Homology Modelling of the small sialidase from Clostridium perfringens: Use of Laser photo CIDNP (chemically induced dynamic nuclear polarization) techniques and site-directed mutagenesis to ascertain the reliability of the modeled structures


Emadeddin Tajkhorshid,1,2 Hans Christian Siebert,3,4 Claus Wilhelm von der Lieth,2 Reinhard G. Kleineidam,5 Susanne Kruse,6 Roland Schauer,6 Robert Kaptein,4 Hans-Joachim Gabius3 and Johannes F. G. Vliegenthart4

  1. School of Pharmacy, Tehran Univ. Med. Sci., Tehran, Iran
  2. German Cancer Research Center; DKFZ, Central Spectroscopy, Heidelberg, Germany
  3. Institut für Physiologische Chemie der Ludwig-Maximilians-Universität, München, Germany;
  4. Bijvoet Center for Biomolecular Research, University of Utrecht, The Netherland;
  5. Biochemisch Laboratorium and Bioson Research Institute, University of Groningen, The Netherland;
  6. Biochemisches Institut, Christian-Albrechts Universität, Kiel, Germany
Correspondence to:
Presented as a non-permanent presentation in 2nd Electronic Glycoscience Conference

INTRODUCTION

Sialidases (N-acylneuraminosyl-glycohydrolases, EC 3.2.1.18) hydrolytically cleave alpha-glycosidically bound sialic acids, derivatives of the amino sugar neuraminic acid (Fig. 1). Sialic acids are mostly found as terminal constituents of oligosaccharides, glycoproteins and glycolipids in higher animals. The sialidases, however, are widely distributed not only throughout the metazoan animals of the deuterostomate lineage, but also among protozoa, viruses, fungi and bacteria, most of which are unable to produce sialic acids by themselves. Remarkably, the enzyme is often produced by microorganisms, which live in close contact with an animal host, whereby the enzyme may serve as a pathogenicity factor, or as an important tool for processing of nutrients [1, 2].

The alignment of eight bacterial sialidases indicates that they belong to one enzyme family [3]. The complete conservation of 25 amino acids in these enzymes gave reason to suspect that they are of functional or structural importance. The tertiary structures of three bacterial sialidases, from Salmonella typhimurium [4], Vibrio cholerae [5] and Micromonospora viridifaciens [6], have been elucidated so far. Regarding high similarity between these enzymes, one can construct a model of three dimensional structure of Clostridium perfringen sialidase which can be used for more rational selection of the positions for site-directed mutagenesis studies as a valuable tool in order to elucidate the molecular aspects of mechanism of sialidase activity [7].

A special NMR method can be used to test the reliability of the modelling approach. The side chains of tyrosine, tryptophan and histidine are able to produce CIDNP (Chemically Induced Dynamic Nuclear Polarization) signals after laser irradiation in the presence of a suitable radical pair-generating dye [8, 9, 10]. Therefore concomitant experimental approach using CIDNP technique results in valuable information about the conformation of particular amino acids. The calculation of surface accessibilities of respective types of side chains, which is possible by applying Connolly surface calculation on 3D model, enables a detailed comparison of these data with, and the interpretation of CIDNP results. This kind of combined studies can be used in prediction of possible conformational changes due to different protein modification such as glycosylation, phosphorylation and/or mutation.
Return to the top of this document ...

MATERIALS AND METHODS

A. Site-directed mutagenesis and purification

Site-directed mutagenesis was performed by using the Sculptor TM in vitro mutagenesis system (Amersham, Braunschweig) and primers (phosphorylated oligonucleotides), including the sequence for the amino acids selected for mutations. Based on the three dimensional models, especially the one of Salmonella typhimurium sialidase, analogeous amino acids were selected for mutations of the "small" Clostridium perfringens enzyme.

