The Salt-Inducible Kinase, SIK, Is Induced by Depolarization in Brain

Authors

  • Jonathan D. Feldman,

    1. Department of Pediatrics, University of Southern California, Los Angeles, California, U.S.A.*Department of Biological Chemistry, University of Southern California, Los Angeles, California, U.S.A.Department of Molecular and Medical Pharmacology, University of Southern California, Los Angeles, California, U.S.A.Molecular Biology Institute, UCLA Center for the Health Sciences, University of Southern California, Los Angeles, California, U.S.A. § Neuroscience Program, University of Southern California, Los Angeles, California, U.S.A.Department of General and Environmental Physiology, University of Naples “Federico II,” Naples, Italy
    Search for more papers by this author
  • Linda Vician,

    1. Department of Pediatrics, University of Southern California, Los Angeles, California, U.S.A.*Department of Biological Chemistry, University of Southern California, Los Angeles, California, U.S.A.Department of Molecular and Medical Pharmacology, University of Southern California, Los Angeles, California, U.S.A.Molecular Biology Institute, UCLA Center for the Health Sciences, University of Southern California, Los Angeles, California, U.S.A. § Neuroscience Program, University of Southern California, Los Angeles, California, U.S.A.Department of General and Environmental Physiology, University of Naples “Federico II,” Naples, Italy
    Search for more papers by this author
  • * Marianna Crispino,

    1. Department of Pediatrics, University of Southern California, Los Angeles, California, U.S.A.*Department of Biological Chemistry, University of Southern California, Los Angeles, California, U.S.A.Department of Molecular and Medical Pharmacology, University of Southern California, Los Angeles, California, U.S.A.Molecular Biology Institute, UCLA Center for the Health Sciences, University of Southern California, Los Angeles, California, U.S.A. § Neuroscience Program, University of Southern California, Los Angeles, California, U.S.A.Department of General and Environmental Physiology, University of Naples “Federico II,” Naples, Italy
    Search for more papers by this author
  • Warren Hoe,

    1. Department of Pediatrics, University of Southern California, Los Angeles, California, U.S.A.*Department of Biological Chemistry, University of Southern California, Los Angeles, California, U.S.A.Department of Molecular and Medical Pharmacology, University of Southern California, Los Angeles, California, U.S.A.Molecular Biology Institute, UCLA Center for the Health Sciences, University of Southern California, Los Angeles, California, U.S.A. § Neuroscience Program, University of Southern California, Los Angeles, California, U.S.A.Department of General and Environmental Physiology, University of Naples “Federico II,” Naples, Italy
    Search for more papers by this author
  • Michel Baudry,

    1. Department of Pediatrics, University of Southern California, Los Angeles, California, U.S.A.*Department of Biological Chemistry, University of Southern California, Los Angeles, California, U.S.A.Department of Molecular and Medical Pharmacology, University of Southern California, Los Angeles, California, U.S.A.Molecular Biology Institute, UCLA Center for the Health Sciences, University of Southern California, Los Angeles, California, U.S.A. § Neuroscience Program, University of Southern California, Los Angeles, California, U.S.A.Department of General and Environmental Physiology, University of Naples “Federico II,” Naples, Italy
    Search for more papers by this author
  • and * Harvey R. Herschman

    1. Department of Pediatrics, University of Southern California, Los Angeles, California, U.S.A.*Department of Biological Chemistry, University of Southern California, Los Angeles, California, U.S.A.Department of Molecular and Medical Pharmacology, University of Southern California, Los Angeles, California, U.S.A.Molecular Biology Institute, UCLA Center for the Health Sciences, University of Southern California, Los Angeles, California, U.S.A. § Neuroscience Program, University of Southern California, Los Angeles, California, U.S.A.Department of General and Environmental Physiology, University of Naples “Federico II,” Naples, Italy
    Search for more papers by this author

  • Lippincott Williams & Wilkins, Inc., Philadelphia

  • Abbreviations used: AMPK, AMP-activated protein kinase; CHX, cycloheximide; CPG, candidate plasticity-related gene; CREB, cyclic AMP-responsive element-binding protein; EGF, epidermal growth factor; GST, glutathione S-transferase; IEG, immediate-early gene; IPTG, isopropyl 1-thio-β-D-galactoside; KA, kainic acid; KID-2, kinase induced by depolarization-2; msk, myocardial sucrose-nonfermenting 1 protein kinase-like kinase; NGF, nerve growth factor; PRMT3, protein-arginine N-methyltransferase 3; RDA, representational difference analysis; SIK, salt-inducible kinase; SIKKD, salt-inducible kinase kinase domain; SNF1, sucrose-nonfermenting 1 protein kinase. The nucleotide sequences discussed in this article are deposited in the GenBank database under the following GenBank accession numbers: KID-2, AF106937 (present study); SIK, AB020480 (Wang et al., 1999); and msk, U11494 (Ruiz et al., 1994).

Address correspondence and reprint requests to Dr. H. R. Herschman at Molecular Biology Institute, UCLA Center for the Health Sciences, 611 Charles E. Young Drive East, Los Angeles, CA 90095-1570, U.S.A. E-mail: hherschman@mednet.ucla.edu

Abstract

Abstract: Membrane depolarization of neurons is thought to lead to changes in gene expression that modulate neuronal plasticity. We used representational difference analysis to identify a group of cDNAs that are induced by membrane depolarization or by forskolin, but not by neurotrophins or growth factors, in PC12 pheochromocytoma cells. One of these genes, SIK (salt-inducible kinase), is a member of the sucrose-nonfermenting 1 protein kinase/AMP-activated protein kinase protein kinase family that was also recently identified from the adrenal gland of rats treated with high-salt diets. SIK mRNA is induced up to eightfold in specific regions of the hippocampus and cortex in rats, following systemic kainic acid administration and seizure induction.

Membrane depolarization activates signal transduction pathways in neurons and leads to induction of gene expression (Arenander and Herschman, 1993; Bliss and Collingridge, 1993; Ghosh and Greenberg, 1995). These signaling mechanisms and the transcription of previously quiescent genes are thought to play important roles in the synaptic plasticity necessary for brain development (Murphy and Segal, 1997), learning and memory (Ghosh and Greenberg, 1995; Wu et al., 1995; Mayford et al., 1996; Tsien et al., 1996), and response to brain injury (Bazan et al., 1995; Katano et al., 1998). Genes induced in neurons by depolarization are therefore candidates for mediators of neuronal plasticity. The identification of genes induced preferentially by depolarization in neurons has been a major focus of work from our laboratory (Vician et al., 1995; Feldman et al., 1998a,b, 2000) and from others (Nedivi et al., 1993; Qian et al., 1993; Yamagata et al., 1993, 1994; Okabe et al., 1996; Babity et al., 1997; Kato et al., 1997; Gorecki et al., 1998; Hevroni et al., 1998).

