F-Actin Binding Domains Abnormalities (NAD 47/89)

Home , Protein domain

Human Lymphocyte-Specific Protein 1, the Protein Overexpressed in Neutrophil Actin Dysfunction with 47-kDa and 89-kDa Protein Abnormalities (NAD 47/89), Has Multiple This information is current as F-Actin Binding Domains of September 26, 2021. Qihong Zhang, Yao Li and Thomas H. Howard J Immunol 2000; 165:2052-2058; ; doi: 10.4049/jimmunol.165.4.2052 http://www.jimmunol.org/content/165/4/2052 Downloaded from

References This article cites 33 articles, 15 of which you can access for free at:

http://www.jimmunol.org/content/165/4/2052.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 26, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2000 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Human Lymphocyte-Speciﬁc Protein 1, the Protein Overexpressed in Neutrophil Actin Dysfunction with 47-kDa and 89-kDa Protein Abnormalities (NAD 47/89), Has Multiple F-Actin Binding Domains1

Qihong Zhang, Yao Li, and Thomas H. Howard2

Human lymphocyte-specific protein 1 (LSP1) is an F-actin binding protein, which has an acidic N-terminal half and a basic C-terminal half. In the basic C-terminal half, there are amino acid sequences highly homologous to the actin-binding domains of two known F-actin binding proteins: caldesmon and the villin headpieces (CI, CII, VI, VII). However, the exact numbers and Downloaded from locations of the F-actin binding domains within LSP1 are not clearly defined. In this report, we utilized 125I-labeled F-actin ligand blotting and high-speed F-actin cosedimentation assays to analyze the F-actin binding properties of truncated LSP1 peptides and to define the F-actin binding domains. Results show that LSP1 has at least three and potentially a fourth F-actin binding domain. All F-actin binding domains are located in the basic C-terminal half and correspond to the caldesmon and villin headpiece homologous regions. LSP1 181–245 and LSP1 246–295, containing sequences homologous to caldesmon F-actin binding site I and II, respectively (CI, CII), binds F-actin; similarly, LSP1 306–339 can bind F-actin and contains two inseparable villin headpiece- http://www.jimmunol.org/ like F-actin binding domains (VI, VII). Although LSP1 1–305, which does not contain VI and VII regions, retains F-actin binding activity, its binding affinity for F-actin is much weaker than that of full-length LSP1. Site-directed mutagenesis of the basic amino acids in the KRYK (VI) or KYEK (VII) sequences to acidic amino acids create mutants that bind F-actin with lower affinity than full-length wild-type LSP1. High KCl concentrations decrease full-length LSP1 binding to F-actin, suggesting the affinity between LSP1 and F-actin is mainly through electrostatic interaction. The Journal of Immunology, 2000, 165: 2052–2058.

he microfilamentous cytoskeleton is a three-dimensional, of an 89-kDa protein. In addition, the NAD 47/89 PMNs have a dynamic network that regulates several types of cell mo- unique morphology and abnormal microfilamentous structure, re- T tility, including pseudopodial extension, cell crawling, sulting from defective actin polymerization in response to chemo- by guest on September 26, 2021 cell shape change, cell locomotion, chemotaxis, and phagocytosis tactic factor stimulation, and this leads to a decreased ability of the in variety of cells including polymorphonuclear neutrophils cells to spread on glass (3, 4). (PMNs).3 The actin-rich PMNs require actin-based motility for Cloning and sequencing of the 47-kDa-protein cDNA revealed their functions in the human immune response and are therefore an a nucleotide sequence nearly identical to that of the lymphocyte- excellent model system in which to study cell motility and cy- specific protein 1 (LSP1), a mouse cDNA sequence reported in toskeletal dynamics. GenBank (5) and subsequently identified as an F-actin binding Direct evidence which links PMN motility and cytoskeletal dy- protein (6). Human LSP1 is a phosphoprotein composed of 339 aa namics are derived in part from studies with pharmacological (7). The gene for this protein was cloned from mouse lymphocyte agents such as cytochalasins and phalloidins and studies of human library using a subtractive hybridization technique (8), and studies cell motility disorders auch as the neutrophil actin dysfunction showed that this gene is expressed in normal murine B and T (NAD) (1, 2). In 1991, Coates and colleagues (3) described a lymphocytes and in transformed B cells but not in T lymphoma unique PMN disorder, NAD with abnormal 47-kDa and 89-kDa cell lines (5). Subsequent studies showed LSP1 is not limited to proteins (NAD 47/89). The NAD 47/89 PMNs could not move and lymphocytes but is a pan-leukocyte protein (9, 10). Human LSP1 contained increased amounts of a 47-kDa and decreased amounts has an acidic N-terminal half which is 53% homologous to its mouse counterpart and a basic C-terminal half which is 85% homologous to its mouse counterpart (6). The basic C-terminal do- Department of Cell Biology, School of Medicine, University of Alabama, Birming- main contains amino acid sequences homologous to two known ham, AL 35295 F-actin binding protein caldesmon and the villin headpiece. Al- Received for publication May 26, 1999. Accepted for publication May 30, 2000. though it is known that LSP1 is an F-actin binding protein and is The costs of publication of this article were defrayed in part by the payment of page a very important regulator of microfilamentous cytoskeleton dy- charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. namics, it is not known which regions of the molecule are required 1 This work was supported by National Institutes of Health Grant R01 HL56144 for the F-actin binding activity. In the studies reported here, two (to T.H.H.). methods were used to define the F-actin binding domains of this 2 Address correspondence and reprint requests to Dr. Thomas H. Howard, Division of molecule. The results show that human LSP1 has at least three Hematology-Oncology, Department of Pediatrics, University of Alabama, 1600 7th F-actin binding domains, all located in the basic C-terminal half, Avenue South, Birmingham, AL 35233. E-mail address: [email protected] corresponding to the homologous regions of F-actin binding do- 3 Abbreviations used in this paper: PMN, polymorphonuclear neutrophil; NAD, neutrophil actin dysfunction; LSP1, lymphocyte-specific protein 1; IPTG, isopropyl ␤-D- mains in two known actin-binding proteins: caldesmon and the thiogalactopyranoside; 125I-F-actin, 125I-labeled F-actin. villin headpiece.