The single stranded template (5'->3'-strand) of the "wild type" sialidase structural gene was obtained by first amplifying the original gene [11] in PCR and specific primers introducing Bam HI-cleavage sites at both ends in order to facilitate cloning. The PCR-product of 1.16 kb was inserted into the phage vector M13mp19 (Boehringer, Mannheim), single stranded DNA was prepared according to [12]. The appropriate strand was selected by autoradiography after overnight hybridization with the 32P-labeled 5'->3'-PCR primer. After confirming the successful mutations by DNA-sequencing [13], double stranded DNA of the phages was prepared from the host cells [12]. The sialidase gene was then inserted into the expression vector pQE-10 (QIAGEN, Hilden) and E. coli BL21(DE3) pLysS [14], a strain with reduced protease activity, was transformed with the construct. This system added a six histidine (His6) affinity tag to the N-terminus of the protein, thus allowing the purification of the sialidase by affinity chromatography on Ni-nitrilo-triacetic acid agarose (Ni-NTA agarose; QIAGEN).

Return to the top of this document ...

B. Characterization of the sialidase enzymes

Sialidase activity was determined according to [15]. In standard assays, 10 l of enzyme solution was incubated with 80 l of 0.1 M sodium acetate buffer, pH 6.1, and 10 l of 1 mM methylum-belliferyl-a-D-N-acetylneuraminic acid (MU-Neu5Ac) for 10 min at 37°C. Free MU was measured in a fluorimeter using excitation at 365 nm and emission at 450 nm. For determination of the kinetic properties and pH-optimum curves, 20 l of enzyme solution were mixed with 45 l water and 25 l of a three buffer system containing 0.2 M acetic acid, 0.4 M triethanolamine and 0.2 M 2(N-morpholino)ethanesulfonic acid [16]. The pH was adjusted with hydrochloric acid or tetraethyl ammonium hydroxide at pH 6.5 for kinetic measurements, and in the pH range of 3.0-9.0 for pH optimum curves. The ionic strength remained constant at 0.1. In kinetic experiments, 10 l of 10 mM, 5 mM, 2 mM, 1 mM and 0.5 mM MU-Neu5Ac were added, while for pH measurements 10 l of 1 mM MU-Neu5Ac were used. The reaction was performed at 30°C for 10 min. Km and Vmax were calculated by fitting the kinetic data to the Michaelis-Menten equation by nonlinear regression with the program Enzfitter (Biosoft, Cambridge, UK). Protein concentrations were determined using the BIO-RAD micro-protein assay (BIO-RAD, München) using bovine serum albumin (Sigma, Deisenhofen) as a standard.

Return to the top of this document ...

C. CIDNP Method

CIDNP experiments were performed at 360 MHz on a Bruker AM-360 NMR spectrometer. The light of a continuous-wave organ ion laser (Spectra Physics, Mountain View, USA) that operates in the multiline mode with principal wave lengths of 488.0 and 514.5 nm was directed to the sample by an optical fiber and chopped by a mechanical shutter controlled by the spectrometer. Typical operation conditions were: 1 s presaturation pulse for water suppression, 0.5 s light pulse (5 W), 5 µs RF pulse (90° flip angle), 1 s acquisition time, and 5 s delay. The duration of the light pulse is short enough to prevent serious heating due to light absorption. 8, 16, or 32 light scans gave an adequate signal-to-noise ratio for the tested samples. CIDNP was generated by using flavin I mononucleotide.

Return to the top of this document ...

D. Knowledge based homology modelling

Bacterial sialidases show significant differences with viral enzymes. However one can find high similarity between Clostridium perfringens small sialidase (CPS) and the structurally solved sialidase of Salmonella typhymurium (STS) [entry 2SIL and 2SIM in the Brookhaven Protein Databank]. Both FastA and BLAST algorithms [17] report for high scores of similarity between these two enzymes. Therefore we started to bulid a three dimensional model for CPS based on the known structure of STS. Alignment of sequences of these two enzymes which was used for further framework construction is shown in figure 2. Although, because of insufficient similarity and locating in the ends of the protein, the first 30 and the last 2 amino acids of CPS have been excluded from further modelling procedure (Fig. 3), the alignment conserves particular amino acids which are located in four Asp boxes and active site of the enzyme. The next steps of computations were performed at the GLAXO institute for molecular biology SA, using the Swiss-Model Automated Protein Modelling service which makes use of ProMod (PROtein MODelling tool) [18, 19]. The program is accessible via internet browsers like Netscape and email as well. The alignments were submitted separately in the optimised mode of the program. The program integrates all the following model building steps :
Three different mutant enzymes were selected for further studies:
  1. Mutant_1 : Y336F

