Skip to main content
NCBI home page
Acta Crystallographica Section F: Structural Biology and Crystallization Communications logoLink to Acta Crystallographica Section F: Structural Biology and Crystallization Communications
. 2006 Nov 30;62(Pt 12):1278–1282. doi: 10.1107/S1744309106049281

N-terminal truncation enables crystallization of the receptor-binding domain of the FedF bacterial adhesin

Maia De Kerpel a, Inge Van Molle a, Lea Brys b, Lode Wyns a, Henri De Greve a, Julie Bouckaert a,*
PMCID: PMC2225354  PMID: 17142917

The N-terminal receptor-binding domain of the FedF adhesin from enterotoxigenic E. coli has been crystallized. This required the deletion of its first 14 residues, which are also cleaved off naturally.

Keywords: FedF, adhesins

Abstract

FedF is the two-domain tip adhesin of F18 fimbriae from enterotoxigenic Escherichia coli. Bacterial adherence, mediated by the N-terminal receptor-binding domain of FedF to carbohydrate receptors on intestinal microvilli, causes diarrhoea and oedema disease in newly weaned piglets and induces the secretion of Shiga toxins. A truncate containing only the receptor-binding domain of FedF was found to be further cleaved at its N-terminus. Reconstruction of this N-terminal truncate rendered FedF amenable to crystallization, resulting in crystals with space group P212121 and unit-cell parameters a = 36.20, b = 74.64, c = 99.03 Å that diffracted to beyond 2 Å resolution. The binding specificity of FedF was screened for on a glycan array, exposing 264 glycoconjugates, to identify specific receptors for cocrystallization with FedF.

1. Introduction

Post-weaning diarrhoea and oedema disease in piglets are caused by pathogenic Escherichia coli strains, including enterotoxigenic E. coli (ETEC) and Shiga toxin-producing E. coli (STEC). These strains produce F18 fimbriae that mediate the adherence of the bacteria to the microvilli of the epithial surface cells in the duodenum (Bertschinger et al., 1990 ▶; Gyles et al., 1998 ▶) and also secrete toxins during the infection process. The Shiga toxin Stx2e variant is secreted by F18ab-positive E. coli and reaches its target host cell via the bloodstream, resulting in the accumulation of fluid (oedema disease) and neurological disorders. F18ac-positive E. coli cause post-weaning diarrhoea by the expression of heat-stable enterotoxins (Kapitany et al., 1979 ▶).

F18 fimbriae are expressed by the fed gene cluster, with fedA encoding the major subunit (Imberechts et al., 1992 ▶), fedB the outer membrane usher, fedC the periplasmic chaperone and fedE and fedF minor pilus subunits. FedE and FedF are the F18 minor subunits essential for fimbrial adhesion and also regulate the length of the F18 fimbriae, which are normally 1–2 µm in length (Imberechts et al., 1996 ▶). FedF was shown to be the F18 tip adhesin (Smeds et al., 2001 ▶). Its pilin domain connects to the FedA polymer via multiple copies of FedE (Hahn et al., 2000 ▶). Tip adhesins can typically be divided into two immunoglobulin-like domains: an N-terminal lectin or receptor-binding module and a C-terminal pilin. Like all major and minor pilins, the pilin domains of adhesins also lack the last β-strand G of their immunoglobulin fold (Ig-fold). Therefore, the adhesin is not a stable protein by itself, but needs to be donor-strand complemented (Choudhury et al., 1999 ▶; Sauer et al., 1999 ▶). The missing strand is donated by a specific chaperone in the periplasm and by the next pilin in the pilus.

In the absence of the chaperone, soluble self-complementing and swapped pilin dimers (Jędrzejczak et al., 2006 ▶; Van Molle et al., in preparation) and trimers (Pettigrew et al., 2004 ▶; Korotkova et al., 2006 ▶) are formed and their crystal structures have been reported. Here, the pilins use different methods to complement their Ig-fold other than using the N-terminal extension of the neighbouring pilin domain. Because of misfolding, the pilins in these oligomers are incompetent for assembly into fimbriae or pili. On the other hand, crystal structures of typical tip adhesins have been mostly determined using genetic constructs that are truncated after the receptor-binding domain of the adhesin. Owing to the compact size of these domains (typically 15–18 kDa), their crystals usually diffract beyond 2 Å resolution. Examples are the fimbrial receptor-binding domains of PapG-II (Dodson et al., 2001 ▶) and of FimH (Bouckaert et al., 2005 ▶) from uropathogenic E. coli and of F17G from ETEC (Buts, Bouckaert et al., 2003 ▶).