PC12 cells are a “neuron-like” clonal cell population derived from a rat pheochromocytoma (Greene and Tischler, 1976). In response to depolarization, synaptic vesicles present in PC12 cells fuse with plasma membranes and release stored neurotransmitters in a calcium-dependent manner (Melega and Howard, 1984; Bauerfeind et al., 1993). PC12 cells are therefore used to study many depolarization-dependent neuronal properties (Melega and Howard, 1984; Kujubu et al., 1987; Milbrandt, 1988; Bartel et al., 1989; Altin et al., 1991; Bradbury et al., 1991; Bauerfeind et al., 1993; Vician et al., 1995).

We use differential screening procedures to isolate, from PC12 cells, cDNAs representing genes preferentially induced by depolarization. These experiments first identified synaptotagmin IV (syt IV) as a depolarization-induced, brain-specific gene (Vician et al., 1995). Using representational difference analysis (RDA), we subsequently identified KID-1, a protein kinase induced in PC12 cells by depolarization and forskolin but not by nerve growth factor (NGF) or epidermal growth factor (EGF). KID-1 expression is induced in the hippocampus and piriform cortex of rats in response to systemic kainic acid (KA) treatment (Feldman et al., 1998b). KID-1 is closely related to the PIM-1 family of protooncogenes. Because of this similarity, we examined PIM-1 expression and found, like KID-1, that (a) PIM-1 is induced in PC12 cells by forskolin but not by NGF or EGF and (b) PIM-1 expression is induced in rat hippocampus in response to KA administration (Feldman et al., 1998a). The induction of KID-1 and PIM-1 in the hippocampus by KA was recently confirmed (Konietzko et al., 1999).

In this report we describe the identification and characterization of a third depolarization-induced protein kinase. We originally named this gene KID-2 (kinase induced by depolarization-2). During the preparation of this article, the identical cDNA named SIK (salt-inducible kinase) (Wang et al., 1999) was reported as a gene induced in the adrenal gland of rats fed high-salt diets. We find that elevated potassium or forskolin treatment rapidly induces SIK expression in PC12 cells. However, neither NGF nor EGF induces SIK expression. Sequence comparisons identify rat SIK as the orthologue of murine msk [myocardial sucrose-nonfermenting 1 protein kinase (SNF1)-like kinase], a partial cDNA identified while characterizing protein kinase expression during murine cardiac development (Ruiz et al., 1994). The sequence of the entire SIK/msk open reading frame predicts that the SIK protein is a member of the SNF1/AMP-activated protein kinase (AMPK) family of serine/threonine protein kinases. A glutathione-S-transferase (GST) fusion protein of SIK catalyzes both autophosphorylation and serine phosphorylation of a synthetic peptide. However, unlike KID-1 and PIM-1, SIK does not phosphorylate histone H1. The SIK gene is an immediate-early gene (IEG); SIK mRNA accumulates in depolarized PC12 cells even in the presence of the protein synthesis inhibitor cycloheximide (CHX). Following seizure induction in rats by KA, SIK mRNA expression is strongly induced in specific areas of the hippocampus and cortex.

MATERIALS AND METHODS

Cell culture, cell treatment, and RNA purification

PC12 cells were cultured and treated as previously described (Feldman et al., 1998b). KCl (50 mM), forskolin (50 μM), NGF (50 ng/ml), or EGF (10 ng/ml) was added to the cultures for the times indicated in the text or figures. To determine if SIK is a primary response gene, CHX (10 μg/ml; Sigma, St. Louis, MO, U.S.A.) was added, either alone or along with the specific ligand, for 1.5 h. Poly(A)+ RNA was isolated using an Oligotex mRNA Kit (Qiagen, Chatsworth, CA, U.S.A.).

In an additional experiment to test the effects of various doses of KCl and forskolin on SIK induction, cells were treated with 20, 35, or 50 mM KCl and 10, 25, or 50 μM forskolin for 0.75 h. In all experiments where KCl was used to depolarize cells, KCl was substituted for NaCl in the culture medium so that total osmolarity was kept constant, to ensure that the effects on gene expression were due to membrane depolarization and not due to changes in osmolarity. As an additional control, some cells were treated with the calcium ionophore A23187 (10 μM) to test the effect of calcium influx, as compared with K+-induced membrane depolarization, on SIK expression.

cDNA synthesis, doping the driver, and RDA

Two micrograms each of the KCl and forskolin poly(A)+ RNAs were pooled to make “tester” cDNA. Similarly, 2 μg each of the NGF and EGF poly(A)+ RNAs were pooled to make “driver” cDNA. Double-stranded cDNA was synthesized from the pooled poly(A)+ RNAs, using a SuperScript Choice cDNA Synthesis kit (GibcoBRL, Gaithersburg, MD, U.S.A.), with a combination of 1.0 μg of oligo(dT)12-18 and 100 ng of random hexamers.

To prevent isolating known genes, northern blots prepared with “tester,”“driver,” and control poly(A)+ RNAs were probed with several genes already reported to be induced by depolarization (Feldman et al., 1998b). The hybridization signals were quantified, and the driver was “doped” with excess NGFI-B (provided by J. Milbrandt, Washington University, St. Louis), secretogranin I (provided by P. Danielson, Scripps Research Institute, La Jolla, CA, U.S.A.), and synaptotagmin IV (Vician et al., 1995) cDNAs as previously described (Feldman et al., 1998b).

Three rounds of RDA were performed as previously described (Vician et al., 1997; Feldman et al., 1998b). Tester: driver hybridization ratios were 1:100 for the first round, 1:1,000 for the second round, and 1:50,000 for the third round. Hybridizations were maintained for ∼60 h.

RNA preparation from tissues and northern analysis

Tissues were processed, and RNA was isolated as previously described (Feldman et al., 1998b). Northern blots were performed as previously described (Vician et al., 1995).