  2. Mutant_2 : Y347F

  3. Mutant_3 : C349S (as control)

Similar procedures were used for mutants by replacement of mutated amino acids in the pattern of alignment.The obtained four structures (wild type and three mutants) were subjected to further calculations before the measurement of surface accessibilities of aromatic amino acids. The generation of hydrogen atoms, automatic assignment of partial charges of each atom and the solvation of the molecule were accomplished using the insightII software. Each molecular ensemble was submitted to a molecular dynamics simulation using the CVFF force field at a temperature of 300 K with an equilibration time of 20 ps. A cutoff of 10 Å and dielectric constant of 4.0 was used in all calculations. A production period of 100 ps and an integration step of 1 fs was applied. Each 250 steps of integration, a conformation was saved. From the 400 saved conformations of the production time the ten with the lowest potential energy were automatically selected and submitted to an energy minmization of the complete ensemble using the conjugate gradient method. For these ten conformations (Fig. 4) the surface accessability of the respective amino acid side chains were calculated with the help of Connolly program (Fig. 5). The assessment of surface command implemented in InsightII pinpoints the accessible exterior part of the relevant portions of the molecule by smoothening the van der Waals surface with a test sphere that displays the average radius of the solvent water (1.5 Å). The dot density was generally set to a value of 1.

Return to the top of this document ...

RESULTS

Table 1: Surface accessibilities (Å2)of aromatic amino acids of Clostridium perfringens sialidase

 
		 WILDTYPE	 MUTANT_1	 MUTANT_2
		 Mean     SD	 Mean     SD	 Mean     SD
Tyr

 35		  24.7     2.0	  12.2    13.3	  13.6	   4.3
 57		  73.2     9.4	  70.2     7.1	  65.7	   8.0
 65		  54.3     6.6	  49.7     9.5	  49.2	  15.0
 82		  63.9     6.4	  43.3     4.5	  65.6	   5.5
 95		  58.7     1.8	  56.1     4.6	  87.6	   8.4
141		  19.1     2.5	  29.1     5.7	  29.9	   5.4
203		  61.5     4.8	  63.0     8.7	  41.5	   5.8
204		  41.4     4.7	   4.5     5.1	   0.6	   1.9
209		  23.4     9.1	  26.1     9.7	  34.1	  14.0
246		   0.0     0.0	   0.0     0.0	   1.8	   2.3
248		  25.1     5.3	  30.4     4.2	  15.6	   4.9
251		  64.2     4.2	  52.7     3.9	  74.7	  13.2
255		   0.0     0.0	   2.1     3.4	   0.9	   2.7
267		  17.0     3.7	  68.6     5.9	  79.6	   6.0
310		  56.3     5.2	 101.2     3.9	  34.4	  13.5
318		  25.4     3.3	  22.6     3.4	  51.1	  18.0
336		  64.8     6.3	  ****     ***	 122.0	   8.7
347		   0.0     0.0	   0.0     0.0	  ****	   ***
361		  36.6     3.6	  27.3     6.6	  36.7	   9.0
369		  40.9    11.0	   3.9     3.7	   3.1	   3.5
375		  38.8     4.5	  91.6     7.1	  98.8	   9.8
376		   8.3     8.7	  48.8     6.3	  71.9	   9.6