In the current study, we report the cloning, expression, purification and crystallization of the receptor-binding domain of the FedF adhesin present on F18 fimbriae. Of all the constructs of FedF produced in this research, only one form, FedF15–165 (151 amino acids, 16.3 kDa), could be crystallized. Crystallization required not only C-­terminal but also N-terminal truncation of the FedF adhesin. Screening delivered one succesful crystallization condition from almost 500 trials using the EMBL high-throughput crystallization facility. Microcrystals could be mounted directly from the screen and flash-frozen for high-resolution data collection at the DESY synchrotron in Hamburg, Germany.

2. Materials and methods

2.1. Cloning, expression and purification and characterization of FedF truncates

The fedF gene from the F18+ E. coli 107/86 strain (Imberechts et al., 1996 ▶; gene accession No. Z26520) was cloned in the pDEST14 vector under a T7 promotor (Gateway technology, Invitrogen). The gene was truncated after Ala110 and Lys165, respectively, in order to express only the receptor-binding domain of FedF (Fig. 1 ▶). This choice was based partly on secondary-structure predictions, to avoid truncation within a clearly defined secondary-structure element (Rost & Liu, 2003 ▶), and partly on amino-acid alignments with fimbrial pilin domains of known three-dimensional structure. The resulting truncates were transformed into E. coli strain C43 (DE3) (Miroux & Walker, 1996 ▶). The C43 (DE3) cells were grown in Luria–Bertani (LB) medium until the OD600 reached 0.5, induced with 1 mM IPTG and incubated for a further 2 h. Cells were subjected to osmotic shock and the periplasmic extract was loaded onto an Ni–NTA affinity column (Qiagen) in 1 M NaCl, 20 mM Tris pH 7.5 and 10 mM imidazole. Bound protein was eluted with the same buffer and 1 M imidazole. C-terminally His-tagged proteins were further purified using gel filtration on a Superdex-75HR column in 10 mM HEPES pH 7.0 and 150 mM NaCl.

Figure 1.

Figure 1

Open in a new tab

Amino-acid sequence alignment of the FedF truncates. The first residue of the mature protein, after cleavage of the signal peptide, is labelled 1. Differences from the FedFv1–165 variant are indicated in bold. The two N-termini found for FedF1–165 are underlined.

Gene constructs encoding truncated FedF proteins without a His tag were expressed in the same way. The periplasmic extract was loaded onto a Source 30S column (Amersham) for cation exchange in 20 mM Tris pH 7.5 and eluted using an NaCl gradient at a concentration of 140 mM NaCl. Subsequently, the same gel-filtration procedure was used as for the His-tagged proteins.

The identity of purified FedF1–165H6 was first confirmed using an anti-His Mab on a Western blot as described in Coligan et al. (2002 ▶). The nitrocellulose membrane was blocked overnight at 277 K with phosphate-buffered saline complemented with 1% casein. The membrane was incubated with the primary antibody at room temperature for 2 h. Goat anti-mouse or anti-rabbit antibodies conjugated with peroxidase were then added for 1 h. The membrane was developed using H2O2. The purified FedF1–165H6 was injected in complete Freund adjuvant intradermally in a rabbit, with the purpose of obtaining a polyclonal rabbit anti-FedF1–165H6 serum. After three weeks, a boost was given by injecting FedF1–165H6 in incomplete Freund adjuvant. Serum containing rabbit anti-FedF was taken after another 10 d.

N-terminal sequencing (Alta Bioscience) was performed after electrotransfer of FedFv1–165 from a 12.5% SDS–PAGE onto a ProBlot PVDF membrane (Applied Biosystems) and excision of the bands corresponding to the estimated molecular weight upon staining with Coomassie R350.

2.2. Crystallization and crystal analysis

Screening for crystallization conditions for FedF15–165 took place at the high-throughput crystallization facility at the EMBL in Hamburg (Germany) after in-house screening with commercially available screens (Crystal Screens 1 and 2 from Hampton Research) remained unsuccessful. Using the robot, 300 nl droplets of protein solution (24 mg ml−1 in 20 mM HEPES pH 7.5 and 140 mM NaCl) were mixed with 300 nl precipitant solution. Five sets of 96 precipitant conditions from both commercial and academically developed screens were used: Index, Crystal Screens I and II, Natrix and MembFac (Hampton Research), the JCSG screen (Newman et al., 2005 ▶) and Jena 1–4 (Jena Bioscience). Crystals of the FedF15–165 truncate were only obtained using the JCSG screen with 0.17 M ammonium sulfate, 25.5%(w/v) PEG 4000 and 15%(v/v) glycerol. Larger crystals were obtained at home using the same conditions and microseeding.

Prior to data collection, a microcrystal of FedF15–165 was directly mounted from the JCSG screen in a nylon cryoloop (Hampton Research) and placed directly into a nitrogen-gas stream at 100 K at beamline X12 of the EMBL Outstation Hamburg, Germany. The data were processed using DENZO, XDISPLAYF and SCALEPACK from the HKL package (Otwinowski & Minor, 1997 ▶) and TRUNCATE from the CCP4 suite (Collaborative Computational Project, Number 4, 1994 ▶). Crystal parameters and data-processing statistics are summarized in Table 1 ▶.

