FEBS Letters 583 (2009) 820–824

journal homepage: www.FEBSLetters.org

Rescue of the neuroblastoma mutant of the human nucleoside diphosphate kinase A/nm23-H1 by the natural osmolyte trimethylamine-N-oxide Florian Georgescauld a,1, Iulia Mocan a, Marie-Lise Lacombe b, Ioan Lascu a,* a b

Institut de Biochimie et Génétique Cellulaires, University of Bordeaux-2, CNRS, 1 Rue C. Saint-Saëns, 33077 Bordeaux Cedex, France UMRS 938 INSERM, Faculté de Médecine Saint Antoine, 27 Rue Chaligny, 75571 Paris, France

a r t i c l e

i n f o

Article history: Received 25 November 2008 Revised 6 January 2009 Accepted 22 January 2009 Available online 30 January 2009 Edited by Miguel De la Rosa Keywords: Trimethylamine-N-oxide Nm23 Molten globule Folding intermediate

a b s t r a c t The point mutation S120G in human nucleoside diphosphate kinase A, identified in patients with neuroblastoma, causes a protein folding defect. The urea-unfolded protein cannot refold in vitro, and accumulates as a molten globule folding intermediate. We show here that the trimethylamine-N-oxide (TMAO) corrects the folding defect and stimulated subunit association. TMAO also substantially increased the stability to denaturation by urea of both wild-type and S120G mutant. A non-native folding intermediate accumulated in the presence of 4.5–7 M urea and of 2 M TMAO. It was inactive, monomeric, contained some secondary structure but no tertiary structure and displayed a remarkable stability to denaturation. Ó 2009 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

1. Introduction NDPKs play a major role in the synthesis of non-adenine nucleoside (and deoxynucleoside) triphosphates [1]. In addition, mammalian NDPKs have been shown to have regulatory functions, as in cancer progression (reviewed in [2]). The S120G mutation of the human nucleoside diphosphate kinase A (product of the nm23-H1 gene) (NDPK-HA) has been identified in patients with high-grade neuroblastomas [3]. Although the mutant protein has enzymatic activity [4,5], its regulatory effects are abnormal [6–8]. We showed that the mutation caused a protein folding defect. While the wild-type enzyme refolded and formed active hexamers, S120G-NDPK-HA could not be renatured in vitro, but accumulated as a molten globule folding intermediate2 (MG) having measurable

Abbreviations: NDPK-HA, human nucleoside diphosphate kinase A (product of the nm23-H1 gene); S120G-NDPK-HA, S120G mutant of the human NDPK-HA; BisANS, 4,40 -dianilino-1,10 -binaphthyl-5,50 -disulfonic acid; K-pn, killer-of-prune; TMAO, trimethylamine-N-oxide; Ksv, the Stern–Volmer constant; MG, the molten globule folding intermediate previously described; MG*, the folding intermediate which accumulated in the presence of 4.5–7 M urea and 2 M TMAO * Corresponding author. E-mail address: [email protected] (I. Lascu). 1 Present address: Max-Planck-Institut für Biochemie, Department of Cellular Biochemistry, Am Klopferspitz 18A, D-82152 Martinsried, Germany. 2 Semantically, the molten globule folding intermediates are better described as an ensemble of conformations than as a single structure, due to their flexible nature. We will use below the MG abbreviation with this understanding.

secondary structure but no tertiary structure [5]. The crystal structures of the wild-type and S120G mutant of NDPK-HA in complex with ADP are identical within the experimental error, rms = 0.29 Å [9]. The cellular effects of the mutated protein are therefore dues to the persistence of a partially folded form of the protein in the cell. There are several other examples in the literature of proteins involved in the control of cancer progression or of their mutants which are misfolded in vitro and possible in vivo [10,11]. The study of the correction of the folding defect by a chemical chaperone is therefore of general biological and pharmaceutical interest [12,13]. We report here the rescue of the S120G mutant structure and enzymatic activity by the natural osmolyte trimethylamine-N-oxide (TMAO). The folding intermediates were often evoked in the recent literature as being essential during formation of amyloid fibrils [14,15]. We discovered that S120G mutant of the human NDPK-HA (S120G-NDPKHA) assembles to amyloid fibrils by moderate heating [24]. Understanding the effect TMAO, on stability by MG is of great interest in this new and exciting perspective. 2. Materials and methods The expression, purification, storage and enzymatic activity of wild-type NDPK-HA and its S120G mutant were reported previously [5,16]. TMAO (Sigma–Aldrich) was freed from fluorescent contaminants by treatment with 10 g active charcoal per liter of 4 M TMAO solution. CD spectra were recorded with a Jasco J-810

