Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. 1997 Oct 14; 94(21): 11623–11626.

Bordetella bronchiseptica dermonecrotizing toxin induces reorganization of actin stress fibers through deamidation of Gln-63 of the GTP-binding protein Rho


Bordetella dermonecrotizing toxin causes assembly of actin stress fibers and focal adhesions in some cultured cells and induces mobility shifts of the small GTP-binding protein Rho on electrophoresis. We attempted to clarify the molecular basis of the toxin action on Rho. Analysis of the amino acid sequence of toxin-treated RhoA revealed the deamidation of Gln-63 to Glu. The substitution of Glu for Gln-63 of RhoA by site-directed mutagenesis caused a mobility shift on electrophoresis, which was indistinguishable from that of the toxin-treated RhoA. Neither mutant RhoA-bearing Glu-63 nor toxin-treated RhoA significantly differed from untreated wild type RhoA in guanosine 5′-[γ-thio]triphosphate binding activity but both showed a 10-fold reduction in GTP hydrolysis activity relative to untreated RhoA. C3H10T1/2 cells transfected with cDNA of the mutant RhoA bearing Glu-63 showed extensive formation of actin stress fibers similar to the toxin-treated cells. These results indicate that the toxin catalyzes deamidation of Gln-63 of Rho and renders it constitutively active, leading to formation of actin stress fibers.

Dermonecrotizing toxin (DNT) produced by bacteria of the Bordetella species has lethal, dermonecrotic, and splenoatrophic activities (13). DNT is considered to be one of the virulence factors responsible for turbinate atrophy in swine atrophic rhinitis and has been suggested to damage bone tissues in turbinates (4, 5). Using osteoblast-like MC3T3-E1 cells, we demonstrated that DNT changes cell morphology and inhibits the differentiation into osteoblasts (6). The morphological changes in DNT-treated cells were accompanied by the assembly of actin stress fibers and focal adhesions, indicating anomalous activation of the small GTP-binding protein Rho (7). It was also found that Rho in the lysate from DNT-treated cells showed shifts in its electrophoretic mobility. Similar mobility shifts of Rho were observed by in vitro treatment of recombinant Rho with DNT. These results indicate that DNT covalently modifies and activates Rho, leading to the assembly of actin stress fibers and focal adhesions. However, the nature of the DNT-induced modification of Rho, which relates to its activation remains unknown.

Here, we show that DNT catalyzes the deamidation of Gln-63 of RhoA and converts it into Glu. We conclude that the deamidated RhoA bearing Glu-63 becomes constitutively active and stimulates the formation of actin stress fibers in DNT-treated cells.


Construction and Expression of the Small GTP-Binding Proteins.

Bovine RhoA cDNA, which was inserted into the NdeI–BamHI site of pET3a (Novagen), was provided by Y. Nemoto (Department of Cell Biology, Yale University School of Medicine, New Haven, CT). Bovine RhoA cDNA encodes an amino acid sequence identical with that of human RhoA. The BamHI–XbaI fragment of pET3a-RhoA was subcloned into pBluescript SK (Stratagene), digested with BamHI and XbaI (pBS-RhoA). A mutant gene of RhoA encoding Glu-63 (RhoAGlu-63) was generated from pBS-RhoA by site-directed mutagenesis with a transformerTM site-directed mutagenesis kit (CLONTECH) and subcloned into the BamHI–XbaI site of pET3a. I.M.A.G.E. Consortium cDNA clones 187804 and 196172 showing homologies to Rac1 and Cdc42Hs, respectively, were identified by searching the dbEST data base and were purchased from Research Genetics (Huntsville, AL). cDNAs of Rac1 and Cdc42Hs were amplified by PCR with 5′ primers containing an NdeI site on the 5′ side of the start codon and 3′ primers containing a BamHI site on the 3′ side of the stop codon. Amplified DNAs were treated with NdeI and BamHI and inserted into NdeI–BamHI sites of pET3a. pET3a containing the RhoA, RhoAGlu-63, Rac1, or Cdc42Hs gene was introduced into Escherichia coli BL21(DE3). The recombinant proteins were produced in the bacteria in the presence of 1 mM isopropylthiogalactoside and extracted by sonic treatment. The bacterial lysates containing RhoA, RhoAGlu-63, and Cdc42Hs dialyzed against 20 mM Tris[center dot]HCl, pH 8.0, 10 mM MgCl2, 1 mM EDTA, and 5 mM DTT were applied to a DEAE Sepharose CL-6B column (1.6 × 20 cm, Pharmacia) equilibrated with the same buffer. The absorbed materials were eluted with a linear gradient of 0 to 0.5 M NaCl. The GTP-binding proteins were eluted with ≈0.1 M NaCl as a single protein peak. The lysate containing Rac1 dialyzed against 66 mM Na2HPO4-33 mM citric acid buffer (pH 5.0) were applied to an SP Toyopearl 650M column (1.6 × 20 cm, Tosoh, Tokyo) equilibrated with the same buffer. The absorbed materials were eluted with a linear gradient of NaCl from 0 to 0.5 M in the same buffer and Rac1 appeared as a single peak at 0.08–0.1 M NaCl.