Trp

 80		  47.4     7.5	  26.6     8.9	  44.9	   7.0
118		   0.0     0.0	   0.2     0.8	   2.8	   2.5
124		  59.9     6.1	  85.4     7.4	  38.5	  13.0
135		   0.0     0.0	  14.0    10.9	  18.4	   4.9
149		  55.3     3.4	  67.2     5.2	  45.0	   5.2
172		   3.4     4.1	  14.9     4.2	   0.0	   0.0
217		   9.4    10.7	   1.8     4.0	  20.8	   4.5
264		   4.3     4.7	  19.9     9.1	  16.0	   4.5

His

 63		  35.9     4.4	  27.9     4.4	  33.2	   5.0
258		  64.8    10.9	  83.3     6.2	  84.3	  10.9
285		  77.1    14.3	  75.3     8.6	  80.5	   6.5
356		  17.4     9.3	  22.8     7.9	   7.6	   5.7
Return to the top of this document ...

FIGURE 2. Sequence Alignmnent of Clostridium perfringens and Salmonella typhimurium sialidases 

                .    :    .    :    .    :    .    :    .    :    .    :    60
NANH_CLOPE  *********************************SKYFRIPNIQLLNDGTILTFSDIRYNG    59
NANH_SALTY  --TVEKSVVFKAEGEHFTDQKGNTIVGSGSGGTTKYFRIPAMCTTSKGTIVVFADARHNT    59

                .    :    .    :    .    :    .    :    .    :    .    :   120
NANH_CLOPE  PDDHAYIDIASARSTDFGKTWSYNIAMKNNRIDSTYSRVMDSTTVITNT---GRIILIAG   116
NANH_SALTY  ASDQSFIDTAAARSTDGGKTWNKKIAIYNDRVNSKLSRVMDPTCIVAN*****TILVMVG   119

                .    :    .    :    .    :    .    :    .    :    .    :   180
NANH_CLOPE  SWNTNGN-WAM-TTSTRRSDWSVQMIYSDDNGLTWSNKIDLTKDSSKVKNQPSNTIGWLG   174
NANH_SALTY  KWNNND***G***DKAPDTDWDLVLYKSTDDGVT**-**ETNIHDIVTKNGTI**--**G   176

                .    :    .    :    .    :    .    :    .    :    .    :   240
NANH_CLOPE  GVGSGIVMDDGTIVMPAQISLRENNENNYYSLIIYSKDNGETWTMGNKVPNSNTSENMVI   234
NANH_SALTY  GVGSGLQLNDGKLVFPVQMVRTKNITTVLNTSF**-**DGITWSLPSGYCEGFGSENNII   235

                .    :    .    :    .    :    .    :    .    :    .    :   300
NANH_CLOPE  ELDGALIMSTRYDYSGYRAAYISHDLGTTWEIYEPLNGKILTGKGSGCQGSFIKATTSNG   294
NANH_SALTY  EFNASL**--**RNSGLRRSFETKDFGKTWTE**-**DKKVDNRNHGVQG**-**IPSGN   291

                .    :    .    :    .    :    .    :    .    :    .    :   360
NANH_CLOPE  HRIGLISAPKNTKGEYIRDNIAVYMIDFDDLSKGVQEICIPYPEDGNKLGGGYSCLSFKN   354
NANH_SALTY  KLVAAHSSAQNKNNDYTR**--**LYAHNLYSGEVKLIDAFYPKVGNASGAGYSCLSYRK   349

                .    :    .    :    .    :    .    :    .    :    .    :   420
NANH_CLOPE  N----HLGIVYEANGNIEYQDLTPYYSLIN--**                             380
NANH_SALTY  ******LYVVYEANGSIEFQDLSRHLPVIKSY                               381
Return to the top of this document ...