0014-5793/$34.00 Ó 2009 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2009.01.043

821

F. Georgescauld et al. / FEBS Letters 583 (2009) 820–824

3. Results and discussion 3.1. Reactivation of NDPK-HA and S120G-NDPK-HA in the presence of TMAO NDPK-HA is fully catalytically active in its hexameric form only. The correctly-folded native protein is almost inactive in its monomeric form [18,19]. ‘‘Reactivation” therefore reflects both correct folding and assembly of the hexamer. Fig. 1 shows the reactivation kinetics of wild-type and S120G mutant NDPK-HA, diluted in buffer after denaturation by 8 M urea. Addition of 2 M TMAO was not required for the reactivation of wild-type NDPK-HA but it did considerably increase its reactivation rate. In contrast, the S120G mutant could not be reactivated without addition of TMAO. Stimulation of S120G-NDPK-HA reactivation was very significant. TMAO therefore assisted subunit folding and association. The effect was largest at 2.0 M while higher concentrations of TMAO were inhibitory (not shown). In the presence of MgATP, where the enzyme becomes phosphorylated [16], reactivation of NDPK-HA and S120G-NDPK-HA was faster (Fig. 1). Since phosphorylation promoted correct folding, it results that TMAO stimulated subunits assembly. Analysis by size exclusion chromatography of the S120G-NDPKHA renatured in buffer demonstrated that it has a size larger than the native subunits (Fig. 2A) and bound BisANS (Fig. 2B) which is typical for the MG state. Native subunits had a size similar to myoglobin used as a marker. In the presence of TMAO, most of the protein was hexameric. Some aggregated material eluted in the void volume, which bound BisANS, suggesting partially folded aggregated species. 3.2. TMAO stabilized the NDPK-HA to denaturation by urea To gain more information about the effect of TMAO on NDPKHA and on S120G-NDPK-HA folding and assembly, the denaturation by urea and subsequent renaturation of NDPK-HA and S120G-NDPK-HA has been studied in the presence and in the ab-

70

Reactivation (% of control)

A

60 50 40 30 20 10 0 0

2

4

6 8 Time (min)

0

2

4

10

12

10

12

80

Reactivation (% of control)

B

70 60 50 40 30 20 10 0 6

8

Time (min) Fig. 1. Effect of 2 M trimethylamine-N-oxide (TMAO) on the reactivation rate of human nucleoside diphosphate kinase A (product of the nm23-H1 gene) (NDPKHA) (A) and S120G mutant of the human NDPK-HA (S120G-NDPK-HA) (B). The enzyme reactivation was followed in the absence of additives (full circles), and in the presence of 2 M TMAO (empty circles), of 1 mM ATP (empty squares) and in the presence of 1 mM ATP and 2 M TMAO (full squares). The Mg2+ concentration was 5 mM. Native hexameric enzyme was used as the control.

sence of the osmolyte. A quantitative thermodynamic analysis of the denaturation and renaturation of hexameric NDPKs can not be made since the reaction was not reversible, in the sense that denaturation and renaturation pathways were not identical [20]. Fig. 3 illustrates the effect of TMAO on the unfolding by urea and on the subsequent refolding of NDPK-HA and S120G-NDPKHA, followed by the intrinsic protein fluorescence (Fig. 3A and B) and by activity measurements (Fig. 3C and D). The persistence of

400

Fluorescence intensity (arbitrary units)

spectropolarimeter. Because both urea and TMAO strongly absorb light in the far-UV range, some measurements were performed in a 10-lm cuvette. This allowed the recording of good quality spectra up to 190 nm. Fluorescence was measured using a LS50B spectrofluorimeter (Perkin–Elmer). Fluorescence quenching by acrylamide was studied using standard methods, with data correction for dilution. The plots of I0/I versus the acrylamide concentrations were linear for the native and MG states, while a modified Stern–Volmer plot which takes into consideration the static quenching [17] fitted better the data for the folding intermediate which accumulated in the presence of 4.5–7 M urea and 2 M TMAO (MG*) and for the unfolded protein. Protein denaturation by urea was analyzed as described [5]. Briefly, native or urea-unfolded protein was diluted in 0–8 M urea solutions to a final concentration of 1 lM; and incubated for 16 h at 25 °C. The following parameters were measured: residual/acquired enzymatic activity, intrinsic protein fluorescence intensity (excitation at 295 nm, emission at 340 nm) and fluorescence intensity of 4,40 -dianilino-1,10 -binaphthyl-5,50 -disulfonic acid (BisANS) added at 1 lM (excitation at 390 nm, emission at 480 nm). Size-exclusion chromatography was performed on a Superdex 75 column equilibrated with buffer A containing 150 mM NaCl. In some experiments BisANS was mixed continuously with the effluent and its fluorescence was measured using a flow-through cuvette. All measurements were performed at 25 °C in buffer A (100 mM Tris–HCl (pH 8.0) containing 1 mM DTT), unless otherwise indicated.