An amino-terminal FLAG-tagged RhoA gene was constructed as follows. The start codons of RhoA and RhoAGlu-63 genes were replaced with a SalI site. A gene encoding FLAG-tagged PIG-A (8) was subcloned into the XhoI–XbaI site of pMEPyori18Sf (9). The PIG-A gene was removed by digestion with SalI and XbaI and replaced by the RhoA or RhoAGlu-63 gene. Untagged RhoA and RhoAGlu-63 genes were amplified by PCR with a 5′ primer containing Kozak’s consensus sequence (10) on the 5′ side of the start codon and a 3′ primer. Amplified DNA was inserted into pBluescript SK treated with EcoRV and alkaline phosphatase. XhoI–XbaI fragments of pBluescript SK containing the RhoA or RhoAGlu-63 gene were subcloned into pMEPyori18sf treated with XhoI and XbaI. C3H10T1/2 cells were transfected with the tagged- or untagged-cDNA samples by the calcium phosphate technique. The sequences of the constructed DNAs were confirmed by nucleotide sequencing.

Treatment of RhoA with DNT and Amino Acid Sequencing.

DNT was purified as described (11). The purity of the toxin was confirmed by SDS/PAGE with silver staining. The recombinant RhoA was incubated with DNT at a molar ratio of 100:1 in 20 mM Tris-HCl, pH 7.5, 10 mM MgCl2, 1 mM EDTA, and 5 mM DTT at 37°C overnight and then digested with 1/100 molar staphylococcal V8 protease or lysyl endopeptidase. Proteolytic fragments were isolated by reversed-phase HPLC on a Cosmosil 5C18-AR 300 column (Nacalai Tesque, Kyoto) with a linear gradient of 5–80% acetonitrile in 0.09% trifluoroacetic acid. Isolated peptides were sequenced with an Applied Biosystems 492 sequencer.

Determinations of Guanosine 5′-[γ-thio]triphosphase GTP[γS] Binding and GTP Hydrolysis Activities.

GTP[γS]-binding activity was measured as follows. Samples of 42 pmol of RhoA were incubated with [35S]GTP[γS] (20 Ci/mmol; 1 Ci = 37 GBq) at various concentrations in 50 μl of 20 mM Tris[center dot]HCl (pH 8.0), 100 mM NaCl, 5 mM MgCl2, 1 mM EDTA, and 5 mM DTT at 30°C for 2 h. Then the samples were filtered through a nitrocellulose membrane (0.45 μm, Sartorius) pretreated with washing buffer (20 mM Tris[center dot]HCl, pH 8.0/100 mM NaCl/25 mM MgCl2) containing 0.1 mM GTP. The filters were washed five times with the washing buffer and dried, and their radioactivities were determined with a scintillation counter. GTP hydrolysis activity was determined by incubating samples of RhoA at 0.8 μM in 20 mM Tris[center dot]HCl (pH 7.5) 5 mM MgCl2, 5 mM EDTA, 2.5 mM DTT, and 1 μM [γ-32P]GTP (2 Ci/mmol) at 37°C for various periods. After incubation, 25-μl aliquots of the reaction mixtures were mixed with 275 μl of ice-cold 50 mM NaH2PO4 and 5% charcoal Norit SX-II, chilled on ice for 15 min, and centrifuged at 10,000 × g for 10 min. The radioactivities in 150 μl of the supernatants were then determined with a scintillation counter.