DNA sequencing and analysis

Sequencing was performed by the core facility of the Department of Pediatrics, University of California, Los Angeles, on an Applied Biosystems instrument. The DNA sequences obtained were compared with the current version of the nonredundant, updated GenBank and EMBL database using BlastN (Altschul et al., 1997). Analysis was performed using MacVector and AssemblyLign (Eastman Kodak, New Haven, CT, U.S.A.).

Protein structure and phylogenetic analysis

The amino acid sequence of SIK was analyzed for predicted coiled coil regions using the Multicoil program (Wolf et al., 1997). Phylogenetic analysis was performed using MegAlign software (DNAStar, Madison, WI, U.S.A.).

cDNA library screening

The full-length SIK cDNA clone was obtained by screening a 9-day-old rat cerebral cortex cDNA library constructed in λ ZAP (Stratagene, La Jolla) and provided by J. Boulter (University of California, Los Angeles). Plaque lifts were probed by standard methods with the cloned fragment isolated from the RDA procedure (Sambrook et al., 1989). Positive clones were converted into plasmid using Exassist helper phage (Stratagene).

Preparation and purification of GST-SIK and GST-SIK kinase domain (SIKKD) fusion proteins and western analysis

By performing PCR with the full-length SIK cDNA as template and the oligonucleotides 5′-CGCGGATCCCGGAATTCCATGGTGATCATGTCGGAGTTCA
GC-3′ and 5′-GCGGCCGCTCGAGTCGACTCACTGTACCAGGACGAACGTC
-3′ as primers, the SIK open reading frame was modified to create a 5′BamHI and a 3′SalI restriction enzyme site. The PCR product was digested with BamHI (New England Biolabs, Beverly, MA, U.S.A.) and SalI (New England Biolabs), and the 2,300-bp fragment containing the SIK open reading frame was isolated from an agarose gel. This fragment was ligated into the BamHI and XhoI sites of the bacterial expression vector pGEX-4T-3 (Pharmacia, Piscataway, NJ, U.S.A.). This insertion created an “in-frame” fusion of the SIK open reading frame cDNA with the isopropyl 1-thio-β-D-galactoside (IPTG)-inducible GST gene.

By performing PCR with the full-length SIK cDNA as template and the oligonucleotides 5′-CGCGGATCCCGGAATTCCATGGTGATCATGTCGGAGTTCA
GC-3′ and 5′-GCGGCCGCTCGAGTCGACTCACTGCTCGTTGTAGTCGCCC
-3′ as primers, the predicted SIKKD was modified to create a 5′BamHI and a 3′XhoI restriction enzyme site. The PCR product was digested with BamHI and XhoI (New England Biolabs), and the 960-bp fragment containing the SIKKD open reading frame was isolated from an agarose gel. This fragment was ligated into the BamHI and XhoI sites of pGEX-4T-3. This insertion created an “in-frame” fusion of the SIKKD open reading frame cDNA with the IPTG-inducible GST gene. Both the pGEX-4T-3+SIK and pGEX-4T-3+SIKKD plasmids were sequenced to confirm the correct reading frame and absence of PCR-generated errors.

The JM109 strain of Escherichia coli was transformed with the GST-SIK and GST-SIKKD plasmids, using standard methods (Sambrook et al., 1989). As a control, an E. coli strain containing a plasmid that encoded a GST fusion protein with protein-arginine N-methyltransferase 3 (PRMT3) (Tang et al., 1998) was processed in parallel with the GST-SIK and GST-SIKKD strains. Overnight cultures of the transformed bacteria were diluted 1:5 in 2× YT containing 100 μg/ml ampicillin and grown for 2 h with shaking at 37°C. Protein expression was induced by adding 500 μM IPTG, and growth was continued for an additional 2 h at 37°C. The bacteria were pelleted by centrifugation, resuspended in phosphate-buffered saline containing 2× complete protease inhibitor cocktail (Boehringer Mannheim, Indianapolis, IN, U.S.A.), and lysed by sonication. The proteins were purified by binding to glutathione-Sepharose beads as described by Pharmacia except that all steps were performed at 4°C. The proteins were eluted from the glutathione-Sepharose beads with 20 mM glutathione and frozen at -20°C. Aliquots of the purified fusion proteins were subjected to electrophoresis on 9% sodium dodecyl sulfate-polyacrylamide gels and stained with Coomassie Brilliant Blue.

For western analysis, the proteins were transferred from acrylamide to nitrocellulose using a Trans-Blot SD semidry transfer cell (Bio-Rad, Hercules, CA, U.S.A.) at 16 V for 40 min. The filter was treated by standard methods (Sambrook et al., 1989). The primary antibody, anti-GST (Pharmacia), was used at 1:1,000 dilution, and the secondary antibody, anti-goat-horseradish peroxidase (Sigma), was used at 1:8,000 dilution. Horseradish peroxidase was detected by ECL (Amersham, Arlington Heights, IL, U.S.A.).

Kinase assays

The in vitro autophosphorylation assay and kinase assay on histone (see Fig. 4) were performed as previously described (Hoover et al., 1991; Feldman et al., 1998b) with the following modifications: Reactions containing 8-15 μl (∼2 μg) of protein, 5-12 μl of phosphate-buffered saline (total volume of protein plus phosphate-buffered saline equal to 20 μl), and 9 μl of kinase buffer [stock concentrations: 25 mM HEPES (pH 7.5), 150 mM NaCl, 10 mM MgCl2, and 2 mM MnCl2] were set up on ice. The reactions were initiated by addition of 10 μCi of [γ-32P]ATP (Amersham) and incubated at room temperature for 30 min. The reactions were terminated by addition of 6 μl of 6× sodium dodecyl sulfate protein loading buffer. The samples were boiled for 5 min, subjected to electrophoresis on 9% sodium dodecyl sulfate-polyacrylamide gels, and stained with Coomassie Brilliant Blue. The gels were dried under vacuum and exposed to film at room temperature. Additional reactions included either 5 μg of histone H1 (Boehringer Mannheim) or 1 μl of purified GST protein (obtained by inducing protein expression from the native pGEX-4T-3 vector).

Figure 4.