A 300

200

100

0

8

B 9

10 11 12 13 14 15 Elution volume (mL)

16

Fig. 2. Size-exclusion chromatography analysis of the reactivated S120G mutant. The urea-unfolded enzyme was diluted into buffer (empty circles) or in the presence of 2 M trimethylamine-N-oxide (TMAO) (full circles), at the final concentration of 10 lg/ml. Two hundred microliters were injected into the column. (A) The protein was followed by its intrinsic fluorescence. (B) In a separate experiment the column effluent was mixed with BisANS at the final concentration of 1 lM and its fluorescence monitored at 480 nm (excitation at 370 nm). No symbols, the native enzyme; dashed line, myoglobin.

822

F. Georgescauld et al. / FEBS Letters 583 (2009) 820–824

E

C

500

700

600

500

400

300

BisANS fluorescence intensity

600 Specific activity (U/mg)

Intrinsic protein fluorescence intensity

A

500 400 300 200 100

200

0

0

1

2

3

4

5

6

1

2

200

100

3

4

5

6

7

0

1

2

[ Urea ], ( M )

[ Urea ], ( M )

D

B

3

4

5

6

7

[ Urea ], ( M )

F

600

500

500

400

300

BisANS fluorescence intensity

600

Specific activity (U/mg)

Intrinsic protein fluorescence intensity

300

0

0

7

400

500 400 300 200 100

200

0 0

1

2

3

4

5

6

7

0

1

2

[ Urea ], ( M )

3

4

[ Urea ], ( M )

5

6

7

400

300

200

100

0 0

1

2

3 4 5 [ Urea ], ( M )

6

7

Fig. 3. Effect of 2 M trimethylamine-N-oxide (TMAO) on the urea-induced unfolding and refolding of human nucleoside diphosphate kinase A (product of the nm23-H1 gene) (NDPK-HA) (A, C, E) and S120G mutant of the human NDPK-HA (S120G-NDPK-HA) (B, D, F). The followed parameters were the intrinsic protein fluorescence intensity (panels A and B), residual/acquired enzymatic activity (C, D) and the fluorescence of BisANS added to the final concentration of 1 lM (E, F). Circles indicate experiments in the absence TMAO while squares indicate the presence of 2 M TMAO. Full symbols, denaturation and empty symbols, renaturation. The lines do not have any theoretical meaning and are shown for clarity only.

the MG intermediate was monitored by the binding of BisANS (Fig. 3E and F). The denaturation and the inactivation curve were coincident with the NDPK-HA. An intermediate appeared in the transition region, as detected by BisANS binding (Fig. 3E). In the presence of 2 M TMAO the active hexamer was considerably stabilized (Fig. 3A). An inactive species accumulated in solutions of 4.5– 7.5 M urea in the presence of 2 M TMAO, which we called MG* (see next section). It unfolds at >8 M urea or 3.6 M GuHCl (not shown). Similar to NDPK-HA, the active S120G-NDPK-HA hexamer was stabilized by 2 M TMAO. The half-inactivation urea concentration increased from 2.5 to 3.5 M in the presence of 2 M TMAO (Fig. 3D). The renaturation pattern changed dramatically. The renaturation curve was non-cooperative in the absence of TMAO, as reported before [5] (Fig. 3B). The non-cooperative renaturation curve is characteristic of some, but not of all, MG folding intermediate states. In the presence of TMAO, the renaturation curve becomes cooperative. This demonstrated that the renatured mutant enzyme was native in the presence of TMAO. The native monomers could associate to active hexamers. Indeed, reactivation level of the S120G-NDPK-HA was almost 100% in the presence of 2 M TMAO while it was less than 5% in its absence (Fig. 3D). As with the wild-type enzyme, an inactive species MG* accumulated in the presence of 2 M TMAO plus >4.5 M urea. The intensity fluorescence plateau was lower than that with the wild-type protein. A tentative explanation on the lower fluorescence intensity would be that MG* was less structured species, as compared to the wild-type protein.