Preparation of Antibody Recognizing the Deamidated GTP-Binding Proteins.

A synthetic peptide shown in Fig. 5a was conjugated to keyhole limpet hemocyanin with m-maleimidobenzoyl-N-hydroxysuccinimide ester. The peptide-keyhole limpet hemocyanin conjugate (250 μg) was emulsified in incomplete Freund’s adjuvant or Ribi adjuvant System R-730 (Ribi Immunochem) and subcutaneously injected three times into rabbits at 21-day intervals. Antibody was purified with Affi-Gel protein A MAPS II kit (Bio-Rad) from serum obtained from the immunized rabbit.


For Western immunoblot analysis, samples separated by SDS/PAGE were transferred to poly(vinylidene difluoride) membranes (Immobilon, Millipore). The membranes were probed with appropriate rabbit antibodies and alkaline phosphatase-conjugated anti-rabbit IgG (Organon Teknika–Cappel). Immunoblots were revealed by CDP-Star reagent (Tropix, Bedford, MA).

Protein amounts were determined by Lowry’s (12) or Bradford’s (13) methods. Statistical significance of data were examined by unpaired and two-tailed Student’s t tests.


To identify the modification of RhoA, we digested DNT-treated RhoA with staphylococcal V8 protease or lysyl endopeptidase, separated the digests by reversed-phase HPLC, and determined the amino acid sequences of the isolated peptides. On analysis of 152 amino acid residues of RhoA (total amino acid residues, 193), we found that Gln-63 was converted to Glu in the modified RhoA. To determine whether the substitution of Gln-63 for Glu results in the same electrophoretic mobility shift as that seen in DNT-treated RhoA, we made a mutant RhoA, RhoAGlu-63, by site-directed mutagenesis, and subjected it to SDS/PAGE. DNT-treated RhoA migrated slower than untreated RhoA as described before (ref. 7, Fig. Fig.11a, lanes 1, 3, and 6). RhoAGlu-63 migrated at the same rate as DNT-treated RhoA (Fig. (Fig.11a, lanes 2 and 3). A mixture of RhoAGlu-63 and DNT-treated RhoA appeared as a single band (Fig. (Fig.11a, lane 5), indicating that they behave as a single molecular species on electrophoresis. RhoAGlu-63 was insensitive to DNT on the basis of the mobility shift (Fig. (Fig.11a, lane 4). These results suggest that in vitro the modification of RhoA by DNT is a deamidation, hydrolysis of the γ-amide of Gln-63. To determine whether RhoAGlu-63 expressed in eukaryotic cells results in similar shifts in electrophoretic mobility, we analyzed lysates of C3H10T1/2 cells transfected with RhoA or RhoAGlu-63 by SDS/PAGE. Rho proteins were located by autoradiography after [32P]ADP ribosylation by Clostridium botulinum C3, a Rho-specific ADP ribosyltransferase (14, 15). The lysates from RhoA- and RhoAGlu-63-transfected cells showed a single and double band, respectively. Of the two bands detected, that with slower mobility was identified as RhoAGlu-63 because it appeared only in the RhoAGlu-63-transfected cells (Fig. (Fig.11b, lanes 1 and 2). A similar slow-moving band was obtained from a DNT-treated lysate of untreated cells and an untreated lysate of DNT-treated cells (Fig. (Fig.11b, lanes 4 and 5), suggesting that Rho proteins in eukaryotic cells were deamidated in vivo as well as in vitro by treatment with DNT. The faster moving band was seen when the cells or the lysate was treated with DNT (Fig. (Fig.11b, lanes 4–6). When the lysate of DNT-treated cells was treated further with DNT, the intensity and mobility of the faster moving band were unchanged whereas a band at a position of normal Rho decreased with increase in intensity of the slower moving band (Fig. (Fig.11b, lanes 5 and 6). These results suggest that Rho of the faster moving band probably had undergone deamidation and was no longer sensitive to DNT. The lysate of DNT-treated cells showed a strong faster moving band (Fig. (Fig.11b, lanes 5 and 6). In contrast, in vitro treatment of a lysate of normal cells with DNT resulted in the appearance of the slower moving band and scarcely any detectable faster moving band (Fig. (Fig.11b, lane 4). Therefore, we consider that some of the deamidated Rho receives further processing in intact cells, which yields the faster moving Rho. We are now attempting to identify the nature of the modification that probably follows or coincides with the deamidation and produces the faster moving Rho.