GST-SIK is a protein kinase. Purified proteins were used in an in vitro kinase assay and subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis as described in the text: lanes 1-3, purified GST-SIK fusion protein; lanes 4-6, purified GST-SIKKD fusion protein; and lanes 7-9, purified GST-PRMT3 fusion protein. Lanes 1, 4, and 7, Coomassie Blue-stained sodium dodecyl sulfate-polyacrylamide gels of the products of the in vitro kinase assay; lanes 2, 5, and 8, autoradiogram of lanes 1, 4, and 7, respectively; and lanes 3, 6, and 9, enhanced chemiluminescence of western blot with anti-GST antibody of lanes from a duplicate gel without radioactive isotope. The 115-kDa GST-SIK fusion protein is barely visible in lane 1 but easily seen in lanes 2 and 3. The 58-kDa GST-SIKKD fusion protein is seen in lanes 4-6. The 110-kDa GST-PRMT3 fusion protein is seen in lanes 7 and 9 but not in lane 8. Lanes 2 and 5 show that both the GST-SIK and GST-SIKKD fusion proteins are able to catalyze autophosphorylation (2-h exposure). Lane 8 shows no phosphorylated species (12-h exposure).

FIG. 4.

The in vitro kinase assays using synthetic peptides as substrate (see Fig. 5) were performed by a modification of the phosphocellulose binding technique (Casnellie, 1991). Reactions containing 5 μl (∼8 μg) of autocamtide-2 (Calbiochem, La Jolla) or syntide-2 (BIOMOL, Plymouth Meeting, PA, U.S.A.), 1 μl (10 μCi) of [γ-32P]ATP, and 12.5 μl of 2× kinase buffer [50 mM HEPES (pH 7.5), 300 mM NaCl, 20 mM MgCl2, and 4 mM MnCl2] were set up at room temperature. The reactions were initiated by addition of 6.5 μl (∼1 μg) of purified GST-SIK or GST-SIKKD and incubated at room temperature for 30 min. Control reactions lacking peptide substrate substituted 5 μl of water, and those lacking GST-fusion protein substituted 6.5 μl of purified bacterial sonicate from cultures not treated with IPTG. The reactions were terminated by addition of 25 μl of 10% trichloroacetic acid and 10 μl of 10 mg/ml bovine serum albumin. The reactions were incubated in ice for 10 min and centrifuged at top speed in a microcentrifuge for 2 min to precipitate the protein. Then 25 μl of the supernatant was spotted onto 2.1-cm-diameter P81 disks (Whatman, Clifton, NJ, U.S.A.). The disks were washed five times in 500 ml of 75 mM H3PO4 for 5 min each, immersed in Bio-Safe II (RPI, Mount Prospect, IL, U.S.A.), and counted for radioactivity in a scintillation counter.

Figure 5.

GST-SIK is a serine protein kinase. Reactions were performed as described in the text. Reactions labeled “IPTG-” included purified bacterial lysates from cultures containing the GST-SIK and GST-SIKKD vectors that were not treated with IPTG. Inset: Coomassie Blue-stained sodium dodecyl sulfate-polyacrylamide gel of the purified bacterial lysates used in the kinase reactions.

FIG. 5.

KA treatment, preparation of brain sections, and in situ hybridization

Experiments were performed according to the rules and regulations of the National Institutes of Health and following review by the University of Southern California institutional animal care committee. These procedures were performed as previously described (Vician et al., 1995; Tocco et al., 1996; Feldman et al., 1998a,b) with the following modifications: The antisense (5′-CTGCTGGTAGTCAGATGAGACGGCTGAGTGGTGCTGTAAC
-3′) and the sense (5′-AGCAGTCTGGAGGTTCCTCAAGAAATTCTCCCATGTGACC
-3′) oligonucleotides were selected from nucleotides 2,118-2,157 and 1,246-1,285 of the SIK cDNA sequence, respectively.

RESULTS

SIK is preferentially induced by depolarization in PC12 cells

In a previous study, we used mRNA from potassium- and forskolin-stimulated PC12 cells as “tester” and mRNA from NGF- and EGF-stimulated PC12 cells as “driver” in an RDA procedure to enrich for cDNA fragments of messages preferentially expressed in the tester (potassium- and forskolin-stimulated) population (Feldman et al., 1998b). The third-round RDA difference product was resolved into a small number of distinct bands by electrophoresis on agarose (Feldman et al., 1998b). The KID-1 gene was cloned from an enriched band of that RDA difference product that migrated at 718 bp (Feldman et al., 1998b). In the current study, we first cloned cDNAs from the third-round RDA product that migrated at ∼400 bp. One clone, a 422-bp sequence, demonstrates a pattern of preferential induction, by northern analysis, in the original “tester” versus “driver” poly(A)+ RNAs (Fig. 1). The 5′ and 3′ sequences of this fragment were scanned against the current version of the nonredundant, updated GenBank and EMBL database and show strong similarity with the murine msk cDNA (Ruiz et al., 1994).

Figure 1.

RDA identifies a cDNA fragment that is preferentially expressed in the tester mRNA population from KCl + forskolin-treated cells. Poly(A)+ RNAs (50 ng per lane) from the original mRNA pools used to prepare the KCl + forskolin “tester” amplicons, NGF + EGF “driver” amplicons, and mRNA from untreated PC12 cells (control) were subjected to electrophoresis on agarose and transferred to a nylon membrane. Probing with the 422-bp fragment cloned from the third-round RDA difference product demonstrates preferential induction in the KCl + forskolin sample (5-day exposure with two intensifying screens). The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) signal (18-h exposure with one screen) is shown to evaluate mRNA loading.

FIG. 1.

The cloned 422-bp fragment was used to probe a cDNA library, prepared from 9-day-old rat cerebral cortex, to isolate longer clones. The nucleotide sequences of the overlapping clones suggest a contiguous sequence of 4,092 nucleotides. The size of the corresponding mRNA, estimated from northern analysis of induced PC12 cells, is ∼4 kb. The predicted open reading frame for this cDNA is 776 amino acids long, coding for an 85-kDa protein. The 5′-untranslated region contains 80% Gs and Cs. The 3′-untranslated region contains four ATTTA motifs and eight TATT motifs, sequences thought to confer mRNA instability (Shaw and Kamen, 1986; Reeves et al., 1987).