The substitution of a serine with a glycine destabilizes the MG* by stabilizing the unfolded state via the increase of its entropy [21]. The size-exclusion chromatographic analysis agrees with this hypothesis (see the next section). BisANS fluorescence reflecting binding to MG* was much lower in the presence of TMAO (Fig. 3F). The MG* unfolds at >8 M urea. 3.3. A non-native intermediate accumulated in the presence of urea and 2 M TMAO The denaturation curves for wild-type and S120G mutant proteins, in solutions of 4.5–7 M urea with 2 M TMAO, revealed the presence of an inactive species displaying the characteristics of a folding intermediate, MG*. This intermediate detected for both wild-type and S120 mutant proteins, possesses a secondary structure as demonstrated by the far-UV CD spectrum (Fig. 4A). However, the absence of a CD signal in the near-UV range (Fig. 4B) demonstrates that the intermediate lacks tertiary structure. Similar CD spectra were obtained for S120G mutant and wild-type NDPK-HA. The folding intermediate MG* had fluorescence properties different from the molten globule intermediate MG. When excited at 295 nm, the maximum emission fluorescence for MG* was at 347 nm; whereas for MG it was 339.5 nm. The maximum emission fluorescence values for the native and unfolded proteins were at 339 and 352 nm, respectively. The enhancement of fluorescence of BisANS upon binding MG* which was lower than that with its

823

A

1 10

A

4 4

1,5 10

4

5000

2

-1

Θ (deg cm dmol )

4

2,5 10 2 10

Intrinsic protein fluorescence (a. u.)

F. Georgescauld et al. / FEBS Letters 583 (2009) 820–824

0 -5000 4

-1 10

4

190

200

210

220

230

240

Wavelength (nm)

B

B

40 20

-1

Θ ( deg cm dm ol )

500 400 300 200 100 0 9

10

11 12 13 14 Elution volume (mL)

9

10

11 12 13 Elution volume(mL)

250

2

0 -20 -40 -60

Intrinsic protein fluorescence (a. u.)

-1,5 10

600

15

800 700 600 500 400 300 200 100 0 14

15

-80 260

270

280

290

300

Wavelength (nm) Fig. 4. Characterization of the folding intermediate which accumulated in the presence of 4.5–7 M urea and 2 M TMAO (MG*) folding intermediate by far-UV (A) and near-UV (B) circular dichroism spectra of the S120G mutant of the human NDPK-HA (S120G-NDPK-HA). The protein was analyzed without additives (full circles) or in the presence of 7 M urea (empty diamonds) or 7 M urea plus 2 M trimethylamine-N-oxide (TMAO) (full diamonds). The protein concentration was 118 lM. The optical path lengths were 10 lm (A) and 1 cm (B).

binding MG. Furthermore, the BisANS emission spectrum was maximal at 510 nm when bound to MG* but at 485 nm when bound to MG. The Stokes radius of the folding intermediate MG* in the presence of 2 M TMAO and 6 M urea was similar to that of MG, but was peaks broader with the S120G mutant (Fig. 5) suggesting that it is heterogeneous. Finally, we used acrylamide quenching experiments to measure tryptophan accessibility (Fig. 6). The plots of I0/I versus acrylamide concentrations were linear. In the presence of 6 M urea and 2 M TMAO, the Stern–Volmer constant of the intermediate species MG* (Ksv 6.9 M1) was higher than in the native protein (Ksv 4.5 M1) but much lower than that of the unfolded protein in 7 M urea (Ksv 11.1 M1). The Ksv for MG was 5.9 M1. Therefore, MG* is less compact than the native protein, but more compact than in the unfolded protein.

sence. TMAO has a dual effect: it stabilizes the native subunits with respect to molten globule state, and promotes their association to the hexamer. In addition, the osmolyte stabilizes the S120G-NDPKHA and NDPK-HA towards denaturation by urea. This was not the result of a simplification of the folding pathway by the disappearance of the folding intermediate MG. Rather, a monomeric folding intermediate MG* accumulated in the presence of 2 M TMAO and 4.5–7 M urea which has an extraordinary stability to denaturation by urea. Similar stabilization of folding intermediates have previously been reported [22,23]. The tendency of the S120G-NDPKHA to populate non-native folding intermediates is probably essential for its assembly in amyloid fibrils [24].