Materials and Methods After incubation with 50 ␮g/ml 125I-F-actin at room temperature for 6 h Reagents during constant agitation, the blots were then washed extensively with TBST (10 mM Tris-HCl (pH 7.5), 0.5% (v/v) Tween 20, and 90 mM NaCl) Restriction enzymes BamHI and HindIII, T4 DNA ligase, and in vitro five times (15–30 min/wash). Blots were air-dried, and F-actin binding mutagenesis kit were purchased from Promega (Madison, WI). QIAEX II proteins were identified by autoradiography (12, 13). For cold probe com- gel extract kit was from Qiagen (Santa Clarita, CA). Proband resin was petition assays, the blots were incubated with either cold F-actin or micro- obtained from Invitrogen (San Diego, CA). PCR primers were synthesized tubule at room temperature for 2 h before the blots were air-dried. at the Department of Biochemistry and Molecular Genetics at the Univer- sity of Alabama (Birmingham, AL), 125I-labeled G-actin was prepared in In vitro site-directed mutagenesis the University of Alabama cancer center. Prokaryotic expression plasmid In vitro site-directed mutagenesis was done according to the protocol de- pET21C was from Novagen (Madison, WI). Isopropyl ␤-D-thiogalactopy- scribed for the Promega GeneEditor in vitro site-directed mutagenesis sys- ranoside (IPTG), ampicillin, gelsolin and phalloidin were purchased from tem. Synthesized oligonucleotides containing the mutated bases were phos- Sigma (St. Louis, MO). Taq DNA polymerase was purchased from Perkin- phorylated by T4 polynucleotide kinase, then the phosphorylated Elmer (Norwalk, CT). Bovine brain tubulin was purchased from Cytoskel- oligonucleotides were annealed with the plasmid which contained the eton (Denver, CO). LSP1 cDNA. The primers were extended by T4 DNA polymerase, and the gap was ligated by T4 DNA ligase, the mutants were selected by DNA Rabbit skeletal muscle G-actin preparation sequencing. The mutated proteins were purified by inserting the mutated Rabbit skeletal muscle G-actin was prepared according to the method of gene into prokaryotic expression vector pET21C and subsequently induced Spudich and Watt (11). G-actin dialyzed against G-buffer (5 mM Tris-HCl with IPTG. The expressed proteins were purified.