The open reading frame for the encoded protein contains sequence similarities to protein kinases at many critical residues and shares similarities with both the serine/threonine and tyrosine kinases (Fig. 2). We previously identified a depolarization-induced kinase that we named KID-1 (Feldman et al., 1998b). Based on its potential kinase activity and its induction in response to depolarization, we tentatively named this new depolarization-induced gene KID-2, for kinase induced by depolarization-2. However, during the preparation of this article, the identical cDNA sequence was identified as a gene induced in the adrenal gland of rats fed high-salt diets. The authors named the gene SIK (salt-inducible kinase) (Wang et al., 1999). SIK/KID-2 shares no significant sequence homology with KID-1, other than at the kinase domains. SIK contains residues in two catalytic subdomains, Lys151 and Glu153 in subdomain VI and Gly185, Ser186, and Tyr189 in subdomain VIII, that are (a) conserved in serine/threonine kinases and (b) thought to be associated with serine/threonine specificity (Hanks et al., 1988). As also concluded by Wang et al. (1999), these sequence similarities suggest that SIK is likely to be a serine/threonine protein kinase.

Figure 2.

SIK is the rat orthologue of msk. The predicted amino acid sequences of SIK and msk (Ruiz et al., 1994) were aligned using the Clustal method and MacVector software. Residues identical to SIK are shaded in gray. Protein kinase catalytic subdomains are indicated by Roman numerals beneath the aligned sequences. Individual residues conserved among serine/threonine protein kinases (+), tyrosine protein kinases (#), and both types of protein kinases (*) are indicated above the aligned sequences (Hanks et al., 1988). Regions of leucines and/or valines spaced seven residues apart are indicated by lines beneath the sequence (see RESULTS for details).

FIG. 2.

The SIK open reading frame is closely related to that of msk, a partial cDNA cloned in a study of protein kinase expression during murine cardiac development (Ruiz et al., 1994). The published msk sequence, which predicts a truncated open reading frame 433 amino acids long, is 94% identical to the rat SIK open reading frame (Fig. 2). The full-length msk cDNA was recently cloned and predicts a protein of 779 amino acids that shares 95% identity with SIK (J. Ruiz, Riley Hospital for Children, personal communication). The first 270 amino acids of SIK and msk share 98% identity, contain all of the consensus subdomains of a serine/threonine kinase, and are closely related to the sequences of the SNF1/AMPK family of serine/threonine protein kinases. The phylogenetic relationships among this family are illustrated in Fig. 3. The remaining 506 and 509 amino acids of SIK and msk, respectively, are 93% identical but have no significant homology with any other entry in the GenBank database. Therefore, we propose that rat SIK and murine msk are species orthologues and may define a new family of serine/threonine protein kinases.

Figure 3.

Phylogenetic analysis of the amino acid sequence of the SIKKD and other members of the SNF1/AMPK family was performed using MegAlign software. The proteins included in the phylogenetic analysis are the eight closest relatives to SIK by BLAST match.

FIG. 3.

The SIK carboxy-terminal region contains five regions of leucines and/or valines spaced seven residues apart, reminiscent of a “leucine zipper” (Landschulz et al., 1988). These residues and their positions relative to each other are conserved in msk, the murine SIK orthologue (Fig. 2). To function as a leucine zipper, the heptad leucine repeat must adopt a coiled coil tertiary structure (Hirst et al., 1996). We analyzed the SIK amino acid sequence for predicted coiled coil regions, using the Multicoil program (Wolf et al., 1997). Despite the conservation of these sequences, the Multicoil analysis does not predict a coiled coil structure in this region of SIK. This finding qualifies these regions of SIK as “nonzippers” (Hirst et al., 1996).

SIK is a serine protein kinase

The full-length GST-SIK fusion protein has a predicted molecular mass of 115 kDa. When expressed in bacteria, GST-SIK fusion protein preparations do contain a faint band of this size by Coomassie staining (Fig. 4, lane 1). The 115-kDa band is not seen in purified lysates of bacteria that were not treated with IPTG (data not shown). Therefore, this band represents an IPTG-induced product. Moreover, an in vitro protein kinase assay demonstrates that the GST-SIK fusion protein is able to catalyze autophosphorylation of this 115-kDa band (Fig. 4, lane 2). Western analysis using an anti-GST antibody confirms that this 115-kDa band is a GST fusion protein (Fig. 4, lane 3). There is no protein kinase activity in the lane containing the control, GST-PRMT3 fusion protein (Fig. 4, lane 8). Additional kinase assays demonstrate that the preparation of GST-SIK used in these assays is not able to catalyze phosphorylation of histone H1 or GST proteins (data not shown).

Several lower-molecular-mass bands ranging between 75 and 100 kDa in the GST-SIK fusion protein kinase assay, which are faint by Coomassie staining, are strongly phosphorylated and react with anti-GST antibody (Fig. 4, lanes 1-3). The dominant band on the Coomassie-stained GST-SIK preparation has an approximate molecular mass of 60 kDa. This is the smallest band that undergoes significant phosphorylation. We presume that these lower-molecular-mass bands between 60 and 100 kDa represent cleavage products of the full-length GST-SIK protein. Retention of the GST moiety (required for detection with anti-GST antibody in lane 3) suggests that progressively larger fragments are cleaved from the carboxy terminus, yielding a family of smaller fragments that still contain the GST sequence.

The consensus kinase domains are contained in the first 270-amino terminal amino acids of the predicted SIK protein. We hypothesized that the remaining amino acid sequence might contain elements that regulate the activity of the kinase domain. Therefore, we tested the kinase activity of a truncated GST fusion protein that included only the kinase domain of SIK (GST-SIKKD). The predicted molecular mass of this GST fusion protein is 58 kDa. Recombinant GST-SIKKD has a dominant 58-kDa band by Coomassie staining (Fig. 4, lane 4). This band retains significant in vitro autokinase activity and reacts with anti-GST antibody (Fig. 4, lanes 5 and 6).

To determine the amino acid specificity for the GST-SIK kinase activity, we performed additional kinase assays using the synthetic peptides autocamtide-2 and syntide-2 as substrates. These peptides both contain the substrate recognition motifs for SNF1 and AMPK (Dale et al., 1995), but autocamtide-2 has a threonine as the phosphate acceptor, whereas syntide-2 has a serine. Both GST-SIK and GST-SIKKD have the substantial ability to phosphorylate syntide-2, whereas neither has the ability to phosphorylate autocamtide-2 (Fig. 5). Thus, under the conditions of this experiment, GST-SIK and GST-SIKKD are serine but not threonine kinases. In this respect, SIK is similar to SNF1 and dissimilar to AMPK (Dale et al., 1995).

