PP1 Regulatory or Targeting Subunits

Protein phosphatase 1 enzymes acquire specificity of function by their association with targeting/regulatory pro teins that direct the enzyme to distinct subcellular structures or compartments in proximity to physiological substrates, confer substrate specificity and/or modulate enzyme activity. Over 40 PPl-associated proteins are currently known. Most interact with PPlc through multiple sites, a shared site recognizable in many of the subunits, and other sites unique to the individual proteins. The common docking site on PP1c is formed by a hydrophobic channel situated opposite the catalytic site and is flanked by an acidic region. This domain accommodates any of the variant forms of the binding motif [R/K] [V/I/L]X[F/W/Y] (often given as RVXF) found in a large number of PPlc-bound proteins. However, the presence of the tetrapetide does not necessarily define a PPlc-binding protein, as this sequence is found in more than

10% of proteins. Furthermore, its presence is not an absolute requirement for association with PPlc. The initial proposal of a mutually exclusive association of polypeptides harboring this motif with PPlc has been weakened by the recent findings that at last two polypeptides containing the motif can simultaneously associate with PPlc [7], supporting the notion that additional sites participate in the binding.

Based on analysis of eukaryotic genome sequences, Ceulemans and coworkers [5] have traced the evolution of l3 families of PPl regulatory/targeting proteins and have suggested the existence of nine additional isoforms not previously recognized. Table l presents a classification of over 40 PPlc-binding proteins, and Fig. l shows a schematic diagram of the structure of representative members of the different groups. Some of the binding subunits function as inhibitors/modulators of activity and do not contain domains for targeting to specific locations. Others function as targeting subunits to direct the phosphatase to specific subcellular structures or substrates and have no known regulatory role. Yet other PPlc-binding proteins may perform both a targeting and a regulatory function.

PPl Inhibitors or Modulators

The proteins in this group inhibit or modulate the activity of PPl but do not contain targeting domains. In fact, PPl was originally defined by its sensitivity to heat-stable protein inhibitor l (I-l) and 2 (I-2). I-l and its brain homolog DARPP-32 (dopamine- and cAMP-regulated phosphopro-tein of apparent Mr 32,000), I-2 , inhibitor-3/HCGV (I-3), and the G substrate inhibit the free PPlc, whereas the PKC-phosphorylated inhibitor protein CPI-l7 and its homologs PHI and KEPI are able to inhibit holoenzymes containing targeting subunits such as the glycogen- and myosin-associated phosphatases [8,9]. Furthermore, I-l, DARPP-32, I-2, and I-3 all contain a variant of the PPlc-binding consensus RVXF motif (Fig. l). The 8KIQFl2 sequence (homologous to RVXF) in I-l and DARPP-32 is essential for inhibition, whereas the equivalent sequence l44KLHYl47 in I-2 is dispensable [l0]. However, the N-terminal l2IKGIl5 residues in I-2 are required for inhibitory activity. This sequence occupies a unique site on PPlc, located adjacent to the hydrophobic groove [ll]. Three additional PPlc-interacting sites have been identified in I-2 (Fig. l) [l0], establishing the paradigm that high-affinity binding may be achieved by multiple contacts. The activity of the majority of the inhibitory/modulatory sub-units is controlled by phosphorylation. I-l and DARPP-32 become potent inhibitors after phosphorylation by the cAMP-dependent protein kinase (PKA) at a conserved threo-nine residue, whereas phosphorylation by cyclin-dependent kinase 5 (Cdk5) prevents phosphorylation by PKA, rendering I-l and DARPP-32 less effective inhibitors. The inhibitory activity of CPI-l7, PHI, and G-substrate is also enhanced by phosphorylation. I-2 does not require phosphorylation and is a complex modulator of PPl activity. Its stable interaction with PPlc at five distinct sites forms the inactive ATP-Mg2+-dependent holoenzyme that is activated by phosphorylation at T72 by glycogen synthase kinase 3 (GSK3), mitogen-activated protein kinase (MAPK), or Cdc2. Reactivation does not cause dissociation, arguing against a proposed chaperone role for I-2. The importance of these inhibitor proteins in the control of the phosphatase activity is highlighted by the phe-notype of the DARPP-32 and I-l knockout mice. DARPP-32 disruption impairs dopamine signaling, and the animals show decreased learning and reduced responses to substances of abuse [l2]. I-l knockout mice lack long-term potentiation at the perforant path-dentate cell synapses and have an impaired cardiac P-adrenergic response that is less severe than that caused by the over expression of PPlc [l3].