Figure 1
Mobility shifts of Rhos on SDS/PAGE. Bands were located by silver staining (a) or autoradiography (b) after electrophoresis. (a) Various recombinant RhoA samples treated or untreated with DNT and subjected to electrophoresis. (b, Left) C3H10T1/2 ...

Gln-63 of RhoA is located in a domain that is highly conserved among members of the Ras superfamily, and is considered to be a key residue in GTP hydrolysis (16, 17). Previous studies have shown that substitution for Gln-61 in Ras results in an 8- to 10-fold reduction in the intrinsic GTP hydrolysis activity (18, 19) and abnormal electrophoretic mobilities (18, 20, 21). Therefore, we examined this point. The rates of GTP hydrolysis by DNT-treated RhoA and RhoAGlu-63 were 10-fold less than those by untreated RhoA (Fig. (Fig.22 Upper). The GTP[γS] binding activities were also compared (Fig. (Fig.22 Lower). Although DNT-treated RhoA and RhoAGlu-63 tended to show lower GTP[γS] binding activities than RhoA, their Kd values and the maximum amounts of the bound nucleotides were not significantly different from those of RhoA (P > 0.05, when both the Kd values and the maximum binding amount of GTP[γS] of RhoA were compared with those of DNT-treated RhoA and RhoAGlu-63). The deamidated RhoA, which reduced GTP hydrolyzing but not GTP[γS] binding activity, must remain in the GTP-bound form and continuously transmit the positive signal downstream of the pathways. This should lead to extensive formation of actin stress fibers in cells intoxicated by DNT. Crystal structure analyses of Ras have shown that the conformation of the domain in which Gln-61, corresponding to Gln-63 of Rho, resides changes markedly upon GDP-GTP replacement (16). This may explain our previous observation that RhoA in the GDP, but not the GTP[γS]-bound form, is sensitive to DNT.

Figure 2
GTP hydrolysis (Upper) and GTP[γS]-binding (Lower) activities of control RhoA (○), DNT-treated RhoA (•), and RhoAGlu-63 ([open triangle]). Three independent experiments were performed and representative data are shown. ...

To determine whether the deamidated Rho contributes to the formation of actin stress fibers, we transfected C3H10T1/2 cells with FLAG-tagged RhoAGlu-63 and examined stress fiber formation. Transfection with RhoAGlu-63 resulted in a marked increase in stress fiber formation similar to that seen in untransfected cells treated with DNT (Fig. (Fig.33 B and F). Transfection with cDNA of wild type RhoA also induced the formation of actin stress fibers, but was less effective than that with RhoAGlu-63 cDNA (Fig. (Fig.33D). Transfection of mutant RhoA bearing Leu-63 has been shown to stimulate stress fiber formation (23) and a variety of mutant proteins of GTPases such as Ras, Rac, and Cdc42 bearing Leu instead of Gln corresponding to Gln-63 of Rho have been used as constitutively active analogs (23, 24). Our results show that RhoAGlu-63, RhoA modified by DNT, also functions as a constitutively active analog like the mutants bearing Leu-63. It was reported that the substitutions in Gln61 of Ras made it insensitive to the GTPase-activating protein (GAP), which strongly stimulates the intrinsic GTP hydrolyzing activity of the GTP-binding protein (19, 25). Therefore, the insensitivity to GAP may contribute to the constitutive activation of RhoAGlu-63 as well as the reduction of the intrinsic GTP hydrolyzing activity.

