipoi aid; Webb, 1966). targets responsible for inhibiti on. We an now orrelate that Attempts at positive identifiation of the labelled 15 kda protein inhibition to uptake and bindiing of the arsenial based on the by immunopreipitation with an antibody speifi for the fatty- i.e. we have shown that aid-binding protein have proved unsuessful. Thus the 15 kda studies with the labelled arsenial, inhibition, uptake and binding are all equally time-dependent. protein that binds the labelled arsenial may merely o-migrate Interestingly, inhibition and binding are reversed almost im- with the fatty-aid-binding protein. mediately by the dithiol reagent even though DMP had no effet In summary, we have been able to synthesize a labelled on uptake of the arsenial. This suggests that uptake depends on trivalent arsenial whih has inhibitory properties that are similar the hydrophobi harater of the arsenial (the phenyl ring) to those of the parent ompound. We have used this analogue to whereas inhibition is protein-imediated. Thus DMP displaes quantify uptake and demonstrate binding to dithiol-speifi [3H]APAO from the protein bu,it does not interfere with hydro- proteins. These results ould be useful in identifying the fator phobi interations. We onludle that one of the labelled proteins whose interation with the dithiol-modifying arsenial prevents whih ehibits reversible bindiing is likely to be responsible for insulin-stimulated gluose transport. In a more general sense, the inhibition of insulin-stimul; ated gluose transport. this labelled arsenial may be useful in analysing the mehanism Unlike DMP, treatment with MC substantially inreased of ation of dithiol-ontaining proteins. uptake and protein-speifi birnding (results not shown) of the labelled arsenial. (It should 1be noted that neither DMP nor We are indebted to Dr. Kenneth Stevenson who offered advie and MC at the onentrations us;ed in these eperiments had any support during the ourse of these eperiments. We would also like to effet on basal or insulin-stimialated transport.) This differene thank Dr. David Bernlohr for providing antibody against the 15 kda distinguishes the hemial speiifiity of these two thiol reagents. fatty-aid-binding protein. This work was supported by grants to S.C.F. 199

7 Uptake of phenylarsine oide from the Amerian Heart Assoiation (Florida affiliate) and the National Institutes of Health (DK39-O1A 1). RFRNCS Baly, D. L., & Horuk, R. (1987) J. Biol. Chem. 262, Bernier, M., Laird, D. & Lane, M. D. (1987) Pro. Natl. Aad. Si. U.S.A. 84, Bernlohr, D. A., Angus, C. W., Lane, M. D., Bolanowski, M. A. & Kelly, T. J. (1984) Pro. Natl. Aad. Si. U.S.A. 81, Blok, J., Gibbs,. M., Lienhard, G.., Slot, J. W. & Geuze, H. J. (1988) J. Cell Biol. 16, Calderhead, D. M. & Lienhard, G.. (1988) J. Biol. Chem. 263, Cushman, S. W. & Wardzala, L. J. (198) J. Biol. Chem. 255, Czeh, M. P., Klarlund, J. K., Yagaloff, K. A., Bradford, A. P. & Lewis, R.. (1988) J. Biol. Chem. 263, bina, Y., llis, L., Jarnagin, K., dery, M., Graf, L., Clauser,., Jing- Hsiung, O., Masiarz, F., Kan, Y. W., Goldfine, I. D., & Rutter, W. J. (1985) Cell 4, llis, L., Clauser,., Morgan, D., dery, M., Roth, R. A. & Rutter, W. J. (1986) Cell 45, llman, G. L. (1959) Arh. Biohem. Biophys. 82, 7-77 Frost, S. C. & Lane, M. D. (1985) J. Biol. Chem. 26, Frost, S. C., Kohanski, R. A. & Lane, M. D. (1987) J. Biol. Chem. 262, Hertel, C., Coulter, S. J. & Perkins, J. P. (1985) J. Biol. Chem. 26, Joost, H. G., Weber, T. M. & Cushman, S. W. (1988) Biohem. J. 249, Kahn, B. B. & Cushman, S. W. (1987) J. Biol. Chem. 262, Knutson, V. P., Ronnett, G. V. & Lane, M. D. (1983) J. Biol. Chem. 258, Laemmli, U. K. (197) Nature (London) 227, Makall, J. S., Student, A. K., Polakis, S.. & Lane, M. D. (1976) J. Biol. Chem. 251, Morgan, D. O., Ho, L., Korn, L. J. & Roth, R. A. (1986) Pro. Natl. Aad. Si. U.S.A. 83, Muekler, M., Caruso, C., Baldwin, S. A., Panio, M., Blenh, I., Morris, H. R., Allard, W. J., Lienhard, G.. & Lodish, H. F. (1985) Siene 229, Ray, L. B. & Sturgill, T. W. (1988) J. Biol. Chem. 263, Saylor, M. J. & Frost, S. C. (1989) J. Cell Biol. 17, 274a Stevenson, K. J., Hale, G. & Perham, R. N. (1978) Biohemistry 17, Suzuki, K. & Kono, T. (198) Pro. Natl. Aad. Si. U.S.A. 77, Webb, J. L. (1966) nzyme and Metaboli Inhibitors, vol 3, pp , Aademi Press, New York Wiley, H. S. & Cunningham, D. D. (1982) J. Biol. Chem. 257, Reeived 18 Deember 1989/6 April 199; aepted 18 April 199