(pH 8.0), 0.2 mM ATP, 0.2 mM DTT, 0.2 mM CaCl2,and1mMNaN3) was further purified by passage over a Sephadex G-75 column. Before use, Results ϫ the G-actin was centrifuged at 100,000 g for 30 min in the Beckman Sequence comparison to known proteins reveals four potential Downloaded from (Palo Alto, CA) Optima TL Ultracentrifuge. actin-binding sequences in LSP1 Construction of truncated LSP1 peptides Database searches of the LSP1 amino acid sequences to define The strategy for creating truncated LSP1 peptides involved preparing PCR regions homologous to the actin-binding domains of known actin- primers which place a BamHI restriction site in front of the 5Ј primer, a binding proteins revealed four regions of interest in the C-terminal stop codon, and a HindIII restriction site behind the 3Ј primer (the LSP1 half of the protein. Two sequences separated by 23 aa were more cDNA contains neither BamHI nor HindIII restriction sites). The PCR than 60% homologous to the F-actin binding sites of caldesmon http://www.jimmunol.org/ products were excised using BamHI and HindIII and purified by QIAEX II gel extract kit. The purified PCR products were ligated into prokaryotic (CI and CII, Fig. 1) (14–17). The CI region contains a 33-aa se- expression vector pET21C, which had been cut with BamHI and HindIII quence (aa 205–237), where 18 of 33 aa are similar; the second and purified. When correctly ligated, this vector added six histidine resi- region of homology, CII, spans 17 aa (aa 260–276), where 9 of 17 dues to the C terminus of the peptide, which was then used to purify the aa are similar. In addition, near the extreme C-terminal end of peptides by nickel affinity chromatography. LSP1, there are also two short stretches of amino acids (VI and Expression of truncated protein and protein purification VII) that are almost identical to the KKEK sequences identified in the headpiece of villin as an F-actin binding domain. This head- The constructed plasmids were transformed into Escherichia coli strain BL21 (DE3), the bacteria were induced with 1 mM IPTG for3htoexpress piece is essential for the bundling activity and its morphologic the proteins, the bacteria were lysed by three freeze-thaw cycles in liquid effect(s) of the villin molecule on cells (18–20) (see Fig. 1). Fur- by guest on September 26, 2021 nitrogen and a 37°C water bath. After centrifugation at 100,000 ϫ g for1h, thermore, computer-based secondary structure analysis reveals a the supernatants were passed through a nickel affinity column and the series of ␣-helices positioned between and following CI and CII. bound proteins were subsequently purified. Also three ␤-pleated sheets that include the VI and VII domains High-speed F-actin cosedimentation assay follow the helical structure and are similar to those found in other F-actin binding proteins including spectrin and ABP280. Truncated and mutated LSP1 proteins in P-buffer (10 mM imidazole (pH 7.0), 75 mM KCl, 0.2 mM DTT, 0.2 mM EGTA, and 0.01% Nonidet P-40) Truncated peptide analysis shows LSP1 has at least three F- were first spun at 100,000 ϫ g at Beckman centrifuge for 45 min to remove aggregates. Assays of F-actin cosedimentation were performed by mixing actin binding domains actin with LSP1 in P-buffer containing 2 mM MgCl2 and 0.1 mM ATP. Our initial results showed human LSP1 purified from Sf9 cells After incubation for 60 min at room temperature, the samples were spun at infected with LSP1 cDNA recombinant baculovirus, similar to 100,000 ϫ g for 45 min. The pellets were washed twice with P-buffer. Both supernatants and pellets were analyzed by SDS-PAGE, and then the bands mouse LSP1, can bind to F-actin. In addition, human LSP1 colo- were visualized by Coomassie blue staining and quantified by a Bio-Rad calizes with F-actin in A7 melanoma cells that stably express LSP1 (Hercules, CA) densitometric scanner. protein from the pCEP4 vector (21). Because the basic C-terminal 125I-labeled G-actin labeling half of LSP1 contains regions highly homologous to the F-actin binding proteins caldesmon and villin headpiece, PCR was utilized 125 Column-purified G-actin was labeled with I-Bolton Hunter reagent to create an array of truncated LSP1 peptide constructs and were (New England Nuclear, Boston, MA). The labeled G-actin was separated from the free 125I by passing over G-75 gel filtration column. The concen- analyzed for F-actin binding activity. The constructs, designated tration of the G-actin was measured by BCA method. LSP1 x-y where x is N-terminal and y is C-terminal amino acid, were designed to include and exclude the V and C homologous F-actin ligand blotting sequences in combinations and alone within unique peptides. 125I-labeled G-actin at 1 mg/ml was polymerized to F-actin in the presence Truncated LSP1 cDNA was cloned into the prokaryotic expression of 0.2 ␮M gelsolin (100:1 mol actin to gelsolin) in polymerizing buffer (20 vector pET21C and transformed into E. coli strain BL21 (DE3) mM PIPES (pH 7.0), 50 mM KCl, 2 mM MgCl , and 50 mM CaCl ) for 2 2 and induced to express truncated peptides by IPTG. Initially, the 10 min on ice, then phalloidin was added to a final concentration of 40 ␮M and polymerization was continued at room temperature for an additional 15 LSP1 truncates expressed in bacteria were screened for F-actin 125 125 min. 125I-labeled F-actin (125I-F-actin) was diluted to a final concentration binding domains by I-F-actin ligand blot. Results of the I- of 50 ␮g/ml with blocking buffer (10 mM Tris-HCl (pH 7.5), 90 mM NaCl, F-actin ligand blot screen for F-actin binding domains are shown 0.5% (v/v) Tween 20, 5% nonfat milk, 1% BSA, and 0.25% gelatin). Trun- in Fig. 2. In whole bacterial lysates of E. coli expressing LSP1 cated LSP1 peptides were expressed in E. coli, and equal OD600 total cells 125 were lysed with Laemmli buffer and E. coli proteins separated on SDS- peptides induced by IPTG, I-F-actin binds to several, but not all, PAGE and transferred to nitrocellulose (NC) membrane. The NC mem- LSP1 peptides, suggesting that LSP1 contain more than one F- brane was first treated with blocking buffer at room temperature for 2 h. actin domain. No actin-binding peptides are present in control 2054 LSP1 HAS AT LEAST THREE F-ACTIN BINDING DOMAINS