3

2.5

0

4. Conclusion

Fig. 5. Size-exclusion chromatographic analysis of the folding intermediate MG*. The human nucleoside diphosphate kinase A (product of the nm23-H1 gene) (NDPK-HA) (A) and S120G mutant of the human NDPK-HA (S120G-NDPK-HA) (B), native (full symbols) or unfolded in 9 M urea (empty symbols), were injected into a Superdex 75 column equilibrated with buffer A containing 0.2 M NaCl (circles) or buffer A containing 0.2 M NaCl plus 6 M urea and 2 M trimethylamine-N-oxide (TMAO) (squares). The elution profile of myoglobin (of same size as the NDKA subunit) is shown in panel A without symbols. The protein was detected by fluorescence (excitation at 280 nm, emission at 340 nm).

IF /IF

250

TMAO corrects the folding defects of several mutated proteins of medical importance: mutant branched-chain a-ketoacid decarboxylase [10]; a1-anti-trypsin [11,22]; defective aquaporin-2 trafficking in nephrogenic diabetes insipidus [11]; and partial restoration of defective chloride conductance in DeltaF508 CF mice [23]. Here we showed that TMAO corrects the folding defect of the S120G mutant protein. It has been suggested TMAO stabilizes the native state because its unfavorable interaction with the peptide bonds of the main chain of the protein in the unfolded state. The urea-unfolded S120G-NDPK-HA may refold and associate into active hexamers in the presence of 2 M TMAO but not in its ab-

2

1.5

1 0

0.05

0.1

0.15

0.2

[Acrylamide], (M) Fig. 6. Quenching of the intrinsic protein fluorescence by acrylamide, of the S120G mutant of the human NDPK-HA (S120G-NDPK-HA). Excitation was at 295 nm. Empty squares, native protein; full circles, protein in 2 M trimethylamine-N-oxide (TMAO) and 6 M urea; full squares; protein refolded in buffer A; triangles, protein unfolded in 6 M urea. The lines correspond to the best fit described in the text.

824

F. Georgescauld et al. / FEBS Letters 583 (2009) 820–824

Acknowledgments We are grateful to Franziska Schneider, Philippe Gonin, and Filip Yabukarski for performing some experiments and to Drs. Angela de Otero, Anna Giartosio, and Alain Dautant for careful reading the manuscript and useful help. References [1] Lascu, I. and Gonin, P. (2000) The catalytic mechanism of nucleoside diphosphate kinases. J. Bioenergy Biomembr. 32, 237–246. [2] Ouatas, T., Salerno, M., Palmieri, D. and Steeg, P.S. (2003) Basic and translational advances in cancer metastasis: Nm23. J. Bioenergy Biomembr. 35, 73–79. [3] Chang, C.L., Zhu, X.X., Thoraval, D.H., Ungar, D., Rawwas, J., Hora, N., Strahler, J.R., Hanash, S.M. and Radany, E. (1994) Nm23-H1 mutation in neuroblastoma. Nature 370, 335–336. [4] Chang, C.L., Strahler, J.R., Thoraval, D.H., Qian, M.G., Hinderer, R. and Hanash, S.M. (1996) A nucleoside diphosphate kinase A (nm23-H1) serine 120-glycine substitution in advanced stage neuroblastoma affects enzyme stability and alters protein-protein interaction. Oncogene 12, 659–667. [5] Lascu, I., Schaertl, S., Wang, C., Sarger, C., Giartosio, A., Briand, G., Lacombe, M.L. and Konrad, M. (1997) A point mutation of human nucleoside diphosphate kinase A found in aggressive neuroblastoma affects protein folding. J. Biol. Chem. 272, 15599–15602. [6] Kim, Y.I., Park, S., Jeoung, D.I. and Lee, H. (2003) Point mutations affecting the oligomeric structure of Nm23-H1 abrogates its inhibitory activity on colonization and invasion of prostate cancer cells. Biochem. Biophys. Res. Commun. 307, 281–289. [7] Forus, A., Zollo, M., et al. (2001) Amplification and overexpression of PRUNE in human sarcomas and breast carcinomas – a possible mechanism for altering the nm23-H1 activity. Oncogene 20, 6881–6890. [8] Almgren, M.A., Henriksson, K.C., Fujimoto, J. and Chang, C.L. (2004) Nucleoside diphosphate kinase A/nm23-H1 promotes metastasis of NB69-derived human neuroblastoma. Mol. Cancer Res. 2, 387–394. [9] Giraud, M.F., Georgescauld, F., Lascu, I. and Dautant, A. (2006) Crystal structures of S120G mutant and wild type of human nucleoside diphosphate kinase A in complex with ADP. J. Bioenergy Biomembr. 38, 261–264. [10] Song, J.L. and Chuang, D.T. (2001) Natural osmolyte trimethylamine-N-oxide corrects assembly defects of mutant branched-chain alpha-ketoacid decarboxylase in maple syrup urine disease. J. Biol. Chem. 276, 40241– 40246.