Control reactions either containing purified bacterial lysate from cultures not treated with IPTG or without peptide substrate showed minimally increased counts over background (Fig. 5). These controls demonstrate that background proteins from the bacterial lysate do not phosphorylate the peptides and that the autophosphorylated recombinant proteins are successfully precipitated by trichloroacetic acid and/or washed off the phosphocellulose disks before counting.

SIK is an immediate early gene that is preferentially induced by forskolin and depolarization in PC12 cells

SIK mRNA levels are relatively low in untreated PC12 cells (Fig. 6A). Exposure to depolarizing levels of KCl induces transient expression of SIK message, which returns to basal levels by 2-4 h. Forskolin treatment leads to a more extensive and longer-lasting induction of SIK mRNA accumulation: Expression is induced by 0.75 h, peaks by 1.5-2 h, and persists, albeit at diminished levels, for at least 4 h. In contrast to the results with forskolin and KCl, stimulation with NGF or EGF does not induce accumulation of SIK mRNA. These two growth factors do induce accumulation of the c-fos and TIS8/egr1 IEGs, demonstrating that the PC12 cells are competent to respond to these growth factors and that the growth factors used in these experiments are active.

Figure 6.

SIK mRNA is induced by KCl, forskolin, and A23187 but not by growth factors. Total RNA (10 μg per lane) from PC12 cells was subjected to electrophoresis on agarose and transferred to nylon membranes. A: Time course of SIK induction. Specific treatments and durations are indicated. Samples from untreated cells are indicated by “0.” Exposure time for SIK was 8 h with two intensifying screens. Hybridization with probes for c-fos (48-h exposure with two screens) and TIS8/egr1 (3-h exposure with one screen) shows that the cells responded appropriately to the growth factor treatments. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) signal (30-h exposure with one screen) is shown to evaluate mRNA loading. B:SIK is induced by KCl, forskolin, and A23187. Specific doses of KCl and forskolin are indicated. The dose of A23187 was 10 μM. All treatments were for 0.75 h. Exposure time for SIK was 16 h and for GAPDH 1 h, each with one intensifying screen.

FIG. 6.

In an additional experiment, we tested the ability of the calcium ionophore A23187 and various doses of KCl and forskolin to induce SIK expression in PC12 cells (Fig. 6B). Depolarizing levels of KCl induce SIK expression in an apparently dose-related fashion. Calcium ionophore, which allows calcium influx independent of depolarization, also induces SIK expression. Forskolin induces SIK expression equally at all doses tested. We conclude that SIK is induced by membrane depolarization and by direct activation of adenylyl cyclase but is not induced by neurotrophins or by growth factors in PC12 cells. These are precisely the criteria for gene identification that we set for the original selection procedures used to create the RDA difference products (Feldman et al., 1998b) (see also Discussion).

The presence of CHX, an inhibitor of protein synthesis, does not prevent SIK induction by KCl and by forskolin (Fig. 7). Therefore, SIK is an IEG/primary response gene (Herschman, 1991) that is preferentially induced by depolarization and forskolin and is unresponsive to NGF or EGF stimulation in PC12 cells. Once again, c-fos and TIS8/egr1 inductions are included to demonstrate the competence of the cells to respond to NGF/EGF and the biological efficacy of the growth factors. In contrast, the induction of transin (Machida et al., 1989), a secondary response gene, in response to NGF is blocked by CHX in the same experiment (data not shown). This control demonstrates the efficacy of CHX in differentiating primary response genes and secondary response genes following stimulation by ligand.

Figure 7.

SIK is an IEG. SIK is induced in the presence of CHX. Total RNA (8 μg per lane) from PC12 cells was subjected to electrophoresis on agarose and transferred to nylon membranes. Specific treatments are indicated. All treatments were for 1.5 h. Exposure time for SIK was 8 h with two intensifying screens. Hybridization with probes for c-fos (5-h exposure with two screens) and TIS8/egr1 (1-h exposure with one screen) shows that the cells responded appropriately to the treatments. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) signal (30-h exposure with one screen) is shown to evaluate mRNA loading.

FIG. 7.

SIK is present in brain and other tissues

Northern blot analysis of RNA preparations from several brain regions and from various other tissues demonstrates that SIK is not unique to brain (Fig. 8). SIK mRNA is present in the adrenal gland, the tissue of origin of PC12 cells. Besides the adrenal gland, SIK mRNA is found in the highest levels in kidney, lung, ovary, and pituitary. Basal levels of SIK mRNA are detectable in the brainstem, cortex, and hippocampus but not in the cerebellum. The tissue expression patterns of two other protein kinases induced by depolarization, KID-1 and PIM-1 (Feldman et al., 1998a,b), are shown for comparison (see DISCUSSION).

Figure 8.

SIK mRNA is present in unstimulated tissues, including brain. Total RNA (10 μg per lane) from tissues of untreated rats was subjected to electrophoresis on agarose, transferred to a nylon membrane, and probed with the 422-bp fragment representing SIK (7-day exposure with two intensifying screens). Similar analyses of KID-1 (Feldman et al., 1998b) and PIM-1 (Feldman et al., 1998a) (7-day exposure with two intensifying screens for each) are included for comparison. The data illustrating the tissue expression of KID-1 and PIM-1 were described previously (Feldman et al., 1998a,b) and were derived by stripping and reprobing the same filters. These data are presented again here to facilitate comparisons. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) signal (15-h exposure with one screen) is shown to verify that the RNA is intact in all samples. Ethidium bromide staining of 18S ribosomal RNA is shown to evaluate total RNA loading.

FIG. 8.