FIGURE 1. Diagram of human LSP1 molecule. Computer analysis of LSP1 identiﬁes two caldesmon-like (CI and CII) (f) and two villin headpiece-like (VI and VII) (z) amino acid sequences in the C- terminal half of the molecule. These sequences are homologous to the actin-binding domains of caldesmon and the villin headpiece. Shown in single letter amino acid code are the location and sequences of the CI, CII, VI, and VII homologous regions. Numbers reveal the amino acids in N- to C-terminal direction for LSP1 (1–339), caldemon (597–667), and villin (805–826). The boundary of truncated Downloaded from peptides are labeled. http://www.jimmunol.org/

IPTG induced bacterial lysates, and no binding was observed with for F-actin binding domains by cosedimentation with F-actin at the nonspecific control, purified BSA protein. Gelsolin, a known high speeds (100,000 ϫ g for 45 min). Results with the high-speed F-actin binding protein on ligand blots (4), is included as a positive F-actin cosedimentation assay show that full-length LSP1 and the control (Fig. 2C). Specific screening results show that the full- C-terminal half of LSP1 (LSP1 181–339) (Fig. 3A) but not the length and C-terminal LSP1 (LSP1 181–339), but not the N-ter- N-terminal half (LSP1 1–180) (Fig. 3B) cosediment with F-actin. minal LSP1 (LSP1 1–180), peptides bind F-actin in the ligand blot. None of these peptides alone sediment under these assay condi- Also, both the LSP1 1–305, which contains both CI and CII, and tions. These results are consistent with the overlay data that F-actin the LSP1 306–339, which contains both VI and VII regions, bind binding domain(s) reside within the C-terminal half of the by guest on September 26, 2021 to F-actin on ligand blots (Fig. 2A). F-actin binding by ligand blot molecule. via CI and CII independently is also demonstrated with LSP1 181– To define more precisely the location and the number of F-actin 245 and LSP1 246–295 peptides (Fig. 2B). The screening results binding domains, additional truncated peptides that include or ex- suggest that LSP1 has at least three F-actin binding domains and clude the CI, CII, and VI, VII domains were created and analyzed. that these binding domains may differ in their apparent affinity for Both LSP1 181–295 that contains both CI and CII and LSP1 306– F-actin. 339 that contains both VI and VII can bind F-actin independently To specifically define the F-actin binding domains in human (Fig. 3C). Thus, LSP1 has at least two F-actin binding domains. LSP1, truncated LSP1 peptides purified from E. coli were analyzed Because the VI and VII domains are separated by only 7 aa, it was

FIGURE 2. 125I-F-actin ligand blotting screens of the LSP1 peptides for F-actin binding domains. Shown are autoradio- graphs of F-actin ligand blot assays which screen for F-actin binding domains by as- saying 125I-F-actin bound to the indicated LSP1 peptides expressed in whole bacterial lysates. Note the ligand blot results suggest there are three F-actin binding domains. Proteins were separated on 10% (A and C) and 15% (B) SDS-PAGE before transfer to nitrocellulose membrane for blotting. Con- trol in A refers to total bacterial protein of IPTG induced without the expression of LSP1 peptides, and while LSP1 1–180 expressed visible evidence of LSP1 peptide in IPTG induced bacterial cells (not shown), there is no binding on 125I-F-actin ligand blot. In C, puriﬁed BSA (10 ␮g) is included as a negative control and puriﬁed gelsolin (1 ␮g) is included as a positive control. The Journal of Immunology 2055