[11] Tamarappoo, B.K. and Verkman, A.S. (1998) Defective aquaporin-2 trafficking in nephrogenic diabetes insipidus and correction by chemical chaperones. J. Clin. Invest. 101, 2257–2267. [12] Morello, J.P., Petaja-Repo, U.E., Bichet, D.G. and Bouvier, M. (2000) Pharmacological chaperones: a new twist on receptor folding. Trends Pharmacol. Sci. 21, 466–469. [13] Papp, E. and Csermely, P. (2006) Chemical chaperones: mechanisms of action and potential use. Handbook Exp. Pharmacol., 405–416. [14] Zhu, L., Zhang, X.J., Wang, L.Y., Zhou, J.M. and Perrett, S. (2003) Relationship between stability of folding intermediates and amyloid formation for the yeast prion Ure2p: a quantitative analysis of the effects of pH and buffer system. J. Mol. Biol. 328, 235–254. [15] Jahn, T.R., Parker, M.J., Homans, S.W. and Radford, S.E. (2006) Amyloid formation under physiological conditions proceeds via a native-like folding intermediate. Nat. Struct. Mol. Biol. 13, 195–201. [16] Mocan, I., Georgescauld, F., Gonin, P., Thoraval, D., Cervoni, L., Giartosio, A., Dabernat-Arnaud, S., Crouzet, M., Lacombe, M.L. and Lascu, I. (2007) Protein phosphorylation corrects the folding defect of the neuroblastoma (S120G) mutant of human nucleoside diphosphate kinase A/Nm23-H1. Biochem. J. 403, 149–156. [17] Lakowicz, J.R. (1999) Principles of Fluorescence Spectroscopy, 2nd edn, Kluwer Academic/Plenum Publishers, New York. [18] Erent, M., Gonin, P., Cherfils, J., Tissier, P., Raschella, G., Giartosio, A., Agou, F., Sarger, C., Lacombe, M.L., Konrad, M. and Lascu, I. (2001) Structural and catalytic properties and homology modelling of the human nucleoside diphosphate kinase C, product of the DRnm23 gene. Eur. J. Biochem. 268, 1972–1981. [19] Mesnildrey, S., Agou, F., Karlsson, A., Bonne, D.D. and Veron, M. (1998) Coupling between catalysis and oligomeric structure in nucleoside diphosphate kinase. J. Biol. Chem. 273, 4436–4442. [20] Lascu, L., Giartosio, A., Ransac, S. and Erent, M. (2000) Quaternary structure of nucleoside diphosphate kinases. J. Bioenergy Biomembr. 32, 227–236. [21] Alber, T. (1986) Mutational effects on protein stability. Ann. Rev. Biochem. 48, 765–798. [22] Devlin, G.L., Parfrey, H., Tew, D.J., Lomas, D.A. and Bottomley, S.P. (2001) Prevention of polymerization of M and Z alpha1-antitrypsin (alpha1-AT) with trimethylamine-N-oxide. Implications for the treatment of alpha1-at deficiency. Am. J. Respir. Cell Mol. Biol. 24, 727–732. [23] Fischer, H., Fukuda, N., Barbry, P., Illek, B., Sartori, C. and Matthay, M.A. (2001) Partial restoration of defective chloride conductance in DeltaF508 CF mice by trimethylamine oxide. Am. J. Physiol. Lung Cell Mol. Physiol. 281, L52–L57. [24] Georgescauld, F., Sabaté, R., Espargaró, A., Ventura, S., Chaignepain, S., Lacombe, M.-L. Lascu, I. Aggregation of the human metastasis suppressor nucleoside diphosphate kinase A/Nm23-H1 into amyloid fibrils, in preparation.