Journal of Muscle Research and Cell Motility 13, 493-496 (1992) REVIEW
Dynamin: motor protein or regulatory GTPase RICHARD
B. V A L L E E
Cell Biology Group, Worcester Foundation for Experimental Biology, Shrewsbury, M A 01545, USA Received 25 April 1992; accepted 5 May 1992
Introduction
Work in recent years has led to the identification of two new 'motor' proteins, kinesin and cytoplasmic dynein, responsible for force-production in opposite directions along cytoplasmic microtubules (reviewed in Vallee & Shpetner, 1990). The discovery of dynamin, which exhibited a number of properties reminiscent of kinesin and cytoplasmic dynein (Shpetner & Vallee, 1989), seemed to herald the existence of a new class of motor molecule. However, while the well established microtubule-associated and actin-associated motor proteins are all ATPases, dynamin is a GTPase. Is it, therefore, the first example of a GTP-driven mechanoenzyme or does its preference for GTP signal a regulatory role, as is thought to be the case for many of the small GTPases which have been identified in recent years? Biochemical and molecular evidence
A number of lines of evidence support the first possibility. The interaction of dynamin with microtubules is nucleotide-sensitive (Shpetner & Valtee, 1989), consistent with a force-producing crossbridge cycle. Both GTP and ATP were effective in dissociating dynamin from microtubules, as is the case for kinesin. When assayed for ATPase activity, the purified protein was inactive; however, together with an 'activating fraction' normally removed during purification, a microtubule-activated ATPase activity was observed. Dynamin induced the formation of microtubule bundles. Addition of ATP and the activating fraction caused the bundles to come apart, with some showing evidence of elongation. A simple interpretation for this behaviour was that dynamin mediated sliding between the crosslinked microtubules. The closely packed, highly ordered microtubules in the bundles were reminiscent of the midzone of the mitotic spindle, suggesting a possible role for dynamin in spindle elongation. Molecular cloning of dynamin (Obar et a]., I990) indicated, however, a profound difference in its primary 0142-4319 9
1992 Chapman & Hall
sequence from that of the motor proteins. In addition to the G(X4)GKS/T consensus element for binding of the gamma-phosphate of nucleoside triphosphates, it contained the two other well-conserved consensus elements found in almost all GTP-binding proteins (Fig. I). Moreover, with the exception of a small region of homology with kinesin surrounding the second GTP-binding consensus sequence element, no relationship was observed between dynamin and kinesin, myosin or, subsequently, axonemal dynein (Gibbons et al., 1991; Ogawa, 1991). Dynamin did exhibit striking homology with two other classes of protein: the interferon-inducible anti-viral factors known as Mx (Staeheli et aI., 1986), and the product of the yeast VPS1 gene, which is involved in sorting of proteins from the Golgi apparatus to the vacuole (Rothman et al., 1990). A fourth related protein has been identified recently, the product of the yeast MGM-1 gene, which is involved in mitochodrial genome replication (Jones & Fangman, 1992). In all cases, the dearest evidence of sequence conservation was observed within an approximately 300 amino acid region at or near the amino terminus, although additional sequence homology was observed between dynamin and the product of the VPS1 gene throughout their length. These findings have clearly identified a novel, high molecular weight class of putative GTP-binding proteins, although a unifying functional theme beyond GTP-binding has been elusive. The mechanism of action of Mx and MGM-1 remains obscure. However, VPS1 is sufficiently well understood to offer some insight into its own mode of action as well as that of the other family members. Mutations in VPS1 block the delivery of proteases, such as carboxypeptidase Y, to the yeast vacuole (Rothman et al., 1990), a degradative organelle functionally comparable to the lysosome. This could reflect a defect in microtubule-associated organelle transport. The vacuole had been reported to become disrupted in the presence of nocodazole (Guthrie & Wickner, 1988), suggesting an association with microtubules. Nevertheless, a direct influence of nocodazole on the rate or efficiency of sorting of proteins to the vacuole was not observed
VALLEE
494 GTP Binding
MT Binding