FIGURE 3. Recombinant LSP1 peptide cosedimentation with F-actin defines three F-actin binding domains. Shown are SDS-PAGE of supernants (S) and pellets (P) from solutions of purified LSP1 peptides alone or LSP1 peptides mixed with F-actin sedi- mented at 100,000 ϫ g for 45 min. Cosedimentation with F-actin of full-length LSP1 (LSP1 1–339), C- terminal one-half of LSP1 (LSP1 181–339), LSP1 181–295, LSP1 296–339, LSP1 306–339, LSP1 181– 245, and LSP1 246–295 peptides but not N-terminal half of LSP1 (LSP1 1–180) indicate there are at least three F-actin binding domains in the C-terminal half of the LSP1 molecule. Downloaded from technically difficult to express and purify VI and VII and analyze (LSP1 1–295) restores F-actin binding, suggesting that the 20 ad- them as independent peptides. In contrast, truncated peptides in- ditional amino acids contribute important secondary structure. cluding CI or CII (LSP1 181–245 and LSP1 246–295) were sufficiently large and sufficiently spaced to allow independent analy- Comparison of affinity of the F-actin binding domains in LSP1 sis of CI and CII to further define F-actin binding domains of truncates and LSP1 mutants

LSP1. The results show these peptides each retain an F-actin bind- Quantative analysis by Scatchard plot of murine LSP1 binding to http://www.jimmunol.org/ ing domain (Fig. 3D). In summary, these results show LSP1 has at F-actin by Jongstra-Bilen et al. (6) suggested that murine LSP1 least three, and possibly four, F-actin binding domains, CI, CII, might contain high- and low-affinity F-actin binding domain. How- and VI/VII. Given the sequence homologies, it is possible that VI ever, these domains were not clearly characterized. Our quantita- and VII can also independently bind F-actin. These results, sum- tive analysis of F-actin binding by cosedimentation shows that marized in Fig. 4, support the contention that LSP1 has at least LSP1 1–305 binding to F-actin is lower affinity than the binding of three F-actin binding domains. Two LSP1 peptides, LSP1 1–275 full-length LSP1 (Fig. 5A). This finding also suggests that LSP1 306– and LSP1 181–275, yielded unexpected results. Although both 339 indeed contain another F-actin binding domain. It should be noted contain the defined CI and CII sequences, neither binds F-actin. that at equal molar ratio of both LSP1 peptides, the quantity of LSP1 Inclusion of 20 additional C-terminal amino acids in these peptides 1–305 which cosediments with F-actin is quantitatively less (Ͻ20% by guest on September 26, 2021

FIGURE 4. Summary of F-actin binding sites within LSP1 molecule. At least three F-actin binding fragments were identiﬁed within LSP1 molecule: one that includes the CI region, the second contains the CII region, and at least one additional binding fragment is retained in the VI and VII region. 2056 LSP1 HAS AT LEAST THREE F-ACTIN BINDING DOMAINS

quences. Similar results are obtained on 125I-F-actin ligand blot (data not shown). The results show both the VI and VII sequences are independently important for F-actin binding. In combination with analysis of LSP1 truncates these results demonstrate that LSP1 contains at least three, and likely four, independent F-actin binding domains in its C-terminal half.

LSP1 binds to F-actin through electrostatic interaction The actin-binding domains of many actin-binding proteins contain clustered basic amino acids. This observation seems to be a re- peating theme in actin-binding proteins, which may reﬂect their evolutionary origin. Actin-binding domains at clustered basic amino acids in proteins include the KKGGKKKG sequence of myosin, DAIKKK sequence in actin depolymerizing factor, coﬁlin and tropomyosin, and KSKLKKT sequence in thymosin ␤4 (22). Vandekerckhove (22) has suggested that the extreme N-terminal peptide Ac-DEDE of actin molecule is the binding site on actin that is the ligand for these positively charged sequences on actin-

binding proteins. Several lines of evidences including chemical Downloaded from cross-linking are consistent with this suggestion (22, 23). Such a situation will inevitably lead to competition between different actin-binding proteins for actin in cells. Because the F-actin binding domains of LSP1 localize in the basic C-terminal half of the molecule, we reasoned that the interaction of LSP1 with F-actin may