SIK is induced in brain following seizure induction by systemic administration of KA

Our major goal is to identify genes that may serve as mediators of neuronal plasticity, in response to in vivo stimulation. Consequently, although we identified SIK as a depolarization-inducible gene in PC12 cells, we used a seizure paradigm to determine whether neuronal stimulation leads to in vivo SIK induction in brain. Using in situ hybridization, we analyzed the regional distribution and time course of SIK expression in rat brain following KA-induced seizure activity. SIK mRNA levels detected by in situ hybridization are extremely low in the brains of control animals (Fig. 9A). The SIK mRNA levels rapidly increase in all hippocampal fields and in the cortex as early as 1 h after seizure onset. In the granule cells of the dentate gyrus, SIK mRNA levels increase to ∼800% of the control level by 1 h and remain higher than in controls up to 8 h after seizure onset (∼500% of the control level). Pyramidal cells in the CA1 and CA3 regions of the hippocampus exhibit a delayed peak expression of SIK mRNA as compared with granule cells of dentate gyrus. In these regions, SIK mRNA levels increase to ∼300% of control at 1 h after seizure onset and reach a maximal induction of ∼550% at 8 h in the CA1 region and 400% at 4 h in the CA3 region. In the cortical layers overlying the hippocampus, SIK mRNA levels increase to ∼200% of control levels at 1 h and remain at that level up to 8 h after seizure onset. In the piriform cortex, SIK mRNA levels increase to ∼200% of control at 4 and 8 h after seizure onset. We do not observe any change in SIK mRNA levels in the dorsal thalamic nuclei. In control experiments, specificity of hybridization is demonstrated by the observation that signals are eliminated when a 100-fold excess of unlabeled probe is added to the hybridization solution, when sections are pretreated with RNase A, and when a sensestrand oligonucleotide is used as a probe (data not shown).

Figure 9.

SIK mRNA is induced in specific regions of the hippocampus and cortex following KA-induced seizures. Male Sprague-Dawley rats (four animals per group) were injected subcutaneously with KA (12 mg/kg) and killed at 1, 4, and 8 h after the onset of seizure activity. An additional four animals (control) were killed without prior handling. Coronal sections (10 μm) were hybridized with an oligonucleotide probe specific for SIK (see Materials and Methods). A: Sections from individual animals. B: The optical density of specific brain regions was determined as previously described (Vician et al., 1995; Tocco et al., 1996). Data are mean ± SEM (bars) values, expressed as percent values compared with the optical density for the brain regions from control animals. ROD, relative optical density. Asterisks indicate statistically significant differences from control: *p < 0.05, **p < 0.001. CA1 and CA3, hippocampal pyramidal cell layers; DG, granule cell layer of the dentate gyrus; CTX, parietal cortex; PIR, piriform cortex; THAL, dorsal thalamic nuclei.

FIG. 9.

DISCUSSION

We, as well as others, are identifying genes induced in neuronal cell populations in response to depolarizing stimuli, on the assumption that the products of such genes will modify subsequent neuronal properties and mediate neuronal plasticity. Studies using various differential and subtractive techniques have identified several IEGs that are induced in brain by depolarization (Nedivi et al., 1993; Qian et al., 1993; Yamagata et al., 1993, 1994; Vician et al., 1995; Feldman et al., 1998b, 2000). Most recently, an extensive screen of cDNAs isolated from the hippocampus of KA-treated rats identified a large number of “candidate plasticity-related genes” (CPGs) that include signal transduction proteins, transcription factors, structural proteins, seven-transmembrane domain receptor-like proteins, protein phosphatases, and protein kinases (Hevroni et al., 1998).

We used PC12 cells as the starting population in our cloning procedures because of their ease of growth and manipulation, their clonal characteristics, and the ability to carefully define the populations of starting mRNA populations from which differentially induced genes are cloned. One selection scheme we used emphasizes a selection for neuron-specific, depolarization-inducible genes. As a result of that selection, we identified syt IV as a brain-specific, depolarization-induced IEG (Vician et al., 1995) that is integrated into synaptic vesicles (Ferguson et al., 1999) and modulates neurotransmitter release (Thomas et al., 1999).

We used a second selection/screening system to isolate cDNAs for genes that are induced by depolarization and/or forskolin stimulation, but not by neurotrophins or growth factors, in PC12 cells. Depolarization and forskolin both activate the transcription factor cyclic AMP-responsive element-binding protein (CREB), by stimulation of calcium/calmodulin-dependent protein kinase in the former case (Sheng et al., 1991) and by activation of protein kinase A via adenylyl cyclase in the latter case (Seamon and Daly, 1981). Both protein kinase cascades phosphorylate CREB at Ser133 and activate its transcription factor properties (Ghosh and Greenberg, 1995). Several biochemical studies, gene disruption studies in mice and flies, and dominant-negative CREB transgene studies have demonstrated that CREB activation plays a critical role in behavioral characteristics such as long-term memory, spatial learning, fear responses, and aggressive behavior (Silva et al., 1992; Bourtchuladze et al., 1994; Chen et al., 1994; Yin et al., 1994). However, the IEGs expressed in response to CREB activation that begin the cascades leading to these phenotypic responses are unknown. We reasoned that genes induced in neurons by depolarization and/or forskolin, but not by neurotrophins and/or growth factors, might be likely to encode molecules that mediate functional plasticity. In contrast to our cloning scheme for depolarization-induced, neuron-specific genes (Vician et al., 1995), genes meeting the current criteria might well modulate alternative phenotypic responses in other tissues, in response to other classes of stimuli, and thus would not a priori be expected to be expressed only in the nervous system.

Our initial examination of the enriched cDNA population created by using amplicons from depolarization/forskolin-stimulated PC12 cells as “tester” and amplicons from NGF/EGF-stimulated PC12 cells as “driver” identified KID-1, a gene that is regulated in PC12 cells in just the fashion for which our enrichment procedure selects. KID-1 is induced in PC12 cells by both depolarization and by forskolin but is not induced by either neurotrophin (NGF) or growth factor (EGF) stimulation (Feldman et al., 1998b). KID-1 is closely related by sequence to PIM-1 (Selten et al., 1986; Reeves et al., 1990; Wingett et al., 1992; Palaty et al., 1997). Like KID-1, PIM-1 is induced in rat hippocampus in response to systemic KA administration (Feldman et al., 1998a,b).

The above RDA experiment also identified rTLE3, a transducin-like enhancer of split and the rat orthologue of mouse and human TLE3. Like KID-1, rTLE3 is also induced by depolarization and forskolin but not by NGF or EGF in PC12 cells (Feldman et al., 2000). Following KA-induced seizures, rTLE3 is induced specifically in the dentate gyrus of the hippocampus. rTLE3 is the only enhancer of split-Groucho homologue demonstrated to be induced by neuronal depolarization (Feldman et al., 2000).