Dynamin VPS1/SP015 Mx MGM1
9~ - V a r . ~
Dynamin: motor protein or regulatory GTPase RICHARD
B. V A L L E E
Cell Biology Group, Worcester Foundation for Experimental Biology, Shrewsbury, M A 01545, USA Received 25 April 1992; accepted 5 May 1992
Introduction
Work in recent years has led to the identification of two new 'motor' proteins, kinesin and cytoplasmic dynein, responsible for force-production in opposite directions along cytoplasmic microtubules (reviewed in Vallee & Shpetner, 1990). The discovery of dynamin, which exhibited a number of properties reminiscent of kinesin and cytoplasmic dynein (Shpetner & Vallee, 1989), seemed to herald the existence of a new class of motor molecule. However, while the well established microtubule-associated and actin-associated motor proteins are all ATPases, dynamin is a GTPase. Is it, therefore, the first example of a GTP-driven mechanoenzyme or does its preference for GTP signal a regulatory role, as is thought to be the case for many of the small GTPases which have been identified in recent years? Biochemical and molecular evidence
A number of lines of evidence support the first possibility. The interaction of dynamin with microtubules is nucleotide-sensitive (Shpetner & Valtee, 1989), consistent with a force-producing crossbridge cycle. Both GTP and ATP were effective in dissociating dynamin from microtubules, as is the case for kinesin. When assayed for ATPase activity, the purified protein was inactive; however, together with an 'activating fraction' normally removed during purification, a microtubule-activated ATPase activity was observed. Dynamin induced the formation of microtubule bundles. Addition of ATP and the activating fraction caused the bundles to come apart, with some showing evidence of elongation. A simple interpretation for this behaviour was that dynamin mediated sliding between the crosslinked microtubules. The closely packed, highly ordered microtubules in the bundles were reminiscent of the midzone of the mitotic spindle, suggesting a possible role for dynamin in spindle elongation. Molecular cloning of dynamin (Obar et a]., I990) indicated, however, a profound difference in its primary 0142-4319 9
1992 Chapman & Hall
sequence from that of the motor proteins. In addition to the G(X4)GKS/T consensus element for binding of the gamma-phosphate of nucleoside triphosphates, it contained the two other well-conserved consensus elements found in almost all GTP-binding proteins (Fig. I). Moreover, with the exception of a small region of homology with kinesin surrounding the second GTP-binding consensus sequence element, no relationship was observed between dynamin and kinesin, myosin or, subsequently, axonemal dynein (Gibbons et al., 1991; Ogawa, 1991). Dynamin did exhibit striking homology with two other classes of protein: the interferon-inducible anti-viral factors known as Mx (Staeheli et aI., 1986), and the product of the yeast VPS1 gene, which is involved in sorting of proteins from the Golgi apparatus to the vacuole (Rothman et al., 1990). A fourth related protein has been identified recently, the product of the yeast MGM-1 gene, which is involved in mitochodrial genome replication (Jones & Fangman, 1992). In all cases, the dearest evidence of sequence conservation was observed within an approximately 300 amino acid region at or near the amino terminus, although additional sequence homology was observed between dynamin and the product of the VPS1 gene throughout their length. These findings have clearly identified a novel, high molecular weight class of putative GTP-binding proteins, although a unifying functional theme beyond GTP-binding has been elusive. The mechanism of action of Mx and MGM-1 remains obscure. However, VPS1 is sufficiently well understood to offer some insight into its own mode of action as well as that of the other family members. Mutations in VPS1 block the delivery of proteases, such as carboxypeptidase Y, to the yeast vacuole (Rothman et al., 1990), a degradative organelle functionally comparable to the lysosome. This could reflect a defect in microtubule-associated organelle transport. The vacuole had been reported to become disrupted in the presence of nocodazole (Guthrie & Wickner, 1988), suggesting an association with microtubules. Nevertheless, a direct influence of nocodazole on the rate or efficiency of sorting of proteins to the vacuole was not observed
VALLEE
494 GTP Binding
MT Binding
Dynamin VPS1/SP015 Mx MGM1
9~ - V a r . ~