CONCLUSION

Clostridium perfringens small sialidase and three of its mutant forms were analysed by CIDNP experiment using the surface exposed aromatic amino acids, tyrosine, tryptophan and histidine as sensitive sensors (Fig. 6) and the results were compared with those of molecular modelling technique, calculating the surface accessibilities (Table 1). It could be shown that in each case, mutation had a significant effect on the pattern and intensities of CIDNP spectra (Fig. 6) as well as the measured Connolly surfaces of aromatic amino acids (Fig. 7 and Table 1). The experimentally derived surface accessibilities for Clostridium perfringens sialidase and its three mutants were in good agreement with those derived from molecular modelling (Fig. 6, 7 and Table 1). The fact that only Tyr 375 shows a similar increase of accessibility in mutant 1 and mutant 2 indicates that the new small Tyr signal in figure 7: b, c is generated by this amino acid residue. Mutant Y336F has a significant lower activity in comparison to the wild type. Mutant Y347F has nearly no activity due the position of Tyr 347 (surface accessibility = 0 Å2) in the catalytic center. The experimental and the theoretical approach in proteins, derived from site-directed mutagenesis confirm the occurrence of significant conformational changes for particular side chains. The CIDNP method allows to decide wether a decrease in sialidase activity is caused by an amino acid exchange in the catalytic center or conformational alterations.

Return to the top of this document ...

REFERENCES

  1. Müller HE (1992) BIOforum 1-2:16-22.

  2. Schauer R (1982) Cell Biology Monographs 10, Springer Verlag, Wien-New York.

  3. Roggentin P, Schauer R, Hoyer LL and Vimr ER (1993) Mol. Microbiol. 9(5):915-921.

  4. Crennell SJ, Garman EF, Laver WG, Vimr ER and Taylor GL (1993) PNAS 90:9852-9856.

  5. Crennell SJ, Garman EF, Laver WG, Vimr ER and Taylor GL (1994) Structure 2:535-544.

  6. Gaskell A, Crennell S and Taylor G (1995) Structure 3:1197-1205.

  7. Roggentin T, Kleineidam R, Schauer R and Roggentin P (1992) Glycoconjugate J. 9:235-240.

  8. Kaptein R, Dijkstra K and Nicolay K (1978) Nature 274:293-294.

  9. Siebert HC, Andrea S, Reuter G, Gabius HJ, Kaptein R and Vliegenthart JFG (1995) FEBS Letters 371:13-16.

  10. Siebert HC, vd Lieth CW, Kaptein R Beintema JJ, Dijkstar K, Nuland N, Soedjanaatmadja UMS, Rice A, Vliegenthart JFG, Wright CS and Gabius HJ (1996) Submitted.

  11. Roggentin P, Rothe B, Lottspeich F and Schauer R (1988) FEBS Letters 238:31-34.

  12. Sambrook J, Fritsch EF and Maniatis T (1989) Cold spring Harbour Laboratory Press, USA.

  13. Sanger F, Nicklen S and Coulson A (1977) PNAS 74:5463-5467.

  14. Studier FW, Rosenberg AH, Dunn JJ and Dubendorff JW (1990) Methods Enzymol. 185:60-89.

  15. Potier M, Mameli L, Belisle M, Dallaire L and Melancon SB (1975) Anal. Biochem. 94:287-296.

  16. Ellis KJ and Morrison JF (1982) Methods Enzymol. 87:405-426.

  17. Altschul SF, Gish W, Miller W, Myers EW and Lipman DJ (1990) J. Mol. Biol. 215:403-410.

  18. Peitsch MC and Jongeneel CV (1993) Int. Immunol. 5:233-238.

  19. Peitsch MC (1995) Bio/Technology 13:658-660.

  20. Greer J (1990) Proteins 7:317-334.

  21. Ponder JW and Richards FM (1987) J. Mol. Biol. 193:775-791.

  22. Brooks BR, Bruccoleri RE, Olafson BD, States DJ, Swaminathan S and Karplus M (1983) J. Comput. Chem. 4:187-217.

  23. Luthy R, Bowie JU and Eisenberg D (1992) Nature 356:83-85.

Return to the top of this document ...
E.Tajkhorshid@DKFZ-Heidelberg.DE
or
E. Tajkhorshid
German Cancer Research Center; DKFZ
Dept. Molecular Biophysics (0810), P.O.Box 101949, 69009 Heidelberg, FRG
Tel: +49 6221 42 2339, FAX: +49 6221 42 2333

Return to the top of this document ...
Last modified: 15/Aug/1996