Additional examination of the amplicon population described above (Feldman et al., 1998b) has now identified a gene we tentatively called KID-2 (kinase induced by depolarization-2) as a third protein kinase induced by depolarization/forskolin but not by neurotrophins/growth factors in PC12 cells. During the preparation of this manuscript, the cloning of SIK from the adrenal glands of rats treated with high-salt diets was reported (Wang et al., 1999). These authors used PCR-coupled subtractive hybridization to identify transcripts that were more prevalent in the adrenal zona glomerulosa of K+-treated rats than in the adrenal zona fasciculata/reticularis and medulla of Na+-treated rats. After isolating the mRNA for SIK, they showed that the SIK transcript was, in fact, induced to a similar extent in the adrenal glands of both K+- and Na+-treated animals. Like KID-1 and PIM-1, KID-2/SIK is also induced by depolarization in brain. KID-1, PIM-1, and SIK are the only protein kinases shown to be induced by depolarization/forskolin, but not by neurotrophins/growth factors, in PC12 cells. CPG16, an additional depolarization-induced protein kinase, was isolated from a subtracted library made from mRNAs from dentate gyrus of KA-stimulated rats (Hevroni et al., 1998). It will be of great interest to see if cpg16 expression in PC12 cells is, like that of KID-1, PIM-1, and SIK, induced by depolarization/forskolin and not by neurotrophins/growth factors.

Following KA-induced seizure, the levels of SIK mRNA increase throughout the hippocampal formation as early as 1 h after seizure onset. The highest induction level is observed in the dentate gyrus, where SIK levels increase eightfold over control values. The pattern of SIK mRNA induction is similar to that of c-Fos, which increases >10-fold within 1 h of seizure onset (Sonnenberg et al., 1989; Schreiber et al., 1993b). The magnitude of increase in SIK expression is substantially greater than many IEGs such as SGP-2 (Schreiber et al., 1993a), KID-1 (Feldman et al., 1998b), and Nurr1 (Crispino et al., 1998). Although these genes also show increased expression throughout the hippocampal formation, their induction is much lower in amplitude, limited to two- to fourfold increase over control levels.

SIK is the rat orthologue of the murine gene msk. msk was identified as a gene thought to be expressed exclusively in the myocardial cells and their progenitors in 7.75-8.5 days postconception murine heart (Ruiz et al., 1994). The name msk (myocardial SNF1-like kinase) is derived from the sequence similarity to the yeast SNF 1 gene (Celenza and Carlson, 1986). Only a partial cDNA clone of msk was initially described (Ruiz et al., 1994). Although the kinase domains of SIK/msk share sequence similarity with many other serine/threonine protein kinases, the remainders of the SIK/msk proteins share no significant homology with any other proteins in the GenBank database. We do not yet know whether SIK/msk is a member of a new family of protein kinases or is a singular protein kinase.

Both GST-SIK and GST-SIKKD are able to catalyze autophosphorylation and serine phosphorylation of syntide-2 but not phosphorylation of histone H1 or threonine phosphorylation of autocamtide-2. These peptide phosphorylation data are consistent with SIK sharing the SNF1 substrate recognition motif (Dale et al., 1995). Wang et al. (1999) also showed that GST-SIK is capable of autophosphorylation. In contrast to SIK, recombinant KID-1 and PIM-1 can phosphorylate histone H1 (Hoover et al., 1991; Feldman et al., 1998b), suggesting that the substrate specificities of the various depolarization-induced neuronal kinases will vary. CPG16, the other depolarization-induced serine/threonine protein kinase, is also able to catalyze autophosphorylation in vitro (Hevroni et al., 1998). Clearly, the subcellular localization and identification of the targets of phosphorylation by these depolarization-induced protein kinases, KID-1, SIK, PIM-1, and CPG16, are important immediate goals of the next phase of this research.

The selection procedure we used to identify KID-1 and SIK does not enrich for mRNAs preferentially expressed in brain (Feldman et al., 1998b). Although KID-1, SIK, and PIM-1 share similarities in their induction by depolarization/forskolin in PC12 cells and KA in brain, their basal expression in brain and other tissues differ from one another (Fig. 8). These results suggest that these three protein kinases are likely to play modulating roles in various cellular processes in different tissues. However, their roles in neuronal responses to depolarizing stimuli may be coordinated, based on their induction by depolarization/forskolin, but not by growth factors or neurotrophins.

Although SIK expression is restricted to one class of inducers in PC12 cells, SIK can be induced by a distinct stimulus in a different cell type. PC12 cells, used in our work, are derived from adrenal medulla, and we induced SIK expression by K+-stimulated membrane depolarization and forskolin. Y1 cells, used by Wang et al. (1999), are derived from adrenal cortex and were stimulated by adrenocorticotropic hormone, a protein hormone. The adrenal medulla and cortex are embryologically and functionally distinct organs. However, membrane depolarization, forskolin, and adrenocorticotropic hormone all lead to activation of adenylyl cyclase, albeit by distinct mechanisms. Therefore, activation of adenylyl cyclase may lead to SIK induction in various different conditions.

SIK joins the protein kinases KID-1 and PIM-1, the transducin-like enhancer of split rTLE3, the synaptic vesicle protein SytIV, the protein dual-specificity phosphatase MKP-1/3CH134, and the orphan nuclear receptor transcription factors NGFIB/nur77/TIS1 and NURR1/HZF-3 as genes known to show much stronger induction by depolarization/forskolin than by neurotrophins or growth factors in PC12 cells (Bartel et al., 1989; Law et al., 1992; Vician et al., 1995; Deortiz and Jamieson, 1996; Feldman et al., 1998a,b, 2000). It will be of interest to see which of the CPGs recently identified by other groups (Okabe et al., 1996; Kato et al., 1997; Pazman et al., 1997; Hevroni et al., 1998; Ingi et al., 1998; Nakayama et al., 1998) demonstrate this sort of preferential induction. It will also be of great interest to determine whether genes induced preferentially by depolarization versus growth factor stimulation in PC12 cells/neurons do, in fact, play generally more important roles in modulating neuronal plasticity than do genes that are also induced in PC12 cells/neurons by a broader range of stimuli.

Acknowledgements

We thank Raymond Basconcillo and Arthur Catapang for technical support, Duncan MacLaren for computer imaging assistance, and the members of the Herschman laboratory for helpful discussions. This work was supported by grants NS28660 (to H.R.H.) and NS18427 (to M.B.) and Physician Scientist Award AR01870 (to J.D.F.) from the National Institutes of Health and a “Dottorato di Ricerca in Neuroscienze” fellowship (to M.C.) from the University of Naples, Italy. Personal communications regarding msk were provided by Joseph C. Ruiz, Riley Hospital for Children, Indianapolis, IN, U.S.A.

Ancillary