Figure 3
Formation of actin stress fibers in DNT-treated and RhoAGlu-63-transfected C3H10T1/2 cells. Actin fibers are stained with rhodamine-phalloidin (A, B, D, and F). Transfected cells were recognized by immunostaining with anti-FLAG M2 antibody (Eastman ...

The Rho subfamily includes RhoA, RhoB, RhoC, Rac1, Rac 2, and Cdc42Hs. Because Gln-63, the target of DNT, resides in a highly conserved domain in members of the Rho subfamily, it is possible that DNT deamidates Gln equivalent to the Gln-63 of RhoA, in the other members. To examine this, we prepared antibody against a synthetic peptide, which corresponds to residues 59 to 73 of RhoA but has Glu-63 instead of Gln (Fig. (Fig.44a). The obtained antibody, designated as anti 63E, recognized DNT-treated RhoA and RhoAGlu-63 but not RhoA (Fig. (Fig.44b, lanes 1–3). Rac1 and Cdc42Hs were also recognized by anti 63E but only after treatment with DNT, indicating that Rac1 and Cdc42Hs are deamidated by DNT (Fig. (Fig.44b, lanes 4–7). It remains unclear whether DNT catalyzes the deamidation of GTP-binding proteins belonging to other subfamilies. An analysis of amino acid sequence is required to detect the deamidation of such proteins, because anti-63E is available only for members of the Rho subfamily, which share a common sequence in the domain containing Gln to be deamidated. In conclusion, we demonstrate that DNT deamidates Gln-63 of Rho, converting it to Glu. This results in the constitutive activation of Rho and eventually the formation of actin stress fibers. Two kinds of enzymes are known to catalyze the deamidation of the γ-amido of protein-bound Gln: peptidyl-glutaminases (26, 27), and transglutaminases, which cause deamidation of Gln only in the absence of a suitable primary amine as an acyl acceptor (28). It is still not clear to which category of enzyme DNT belongs, but it should be recognized as a novel toxin that elicits its cellular effects by enzymatically causing amino acid substitutions in GTP-binding proteins.

Figure 4
Detection of DNT-catalyzed deamidation of RhoA, Rac1, and Cdc42Hs by anti-63E. (a) Sequence alignment of the synthetic peptide (antigen) and the members of the Rho subfamily. Asterisks indicate identical residues among Rho, Rac, and Cdc42 proteins. The ...


We thank Dr. T. Kinoshita for critical comments on the manuscript, Dr. Y. Nemoto for supplying cDNA of RhoA, Dr. Y. Maeda for supplying C3H10T1/2 cells and helpful discussion and suggestions, and Dr. S. Kozaki for supplying C3 exoenzyme.



After this manuscript had been completed and submitted for publication, two groups (31, 32) reported that E. coli cytotoxic necrotizing factor 1 (CNF1) deamidates Gln-63 of Rho (29, 30). It is worthy of further study to clarify the structure-function relationship of CNF1 and DNT, because these toxins were reported to show homology in their C-terminal regions.


This paper was submitted directly (Track II) to the Proceedings Office.

Abbreviations: DNT, dermonecrotizing toxin; GTP[γS], guanosine 5′-[γ-thio]triphosphase.