also be through electrostatic interaction and therefore that LSP1 http://www.jimmunol.org/ may compete with other actin-binding proteins. To test the possibility that the LSP1 to actin interaction is controlled electrostati- cally, we increased the concentration of KCl in the F-actin cosedimentation assay to determine whether increased ionic strength would dissociate LSP1 from F-actin. The increase in KCl concentration did not affect the solubility of LSP1 protein (data not FIGURE 5. Both the deletion mutation and point mutation in the villin shown). Increasing KCl concentration did not affect the polymer- headpiece-like regions weaken LSP1 binding to F-actin. Shown are the ization of actin, as shown in Fig. 6, which was consistent with the results of F-actin cosedimentation assay of LSP1 1–305 truncate (A) and fact that actin-actin interaction is mainly hydrophobic (24). How- by guest on September 26, 2021 LSP1 K326E mutant (B). Compared with full-length wild-type LSP1 pro- ever, when the concentration of KCl was increased, the binding of tein, both LSP1 1–305 and LF K326E bind to F-actin weakly, though the LSP1 to F-actin decreased. This observation suggests that at least amount of proteins and actin is similar in each assay. C, Saturation binding one of the three F-actin binding domains associate with F-actin curves and Kd of wild-type and mutated LSP1 to F-actin. by cosedimentation) than that of LSP1 1–339. The result indicates the CI, CII binding to F-actin is weaker than full length and suggests the V regions may represent strong binding domains while C regions may represent relatively weaker binding domains. It is also possible that VI and VII domains may cooperate with each other and with CI and CII to strengthen F-actin binding. To exclude the possibility that the weaker binding of LSP1 1–305 to F-actin was due to the conformational change accompa- nying peptide truncation and to determine whether VI and VII can bind F-actin independently, full-length LSP1 mutants were created by site-directed mutagenesis to assay the contribution of VI and VII to F-actin binding independently. The results show both the VI and VII sequences are independently important for F-actin binding of LSP1 (Fig. 5B). VI and VII respectively contain the sequences KYEK and KRYK that are similar to the villin headpiece sequence KKEK, the F-actin binding domain of villin. Mutation of a single positively charged amino acid to glutamic acid (E) in the VII domain (KYEK to KYEE) weakens full-length LSP1 binding to F- actin (see Fig. 5B). Similar results were obtained with point mu- FIGURE 6. LSP1 binds to F-actin through electrostatic interaction. tations in the VI region. Furthermore, quantitative assay of LSP1 LSP1 has greatly decreased binding to F-actin with increasing KCl con- wild-type and VI or VII mutant binding to F-actin over a range of centration, whereas KCl concentration does not affect the polymerization molar ratios of LSP1 to F-actin (Fig. 5C) shows LSP1 afﬁnity for of actin. A, SDS-PAGE analysis of LSP1 in the presence of increasing KCl and saturation binding to F-actin is dramatically altered by basic concentrations (100–500 mM). B, The amount of actin and LSP1 in the amino acid to acidic amino acid changes in either VI or VII se- pellet determined by scanning of Coomassie blue stained gels. The Journal of Immunology 2057

the molecule: aa 181–245, aa 246–295, and aa 306–339. Each domain contains amino acid sequences homologous to those in the F-actin binding proteins caldesmon or the villin headpiece. Two complementary methods were utilized to define the F-actin binding domains. The high-speed F-actin cosedimentation assay utilizes purified native proteins and excludes specific or nonspecific inter- ference from other proteins. The F-actin ligand blotting utilizes 125I-radioactive labeled F-actin. This method is more useful in screening for potential F-actin binding domains including those peptides that have weak binding to F-actin and those peptides which might form homologous aggregates or easily precipitate under the conditions of F-actin cosedimentation assay condition. With the F-actin cosedimentation assay, it is difficult to determine whether homologous aggregates might explain cosedimentation with F-actin. In contrast, 125I-F-actin ligand blot screens independent of peptide aggregates. One limitation of the F-actin ligand blotting is that this method uses denatured and renatured proteins, resulting in the uncertainty that some proteins may not be properly

renatured. Combining these two methods provides the most accu- Downloaded from rate and easily interpreted results. FIGURE 7. LSP1 binds to F-actin specifically. Shown are autoradiog- 125 125 Although domain mapping through creation of actin-binding raphy of I-F-actin binding to wild-type LSP1. Note that I-F-actin protein truncates is a standard method for defining the functional bound to LSP1 is displaced by 50- to 100-fold excess of cold F-actin, but domains of actin-binding proteins, it has its limitation. For exam- not by excess cold taxol stabilized microtubules. Therefore, F-actin binding to LSP1 is specific for that polymer. ple, truncates may have a significantly altered conformation re-