Glycogen Targeting Subunits

Four glycogen-targeting subunits have been characterized and three more putative forms, encoded by PPP1R3E, F, and G, have been identified in the human genome based on homology to PPlc-binding and targeting motifs [5]. Whether or not they represent bona fide PPl glycogen-targeting components remains to be determined. RGL, also called Gm, was the first glycogen-binding subunit of PPlc identified and is exclusively expressed in striated muscle [l4,l5]. The N-terminal region contains binding sites for PPlc, glycogen, and possibly glycogen synthase (GS), whereas a hydrophobic region in the C-terminus anchors the protein to membranes. Interaction with PPlc most likely involves multiple contacts, one of which is the 65RVSF68 sequence.

It has been proposed that muscle PPlG/RGL plays a major role in insulin and epinephrine control of glycogen metabolism via phosphorylation of the targeting subunit [l]. Insulin would cause phosphorylation of RGL at S48 and activation toward glycogen synthase, whereas epinephrine would

Table I Mammalian Regulatory/Targeting Subunits of Protein Phosphatase 1

Regulatory/targeting subunit

Gene namea

Established or putative cellular function controlled and/or effect on PP1c

Inhibitors/modulators

Inhibitor-1

DARPP-32

Inhibitor-2

Inhibitor-3

CPI-17

KEPI

G substrate Glycogen targeting

PPP1R1A

PPP1R1B

PPP1R2

PPP1R11

PPP1R14A

PPPP1R14B

PTG/R5 R6

Myosin targeting Mypt1/M110 Mypt2 p85

Nuclear targeting NIPP1 Sds22 PNUTS AKAP350 Nck2 PSF1 HCF

Membrane/cytoskeleton targeting Neurabin I

Neurabin II/spinophilin Neurofilament-L AKAP220 Yotiao AKAP149

Endoplasmic reticulum/ribosome targeting GADD34 GRP78 L5

RIPP1

Others Rb

53BP2 Hox11 Bcl2 PFK

Ryanodine receptor BH-protocadherin-c NKCC1 PRIP-1 PP1bp80 TIMAP Mypt3

I pp2a I pp2a

II I2

PPP1R3A PPP1R3B PPP1R3C PPP1R3D

PPP1R12A PPP1R12B PPP1R12C

PPP1R8 PPP1R7 PPP1R10

PPP1R9A PPP1R9B

PPP1R15A

PPP1R13A

PPP1R16B PPP1R16A

PKA-mediated PP1c inhibition Neurotransmitter signaling Modulation of PP1c Inhibition of PP1c Smooth muscle contraction Inhibition of PP1 holoenzymes Morphine signaling Inhibition of PP1c

Muscle glycogen metabolism Liver glycogen metabolism Glycogen metabolism Glycogen metabolism

Smooth muscle contraction, cell Shape and motility Skeletal muscle contraction Actin cytoskeleton

Pre-mRNA splicing Exit from mitosis Transcription/RNA maturation Centrosome and Golgi function Centrosome separation Pre-mRNA spicing Transcription

Neurite outgrowth Dendritic spine formation Neuronal morphology Peroxisome/cytoskeletal activities NMDA/ion channel activity Nucleus reassembly

Stress responses Chaperone/protein folding Protein synthesis Protein synthesis

Cell cycle

Cell cycle/apoptosis

Cell cycle

Apoptosis

Glycolysis

Calcium channel activity Cell adhesion Ion transport Ins(1,4,5)P3 signaling Unknown TGFß signaling TGFß signaling

Stimulation of PP1c/substrate specificity

RGL/GM

aThe human genome nomenclature recently redefined by Ceulemans et al. [5] for the PP1c-binding proteins is indicated.

Figure 1 Domain structure of representative members of PP1c-binding proteins. Phosphorylation sites and various domains are indicated. The filled black box indicates the common PPlc-binding motif. Targeting domains are indicated by various patterns and abbreviations: FHA, forkhead associated; PB, polybasic; RB, RNA-binding; GT, glycogen targeting; S-B?, putative substrate-binding; TM, transmembrane; ANK, ankyrin repeats; MYT, myosin targeting; M20B, M20-binding; F-AB, F-actin binding; CC, coiled-coil; SAM, sterile alpha motif; TFIIS, transcriptional factor IIS; ZF, zinc finger.

