Interaction of tubulin with guanosine 5'-O-(1-thiotriphosphate

Mar 9, 1994 - Departments of Psychiatric Medicine and Neurology, University of Virginia School of Medicine, Charlottesville, Virginia 22908. Received ...
0 downloads 0 Views 851KB Size
Biochemistry 1994, 33, 11884-1 1890

11884

Interaction of Tubulin with Guanosine 5’-0-(1 -Thiotriphosphate) Diastereoisomers: Specificity of the &-Phosphate Binding Region? Shihua Xu* and Felicia Gaskin’.$,$ Departments of Psychiatric Medicine and Neurology, University of Virginia School of Medicine, Charlottesville, Virginia 22908 Received March 9, 1994: Revised Manuscript Received July 26, 1994”

The exchangeable nucleotide-binding site of tubulin has been studied using diastereoisomers A (S,) and B (R,)of guanosine 5’-0-( 1-thiotriphosphate) (GTPaS) in which the phosphorus atom to which sulfur is attached is chiral. GTPaS(A) (10 pM) nucleated assembly of purified tubulin (20 pM) into microtubules in buffer containing 0.1 M 2-(N-morpholino)ethanesulfonic acid with 3 m M Mg2+and 1 m M EGTA, p H 6.6 at 37 “C. With 0.2 m M GTPaS(A), the critical concentration (Cc; minimum protein concentration required for assembly) was 8 p M tubulin. Neither 0.2 m M G T P nor GTPaS(B) promoted microtubule assembly in buffer with 0.5-6.75 m M Mg2+and 20-70 p M tubulin. The Ccvalues for GTPaS(A)-induced assembly of tubulin in buffer with 30% glycerol and of microtubule protein (tubulin and microtubule-associated proteins) in buffer were lower than for GTP. GTPaS(A)-induced microtubules were more stable to the cold and to e a 2 + . GTPaS(A) and G T P but not GTPaS(B) bound tightly to tubulin at 4 OC. Although GTPaS(B) did not nucleate assembly, it did bind to tubulin since it was incorporated into the growing microtubule. Both isomers were hydrolyzed in the microtubules. These studies show that GTPaS(A) promotes tubulin assembly better than G T P and GTPaS(B) and that there is stereoselectivity at the a-phosphate binding region of tubulin. The stereoselectivity may be due to different MgGTPaS(A) and -(B) interactions with tubulin. ABSTRACT:

Tubulin is a unique GTP binding protein and has two GTP binding sites, one nonexchangeable and one exchangeable (Weisenberg et al., 1968; Jacobs et al., 1974). In the microtubule, the exchangeable GTP is hydrolyzed and both nucleotides are nonexchangeable. Since GTP at the exchangeable binding site may have several roles [reviewed by Bayley et al. (1993)], Le., to promote nucleation and/or propagation of the growing microtubule, to effect dynamic instability of microtubules [also reviewed by Caplow (1992), Erickson and O’Brien (1992); Mandelkow and Mandelow (1992)], and to transfer GTP to a G-binding protein (Roychowdhury et al., 1993), it is important to understand the active site of GTP. Mg2+ is required for GTP to bind to tubulin (Correia et al., 1987). Mg2+ and other metals have dramatic effects on microtubule assembly, and some of these are due to the nucleotide-metal bound at the active site on tubulin (Gaskin, 1981; O’Brien et al., 1990; Shearwin & Timasheff, 1992; Hamel et al., 1992). Whether Pi from GTP is released slowly or rapidly after incorporation has not been resolved. Melki et al. (1990) report a slow release of Pi with a t l p of 50 s. Stewart et al. (1990) found no evidence for transiently bound Pi either as GTP or GDP-Pi and attribute their results to be consistent with a slow conformational change in tubulin subunits after GTP hydrolysis and Pi release. Studies with diastereoisomers should be helpful in understanding the role of GTP and metals at the active site of tubulin. Diastereoisomers A (S,) and B (Rp) of GTPaS’ (Figure 1) and of GTPPS are GTP analogues which contain sulfur in place of a nonbridging oxygen atom in one of the phosphoryl units of GTP. Many kinases and nucleoside triphosphatases This work was supported by NIH Grant NS24489.

* Correspondence should be addressed to this author at the Psychiatry

Department, Box 203, University of Virginia Health Sciences Center, Charlottesville, VA 22908. t Department of Psychiatric Medicine. Department of Neurology. Abstract published in Aduance ACS Abstracts, September 1 , 1994. @

0006-2960/94/0433-11884$04.50/0

B (RP) FIGURE1: Diastereoisomers of GTPaS. G is guanine.

bind their substrates as P,y-bidentate or as a,P,y-tridentate chelates with Mg2+(Eckstein, 1985; Leyhet al., 1985). Since it has been reported that Mg2+chelates oxygen 3 1 000 times more strongly than sulfur (Pecoraro et al., 1984), screw sense specificity for binding may be due in part to the Mg-nucleotide complex. Previous studies showed that most of the GTPPS(B) was not hydrolyzed in polymers (bundles of sheets and microtubules) assembled from tubulin in buffer with 30% glycerol whereas GTPPS(A) promoted assembly into microtubules and I Abbreviations: Cc, critical concentration; CrGTP, chromium guanosine S’-triphosphate;EGTA, [ethylenebis(oxyethylenenitrilo)]tetraacetic acid; EM, electron microscopy; FPLC, fast performance liquid chromatography; GDPaS, guanosine 5’-0-( 1 -thiodiphosphate);GMPS, guanosine 5’-0-(thiophosphate); GTPaS, guanosine 5 ’ - 0 ( 1-thiotriphosphate);GTPBS,guanosine5’-Q(2-thiotriphosphate);GTPyS, guanosine 5’-0-(3-thiotriphosphate); HPLC, high-performanceliquid chromatography; MAPS,microtubule-associatedproteins; Mes, 2-(N-morpholino)ethanesulfonic acid; MTP, two-cycle microtubule protein (tubulin + microtubule-associated proteins); PAGE, polyacrylamide gel electrophoresis; SDS, sodium dodecyl sulfate.