sulting from deleting adjunct sequences, and therefore may pro- http://www.jimmunol.org/ vide little insight into the function of full-length actin-binding through electrostatic interaction. Such electrostatic interactions proteins. An example is the microtubule associated protein 2 may be nonspecific. However several lines of evidence show the (MAP2) and tau molecules binding to microtubule. It was shown LSP1 interaction with actin is specific. As shown in Fig. 7, 125I- that the extreme C-terminal hydrophobic ␣-helix in these proteins F-actin ligand blot demonstrate cold F-actin at Ͼ20-fold excess is responsible for the microtubule bundling activity of MAP2 and significantly dissociates 125I-F-actin from binding. This effect is tau by domain mapping (32). However, in subsequent experiments specific to F-actin polymer since 50- to 100-fold excess cold taxol exploiting site-directed mutagenesis and deletion mutation exper- stabilized microtubule does not compete away F-actin binding. iments, it was shown that the regions identified by truncation are not directly responsible for microtubule bundling activity of these by guest on September 26, 2021 Discussion proteins (33). Therefore, caution should be taken in interpreting The polymerization and depolymerization between monomeric G- data from truncated peptides alone to map the functional domains actin and filamentous F-actin is regulated by a number of actin- of a molecule as the sole method for defining actin-binding do- binding proteins. According to their mode of interaction with actin mains. The same situation has been encountered in this study with in vitro, these actin-binding proteins are classified into several analysis of LSP1 actin-binding domains. Clearly, the results show groups: 1) monomer binding proteins, such as profilin, that bind to the truncated peptide LSP1 181–245 and peptide LSP1 246–295, G-actin and sequester G-actin to limit actin polymerization; 2) F- which respectively contain CI and CII alone, can each bind to actin capping proteins such as gelsolin, that cap the rapidly grow- F-actin independently, and LSP1 1–295, which include these two ing plus-end of F-actin filaments, so F-actin cannot grow from the fragments, can also bind to F-actin. However, LSP1 1–275, which barbed end; 3) F-actin severing proteins, such as severin, that cut include LSP1 181–245 and part of LSP1 246–295, does not bind longer F-actin filaments into shorter F-actin filaments; and 4) F- to F-actin either in the F-actin cosedimentation assay or the F-actin actin bundling proteins like villin, fimbrin, that cross-link actin ligand-blotting assay. This result suggests aa 275–295 may modify filament into a parallel array (25–27). The ability of these proteins the conformation of the truncates and therefore affect the F-actin to bind actin is regulated by receptor-activated signaling events binding. Thus the three-dimensional structure of the molecule is as including [Ca2ϩ], phospholipid, and protein phosphorylation (28– important as the primary amino acid sequences. To fully confirm 31) and thereby links receptor signaling to cytoskeletal dynamics. the findings reported here, the crystal structure of the LSP1 would Each actin-binding protein has its own unique binding affinity for be required. actin, some of them may bind to the same region of the actin molecule, and therefore they may compete with each other. The Acknowledgments competition, cooperation, and regulatory interactions among these We thank Dr. William Nauseef for the completing initial screen for binding actin-binding proteins modulate microfilamentous cytoskeleton sites with the 125I-F-actin ligand blot assay. dynamics in nonmuscle cells. To understand how these actin-binding proteins regulate microfilamentous cytoskeleton dynamics, it is essential to define the actin-binding domains in each actin-binding References protein. Domain mapping is still the major approach to explore the 1. Boxer, L. A., E. T. Hedley-Whyte, and T. P. Stossel. 1974. Neutrophil actin dysfunction and abnormal neutrophil behavior. N. Engl. J. Med. 291:1093. functional domains of actin-binding proteins. 2. Southwick, F. S., G. A. Dabiri, and T. P. Stosse. 1988. Neutrophil actin dysfunc- In this study, we have mapped the F-actin binding domains of tion is a genetic disorder associated with partial impairment of neutrophil actin human LSP1, an F-actin binding protein. Overexpression of LSP1 assembly in three family members. J. Clin. Invest. 82:1525. 3. Coates, T. D., J. C. Torkildson, M. Torres, J. A. Church, and T. H. Howard. 1991. modifies the cytoskeletal structure and motility of PMNs. The F- An inherited defect of neutrophil motility and microfilamentous cytoskeleton actin binding domains of LSP1 were mapped to three regions of associated with abnormalities in 47-kD and 89-kD proteins. Blood 78:1338. 2058 LSP1 HAS AT LEAST THREE F-ACTIN BINDING DOMAINS