Figure 1 Domain structure of representative members of PP1c-binding proteins. Phosphorylation sites and various domains are indicated. The filled black box indicates the common PPlc-binding motif. Targeting domains are indicated by various patterns and abbreviations: FHA, forkhead associated; PB, polybasic; RB, RNA-binding; GT, glycogen targeting; S-B?, putative substrate-binding; TM, transmembrane; ANK, ankyrin repeats; MYT, myosin targeting; M20B, M20-binding; F-AB, F-actin binding; CC, coiled-coil; SAM, sterile alpha motif; TFIIS, transcriptional factor IIS; ZF, zinc finger.

induce phosphorylation at S67, mediated by PKA, and cause dissociation of PPlc. The released PPlc would be less active toward glycogen-associated and perhaps membrane-bound substrates. Furthermore, activation of PKA would lead to phosphorylation of I-l, which then becomes a potent inhibitor of the released PPlc. However, recent studies have shown that Rgl is not phosphorylated at S48 in response to insulin [l6,l7], and disruption of the RGL and I-l genes has shown that neither is required for either insulin or epinephrine control of glycogen metabolism [l7-l9]. RGL, though, is essential for control of GS by exercise and muscle contraction [20]. Direct phosphorylation of GS by GSK-3 and PKA, the protein kinases regulated by insulin and epinephrine, respectively cannot account for the effects on glycogen metabolism, as changes in phosphorylation at the GS sites recognized by these individual kinases are insufficient to account for the alterations of activity caused by the two hormones.

Of the other three glycogen-targeting subunits, GL was thought to be liver specific, but a recent report describes GL in human, but not in rodent, skeletal muscle [2l]. PTG expression is higher in skeletal muscle, liver, heart, and fat, whereas R6 is more ubiquitously expressed. GL, PTG, and R6 share homology to the N-terminal region of RGL and lack the extended C-terminal tail and the membrane-binding domain [22,23]. All three contain the PPlc-binding motif and the glycogen-binding domain, but the two phosphory-lation sites of RGL are not conserved. PTG interacts with glycogen metabolizing enzymes but, unlike the liver-specific PPlG/GL complex, PPlG/PTG is not controlled allosterically by phosphorylase a. Expression of the GL subunit is downregulated in diabetic rats. Overexpression of PTG in Chinese hamster ovary cells increases the basal and insulin-stimulated GS activity, but neither insulin nor forskolin induce detectable PTG phosphorylation [23]. A mechanism has been proposed whereby PTG would affect PPl activity by relieving inhibition by DARPP-32. However, neither I-l nor DARPP-32 is required for insulin activation of glycogen synthase [l9]. Thus, the mechanisms for control of PTG- and R6-containing phosphatases are largely unknown. Homozygous disruption of PTG results in embryonic lethality. Heterozygous PTG knockout mice retain activation of GS by insulin in skeletal muscles but appear to develop impaired glucose disposal with aging [24].

Myosin Targeting Subunits

Three subunits target PPlc to myosin [25]. The best characterized, Myptl/MllO, interacts with myosin type II and is involved in control of smooth muscle contraction, cell shape, and migration. The myosin phosphatase complex is a heterotrimer containing also a smaller polypeptide, M20 (~ 20 kD), that does not bind to PPlc but interacts with Myptl and myosin. In addition to a targeting function, Myptl confers substrate specificity, enhancing phosphatase activity toward the regulatory myosin light chains while decreasing it toward phosphorylase. Two other actin-binding proteins, adducin and moesin, are also associated with and regulated by Myptl in non-muscle cells, supporting a role in actin cytoskeleton organization. The RVXF motif in the N-terminal region of Myptl is followed by seven or eight ankyrin repeats that may be involved in interaction with PPlc and/or myosin. The C-terminus harbors domains that bind to myosin and M20. Phosphorylation by a RhoA-activated kinase (ROCK), ZIP-like kinase, or the myotonic dystrophy protein kinse (DMPK) at T697 inhibits phosphatase activity, leading to increased light-chain phosphorylation and contraction in smooth muscle in the absence of changes in intracellular calcium levels. In contrast, phosphorylation at S435 during mitosis increases myosin light-chain phosphatase activity. Another feature of this phosphatase is that it is potently and specifically inhibited by CPI-17 [25], providing an additional mechanism for enhancement of myosin phosphorylation and smooth-muscle contraction. CPI-17 does not contain an RVXF motif. Phosphorylation by PKC causes a conformational change that appears to expose a region that may interact with PPlc.