1. Horiguchi Y, Nakai T, Kume K. Microb Pathog. 1989;6:361–368. [PubMed]
2. Horiguchi Y, Matsuda H, Koyama H, Nakai T, Kume K. FEMS Microbiol Lett. 1992;90:229–234. [PubMed]
3. Iida T, Okonogi T. J Med Microbiol. 1971;4:51–61. [PubMed]
4. Hanada M, Shimoda K, Tomita S, Nakase Y, Nishiyama Y. Jpn J Vet Sci. 1979;41:1–8.
5. Horiguchi Y, Okada T, Sugimoto N, Morikawa Y, Katahira J, Matsuda M. FEMS Immunol Med Microbiol. 1995;12:29–32. [PubMed]
6. Horiguchi Y, Nakai T, Kume K. Infect Immun. 1991;59:1112–1116. [PMC free article] [PubMed]
7. Horiguchi Y, Senda T, Sugimoto N, Katahira J, Matsuda M. J Cell Sci. 1995;108:3243–3251. [PubMed]
8. Watanabe R, Kinoshita T, Masaki R, Yamamoto A, Takeda J, Inoue N. J Biol Chem. 1996;271:26868–26875. [PubMed]
9. Ohishi K, Kurimoto Y, Inoue N, Endo Y, Takeda J, Kinoshita T. Genomics. 1996;34:340–346. [PubMed]
10. Kozak M. Nucleic Acids Res. 1987;15:8125–8148. [PMC free article] [PubMed]
11. Horiguchi Y, Nakai T, Kume K. FEMS Microbiol Lett. 1990;66:39–43. [PubMed]
12. Lowry O H, Rosebrough N J, Farr A L, Randall R J. J Biol Chem. 1951;193:265–275. [PubMed]
13. Bradford M M. Anal Biochem. 1976;72:248–254. [PubMed]
14. Ridley A J, Hall A. Cell. 1992;70:389–399. [PubMed]
15. Sekine A, Fujiwara M, Narumiya S. J Biol Chem. 1989;264:8602–8605. [PubMed]
16. Bourne H R, Sanders D A, McCormick F. Nature (London) 1991;349:117–127. [PubMed]
17. Hirshberg M, Stockley R W, Dodson G, Webb M R. Nat Struct Biol. 1997;4:147–152. [PubMed]
18. Der C J, Finkel T, Cooper G M. Cell. 1986;44:167–176. [PubMed]
19. Krengel U, Schlichting I, Scherer A, Schumann R, Frech M, John J, Kabsch W, Pai E F, Wittinghofer A. Cell. 1990;62:539–548. [PubMed]
20. Srivastava S K, Yuasa Y, Reynolds S H, Aaronson S A. Proc Natl Acad Sci USA. 1985;82:38–42. [PMC free article] [PubMed]
21. Temeles G T, Gibbs J B, D’Alonzo J S, Sigal I S, Scolnick E M. Nature (London) 1985;313:700–703. [PubMed]
22. Self A J, Hall A. Methods Enzymol. 1995;256:3–10. [PubMed]
23. Khosravi-far R, Chrzanowska-Wodnicka M, Solski P A, Eva A, Burridge K, Der C J. Mol Cell Biol. 1994;14:6848–6857. [PMC free article] [PubMed]
24. Lamarche N, Tapon N, Stowers L, Burbelo P D, Aspenström P, Bridges T, Chant J, Hall A. Cell. 1996;87:519–529. [PubMed]
25. John J, Frech M, Wittinghofer A. J Biol Chem. 1988;263:11792–11799. [PubMed]
26. Kikuchi M, Hayashida H, Nakano E, Sakaguchi K. Biochemistry. 1971;10:1222–1229. [PubMed]
27. Vaintraub I A, Kotova L V, Shaha R. FEBS Lett. 1992;302:169–171. [PubMed]
28. Folk J E, Finlayson J S. Adv Protein Chem. 1977;31:1–133. [PubMed]
29. Schmidt G, Sehr P, Wilm M, Selzer J, Mann M, Aktories K. Nature (London) 1997;387:725–729. [PubMed]
30. Flatau G, Lemichez E, Gauthier M, Chardin P, Paris S, Fiorentini C, Boquet P. Nature (London) 1997;387:729–733. [PubMed]
31. Walker K E, Weiss A A. Infect Immun. 1994;62:3817–3828. [PMC free article] [PubMed]
32. Lemichez E, Flatau G, Bruzzone M, Boquet P, Gauthier M. Mol Microbiol. 1997;24:1061–1070. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...