4. Howard, T., Y. Li, M. Torres, A. Guerrero, and T. Coates. 1994. The 47-kD 18. Friederich, E., C. Huet, M. Arpin, and D. Louvard. 1989. Villin induces mi- protein increased in neutrophil actin dysfunction with 47- and 89-kD protein crovilli growth and actin redistribution in transfected fibroblasts. Cell 59:461. abnormalities is lymphocyte specific protein. Blood 83:231. 19. Friederich, E., E. Pringault, M. Arpin, and D. Louvard. 1990. From the structure 5. Jongstra, J., G. F. Tidmarsh, J. Jongstra-Bilen, and M. M. Davis. 1988. A new to the function of villin, an actin-binding protein of brush border. BioEssays ϩ lymphocyte-specific gene which encodes a putative Ca2 -binding protein is not 12:403. expressed in transformed T lymphocyte lines. J. Immunol. 141:3999. 20. Friederich, E., K. Vancompernolle, C. Huet, M. Goethals, J. Finidori, 6. Jongstra-Bilen, J., P. A. Janmey, J. H. Hartwig, S. Galea, and J. Jongstra. 1992. J. Vandekerckhove, and D. Louvard. 1992. An actin-binding site containing a The lymphocyte-specific protein LSP1 binds to F-actin and to the cytoskeleton conserved motif of charged amino acid residues is essential for the morphogenic through its COOH-terminal basic domain. J. Cell Biol. 118:1443. effect of villin. Cell 70:81. 7. Jongstra-Bilen, J., A. J. Young, R. Chong, and J. Jongstra. 1990. Human and 21. Howard, T. H., J. Hartwig, and C. Cunningham. 1998. Lymphocyte-specific pro- mouse LSP1 genes code for high1y conserved phosphoproteins. J. Immunol. tein 1 expression in eukaryotic cells reproduces the morphologic and motile ab- 144:1104. normality of NAD 47/89 neutrophils. Blood 91:4786. 8. Jongstra, J., and M. M. Davis. 1988. Molecular genetic analysis of mouse B 22. Vandekerckhove, J. 1990. Actin-binding proteins. Curr. Opin. Cell Biol. 2:41. lymphocyte differentiation. UCLA Symp. Mol. Cell. Biol. (New Ser.) 56:261. 23. Highsmith, S. 1990. Electrostatic contributions to the binding of myosin and 9. Jongstra, J., M. E. Ittel, N. N. Iscove, and G. Brady. 1994. The LSP1 gene is myosin-MgADP to F-actin in solution. Biochemistry 29:10690. expressed in cultured normal and transformed mouse macrophages. Mol. Immu- 24. Erickson, H. P. 1989. Co-operativity in protein-protein association. J. Mol. Biol. nol. 31:1125. 206:465. 10. Li, Y., A. Guerrero, and T. H. Howard. 1995. The actin-binding protein, lym- 25. Hartwig, J. H., and D. J. Kwiatkowski. 1991. Actin-binding proteins. Curr. Opin. phocyte-specific protein 1, is expressed in human leukocytes and human myeloid Cell Biol. 3:87. and lymphoid cell lines. J. Immunol. 155:3563. 26. Matsudaira, P. 1991. Modular organization of actin crosslinking proteins. Trends 11. Spudich, J. A., and S. Watt. 1971. The regulation of rabbit skeletal muscle con- Biochem. Sci. 16:87. traction: biochemical studies of the interaction of the tropomyosin-troponin complex with actin and the proteolytic fragments of myosin. J. Biol. Chem. 246:4866. 27. Otto, J. J. 1994. Actin-bundling proteins. Curr. Opin. Cell Biol. 6:105. 12. Chia, C. P., A. L. L. Hitt, and E. J. Luna. 1991. Direct binding of F-actin to 28. Yamashiro, S., Y. Yamakita, R. Ishikawa, and F. Matsumura. 1990. Mitosis- ponticulin, an integral plasma membrane glycoprotein. Cell Motil. Cytoskeleton specific phosphorylation causes 83K nonmuscle caldesmon to dissociate from 18:164. microfilaments. Nature 344:675. Downloaded from 13. Andrews, R. K., and J. E. B. Fox. 1992. Identification of a region in the cyto- 29. Yin, H. L., K. Iida, and P. A. Janmey. 1988. Identification of a polyphosphoi- plasmic domain of the platelet membrane glycoprotein Ib-IX complex that binds nositide-modulated domain in gelsolin which binds to the sides of actin filaments. to purified actin-binding protein. J. Biol. Chem. 267:18605. J. Cell Biol. 106:805. 14. Bryan, J., M. Imai, R. Lee, P. Moore, R. G. Cook, and W.-G. Lin. 1989. Cloning 30. Yu, F.-X., P. A. Johnston, T. C. Sudhof, and H. L. Yin. 1990. GCap39, a calcium and expression of a smooth muscle caldesmon. J. Biol. Chem. 264:13873. ion- and polyphosphoinositide-regulated actin capping protein. Science 15. Hayashi, K., Y. Fujio, I. Kato, and K. Sobue. 1991. Structural and functional 250:1413. relationships between h- and l-caldesmons. J. Biol. Chem. 266:355. 31. Janmey, P., and T. Stossel. 1987. Modulation of gelsolin function by phospha- tidylinositol 4,5-bisphosphate. Nature 325:362. 16. Wang, C.-L. A., S. Xu, R. C. Lu, V. Saavedra-Alanis, and J. Bryan. 1991. Lo- http://www.jimmunol.org/ calization of the calmodulin and the actin-binding sites of caldesmon. J. Biol. 32. Lewis, S. A., I. E. Ivanov, G. H. Lee, and N. J. Cowan. 1989. Organization of Chem. 266:9166. microtubules in dendrites and axons is determined by hydrophobic zipper in 17. Zhan, Q., S. S. Wong, and C.-L. A. Wang. 1991. A calmodulin-binding peptide microtubule-associated proteins MAP2 and tau. Nature 342:498. of caldesmon. J. Biol. Chem. 266:21810. 33. Lewis, S. A., and N. Cowan. 1990. Microtubule bundling. Nature 345:674. by guest on September 26, 2021