Of the other two members of the family, Mypt2 is expressed in skeletal muscle, heart, and brain, whereas p85 is more widely distributed and appears to be required for assembly of the actin cytoskeleton. Both share structural similarity with Myptl, including the PPlc-binding motif and the N-terminal ankyrin repeats. Interestingly, the small M20 subunit appears to be generated by alternative splicing of PPP1R12B, the gene that codes for Mypt2. A newly identified protein was named Mypt3, due to the presence of ankyrin repeats in addition to the RVXF motif [26]. However, based on the absence of a myosin-binding domain and on gene structural similarity, Ceulemans and coworkers [5] have reclassified it as a member of the TIMAP family, which may be involved in TGFP signaling (Table l).

Nuclear Targeting Subunits

Protein phosphatase l is abundantly expressed in the nucleus, where it is complexed with a variety of regulatory subunits to control such processes as cell-cycle progression and division, transcription, and pre-mRNA splicing. Sds22, a protein required for exit from mitosis, and HCF-l (human factor Cl) have no discernable RVXF motif [27]. Disruption of the PPlc hydrophobic docking channel does not impair Sds22 binding, indicating that interaction involves other sites. Multiple contacts are also established between PPlc and NIPPl, nuclear inhibitor of PPl (Fig. l) [28]. Binding of the RVXF motif is not inhibitory by itself. A polybasic sequence preceding the common motif as well as a C-terminal region are required for high potency inhibition. Similar to I-2, phosphorylation is not required for inhibitory activity, and the action of two protein kinases, PKA and casein kinase II, relieves the inhibition without causing dissociation of the NIPPl/PPlc complex. NIPPl is localized to nuclear "speckles" where it interacts with splicing factors through its N-terminal forkhead-associated domain. The C-terminus binds RNA, supporting a role of NIPPl in pre-mRNA splicing [29]. The splicing-factor-associated protein PSFl also binds PPlc, perhaps allowing control of SFl, which inhibits early spliceosome formation once phosphorylated. Two proteins, AKAP350 and Nek2 (NIMA-related protein kinase 2), may localize PPlc to the centrosome [30,3l]. Nek2 has been implicated in centrosome separation and, together with its substrate C-Napl, is a PPl substrate. PNUTS/p99 is another putative nuclear targeting subunit of PPlc [32]. An N-terminal sequence related to domains present in other transcriptional factors, TFIIS, elongin A, and CRSP70, and the presence of a zinc finger motif in the C-terminus may account for its chromatin association and a potential role in transcription.

Membrane or Cytoskeleton Targeting Subunits

A number of proteins associated with membrane and cytoskeletal structures have been shown to bind PPlc. Neurabin I and neurabin II (also known as spinophilin) are actin cross-linking proteins enriched in postsynaptic densities and dendritic spines [33,34]. Recent studies have shown that both neurabins and neuofilment-L display binding selectivity for the PPlc a and yl isoforms [6]. Neurabins contain an N-terminal F-actin-associating domain that accounts for localization at the actin cytoskeleton, C-terminal coiled-coil and SAM (sterile alpha motif) modules that mediate homo-and heterodimerization, and a central PDZ protein-binding domain that may link PPlc to transmembrane proteins (Fig. l). Phosphorylation of neurabin I by PKA at S46l, immediately C-terminal to the RVXF sequence, reduces PPlc-binding, and a S46lE mutation decreases inhibitory potency, suggesting regulation by cAMP signaling. Spinophilin knockout mice have provided evidence for a role in a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) and N-methyl-D-aspartate (NMDA) receptor channel activity and in dendritic spine development [35].