0 1994 American Chemical Society

Interaction of Tubulin with GTPaS Diastereoisomers was hydrolyzed (Roychowdhury & Gaskin, 1988). The GTPPS(B)-induced polymers were more cold stable than the GTPPS(A)-induced microtubules, which were slightly more stable than the GTP-induced polymers. Mg2+ (2-5 mM) had minimal effects on the tubulin polymersinduced by GTP@ isomers. With microtubule protein (MTP) in buffer, both isomers promoted assembly into microtubules and 75% of the bound GTPPS(B) isomer and all of the GTPPS(A) were hydrolyzed. Studies with MTP and Mg2+ showed that GTPPS(A) required lower concentrations of Mg2+to induce rings and cross-linked rings which were inactive in assembly. Thus, there is stereoselectivity at the P-phosphorus of the exchangeable nucleotide-binding site on tubulin. Studies in this paper using GTPaS diastereoisomers demonstrate that there is also stereoselectivity at the a-phosphorus group. In particular, GTPaS(A) binds tighter than GTPaS(B) and promotes assembly much better than GTP and GTPaS(B). Both A and B isomers of GTPaS can be incorporated into the growing microtubule and are hydrolyzed.

EXPERIMENTAL PROCEDURES Preparation of Microtubule Protein and Tubulin. Microtubule protein (MTP) was prepared from porcine brain by two cycles of assembly-disassembly as described previously (Shelanski et al., 1973). Tubulin was purified from MTP by Mono Q anion-exchange column chromatography using fastperformance liquid chromatography (FPLC) (Roychowdhury & Gaskin, 1986b). This tubulin is 99% pure on the basis of SDS/PAGE and contains little nucleosidediphosphate kinase, adenylate kinase, and ATPase activities. Aliquots of protein in 0.1 M 2-(N-morpholino)ethanesulfonic acid (Mes) with 0.5 mM MgC12 and 1 mM EGTA, pH 6.6 (buffer A), were stored in liquid nitrogen until use. On the day of the experiment, 1-2 mL of protein was dialyzed for 2 h against 2 X 500 mL of buffer A (buffer change hourly) in a CrowEnglander-type thin-film microdialyzer at 4 "C. The dialyzed sample was warmed to 37 OC for 30 min, cooled to 4 OC for 30 min, and centrifuged at 260 OOOg for 6 min at 4 OC in a Beckman TL-100 centrifuge. The supernatant was diluted with buffer A or buffer containing 30% (v/v) glycerol to protein concentrations of approximately 2 mg/mL for tubulin and 1.8 mg/mL for MTP. Protein concentrations weredetermined by the method of Lowry (Lowry et al., 1951) using a BSA standard curve. Assembly Studies. Tubulin or MTP in buffer A with or without 30% glycerol and with the addition of MgC12 or MgS04 (to 3 mM, unless indicated otherwise for tubulin) was incubated with different nucleotides at 4 "C for 25 min. The incubated protein was transferred to cold cuvettes and placed in a thermostatable sample changer at 37 OC. Microtubule assembly was monitored by turbidity at 350 nm by using the Varian DMS 300 recording spectrophotometer as previously described (Gaskin et al., 1974). Microtubule disassembly and reassembly was done by transferring the cuvettes with protein to 4 OC and placing them back to 37 OC. The assembly of tubulin was also monitored by centrifugation (Gaskin et al., 1974) and by negative staining electron microscopy. The JEOL lOOCX was used to analyze the grids. Length determinations were done at time 10 and 30 min as described by Roychowdhury and Gaskin (1988). The error is 1 S D assuming a Gaussian distribution. The critical concentrations (Cc) for assembly were determined by using turbidimetric and sedimentation studies as previously described (Gaskin et al., 1974). Taxol (supplied by Dr. M. Suffness, Natural Products Branch, NCI) was used in taxol-induced assembly of tubulin in buffer at a final concentration of 30 pM.

Biochemistry, Vol. 33, No. 39, 1994

11885

Nucleotides. GTPaS(A) and GTPaS(B) were synthesized asdescribed for ATP& (Eckstein &Good, 1976). Guanosine 5'-O-(thiophosphate) (GMPS) was the starting material and was chemically synthesized from guanosine (Fisher) as previously described (Connolly et al., 1982). Tributylammonium pyrophosphate was purchased from Sigma. The isomers were separated by DEAE-Sephadex A-25 column chromatography (100 X 2.5 cm). HPLC on an anionexchange Nucleosil 10 SB column from Mackry-Nagel was used to determine purity as previously described (Connolly et al., 1982). B isomer was not detected in the A isomer preparation which was 98% pure. However, the B isomer preparation with 95-99% purity contained up to 2% A isomer (depending on the preparation). Two methods were used to further purify GTPaS(B). The first was rechromatography through the DEAE-Sephadex A-25 column. No GTPaS(A) was detected by HPLC which was capable of detecting 0.5% GTPaS(A). Since we found tubulin binds GTPaS(A) tightly but GTPaS(B) weakly at 4 OC, a mixture of 0.4 mM nucleotide GTPaS(B) (98%) and GTPaS(A) (2%) was incubated with tubulin (7.4 mg/mL) in buffer A (3 mM Mg2+) at 4 OC for 30 min. Free nucleotide was separated from tubulin by Sephadex G-25 chromatography. HPLC showed no GTPaS(A) (