A defining feature of the AKAPs is the localization of PKA to different intracellular structures, including cytoskele-ton, mitochondria, nucleus, membrane, and vesicles [2]. Several AKAPs also interact with protein phosphatases, resulting in colocalization of enzymes that potentially exert opposite effects [30]. The AKAP Yotiao, an NMDA-receptor-associated protein, binds PPlc. Although Yotiao contains the RVXF motif, this sequence is not essential for interaction. The tethered PPlc is active and may negatively regulate receptor activity. Recently, it has been shown that Yotiao, complexed with PKA and PPlc, also associates with cardiac potassium channels. Mutations that disrupt the interaction correlate with inherited cardiac arrhythmias [36]. AKAPl49, an integral protein of the endoplasmic reticulum/nuclear envelop network, recruits PPlc to the lamina of the nuclear membrane where it may function to dephosphorylate B-type lamins for reassembly of the nuclear envelop at the end of mitosis. AKAP220 binds to and inhibits PPlc through multiple contacts and recruits the phosphatase to vesicles.

Endoplasmic Reticulum/Ribosome Targeting Subunits

Biochemical and genetic studies have implicated PP1 in the control of protein synthesis. Several components of the translational machinery are controlled by phosphorylation. In response to various stress stimuli, such as ultraviolet irradiation, viral infection, and nutrient deprivation, the eIF2a subunit becomes phosphorylated and translation initiation is inhibited. GADD34, a growth arrest and DNA-damage-induced gene, has been implicated in growth arrest and apoptosis induced by endoplasmic reticulum (ER) stress signals and has been shown to be associated with reticular structures. In response to stress signals, protein synthesis is shut off through phosphorylation of eIF2a. GADD34 forms a complex with PP1c that specifically promotes the dephos-phorylation of eIF2a. The stress-dependent expression of GADD34 implies that it may provide for a negative feedback mechanism to evade or promote recovery from the transla-tional inhibition. GADD34 interacts with PP1c through its C-terminus, a region homologous to the herpes virus ICP34.5 protein domain that also interacts with PP1c and that redirects the phosphatase to dephosphorylate eIF-2, enabling continued protein synthesis in virally infected cells. Interestingly, I-1 binds both GADD34 and PP1c via different domains [7]. The C-terminus of I-1 is essential for interaction with a central region of GADD34, whereas the N-terminus binds PP1c, raising the possibility for formation of a heterotrimeric complex containing two PP1 regulatory components, each of which harbors a canonical docking site. Similarly, hSNF5/ INI1, a member of the hSWI/SNF chromatin remodeling complex, binds to free GADD34 and PP1c as well as to the GADD34/PP1c complex to form a stable heterotrimer [37]. These findings indicate that association of different regulatory subunits with PP1c may not necessarily be mutually exclusive even if each contains an RVXF sequence. Another ER protein that binds PP1c is the glucose-regulated protein chaperone GRP78 [38], which is involved in protein translocation and folding and is induced by ER stress. Although the role of this phosphatase in the ER is not clearly understood, it may be part of a general mechanism to overcome the inhibition of translation in response to cellular stress conditions.

Other PP1c-binding proteins that may be involved in control of translation are the large ribosomal protein L5 and RP111. L5 [39] is located at the interphase between the small and large ribosomal subunits where translation initiation takes place and is therefore positioned for potential control of this step. In addition, the phosphorylated S6 (small ribosomal subunit) protein promotes the preferential recruitment of polypirimidine-track-containing mRNAs. The phosphatase that dephosphorylates S6 is a type 1 enzyme.

Other PP1c-Binding Proteins

Although much has been learned about localization and function of PP1, the precise roles of most PP1c-binding proteins are not completely understood. Not all physically target the enzyme to subcellular compartments. Some of the reported binding proteins may simply be substrates. Included would be the retinoblastoma protein pRb, phos-phofructokinase (PFK), the ryanodine receptor, and the Na-K-Cl co-transporter NKCC1, all of which are controlled by phosphorylation and some of which do not contain any recognizable PP1c docking site. The 53BP2 [40] interacts with p53 and Bcl2 and may specifically direct the phos-phatase activity toward proteins whose phosphorylation state is critical for the control of apoptosis. The ability of PRIP-1 to bind inosito1,4,5-trisphosphate [41] may allow recruitment of PP1c to membranes in response to stimuli that generate the phospholipid. The two heat-stable inhibitors of PP2A, I1PP2A and I2PP2A have recently been shown to enhance in vitro PP1c activity toward specific substrates [42]. These polypeptides also do not contain a recognizable PP1c-binding site. However, whether or not they can function as activators of PP1 in vivo remains to be determined.

Delicious Diabetic Recipes

Delicious Diabetic Recipes

This brilliant guide will teach you how to cook all those delicious recipes for people who have diabetes.

Get My Free Ebook


Post a comment