Table 1. Data-collection and processing for FedF15–165 .

Values in parentheses are for the highest resolution shell (2.07–2.00 Å).

Space group P212121
Unit-cell parameters (Å) a = 36.20, b = 74.64, c = 99.03
Unit-cell volume (Å3) 267578
Resolution range (Å) 42–2.0
Reflections (total/unique) 100446/18666
Completeness (%) 98.9 (98.8)
Rmerge† (%) 7.5 (12.1)
〈I/σ(I)〉 18.5 (12.4)
Matthews coefficient (Å3 Da−1) 2.05
Solvent content (%) 40.1
FedF molecules per ASU 2
Open in a new tab
†

R merge = Inline graphic Inline graphic, where I l is the lth observation of reflection h and 〈I h〉 is the weighted average intensity for all observations l of reflection h.

2.3. Screening of carbohydrate binding of FedF on a glycan microarray

The carbohydrate-specificity of the FedF1–165 receptor-binding domain of the FedF fimbrial adhesin from the F18+ E. coli 107/86 strain was screened against printed array version 2. Slides (Schott Nexterion H slides, catalogue No. 1070936B) were printed at the Core D carbohydrate-synthesis/protein-expression consortium located at the Scripps Research Institute. The slides exposed 264 immobilized glycans spotted at a concentration of 100 µM and in six replicates (Blixt et al., 2004 ▶). Preprinted slides were soaked in de­ionized water for 5 min at room temperature and dried in a stream of nitrogen. FedF1–165 was diluted to the assay concentration of 100 µg ml−1 in 20 mM Tris–HCl pH 7.4, 150 mM NaCl, 2 mM CaCl2 and 2 mM MgCl2, the wash buffer, in the presence of 0.05% Tween-20 and 1% BSA, and 50 µl was applied to the printed surface and covered with a cover slip. The slide was incubated, protected from light, in a humidified chamber for 1 h at room temperature. A first wash step used the wash buffer with 0.05% Tween-20 and wash buffer alone for 10 s each. For detection, 50 µl rabbit anti-FedF was applied at 10 µg ml−1 and the slide was subsequently incubated and washed as described above. Finally, 50 µl labelled goat anti-rabbit IgG-488 was applied, incubated and washed again in the same way. After another 10 s wash in deionized water, the slide was dried in a nitrogen stream. The binding image was read by a Perkin–Elmer Microscanarray XL4000 scanner. Image analysis was performed using Imagene v.6 image-analysis software. The raw data results files were uploaded to the Consortium for Functional Glycomics database and webpages (http://www.functionalglycomics.org).

3. Results and discussion

Of the two FedF proteins, FedF1–110H6 and FedF1–165-H6, that were C-­terminal truncates followed by a six-histidine tag, only FedF1–165H6 was soluble (Fig. 1 ▶). However, it did not crystallize. Therefore, FedF1–­165 and two more truncates, FedF1–160 and FedF1–162, which are five and three amino-acid residues shorter (Fig. 1 ▶), respectively, were produced without a His tag. The construct for the shortest C-terminal truncate, FedF1–160, did not express. FedF1–162 and FedF1–165 were purified as soluble proteins and set up for crystallization. However, no crystals could be obtained. In an attempt to make the FedF receptor-binding domain more amenable to crystallization (Buts et al., 2005 ▶), we purified FedFv1–165 (Fig. 1 ▶) originating from a natural variant of the adhesin from the F18+ E. coli strain 189P (Imberechts, CODA, Belgium). FedFv1–165 also did not crystallize.

Remarkably, for all of the expressed soluble FedF truncates, the protein appeared as two close bands on SDS–PAGE, near the expected molecular weight of 18 kDa (Fig. 2 ▶). Moreover, both bands were also detected on Western blotting using the polyclonal rabbit antibody against FedF. The addition of reducing agents, such as 20 mM DTT, to the sample-loading buffer did not change this feature, indicating that this was not a consequence of incomplete disulfide reduction. N-terminal sequencing revealed two different N-termini, starting either at the expected residue or 14 residues later (Fig. 1 ▶). These results demonstrate that the FedF constructs were subject to a naturally occurring N-terminal cleavage. It was reasoned that the N-­terminal heterogeneity could disturb the crystallization process. Two constructs encoding FedF13–165 and FedF15–165 were designed to match the shorter sequence of FedF1–165 (Fig. 1 ▶). The signal sequence of the F4 fimbrial adhesin from E. coli was used because it is known to allow high levels of periplasmic FaeG. In FedF13–165, the two residues preceding Asn15 were kept to resemble the cleavage site for the signal peptidase in FaeG. Nevertheless, the N-terminally truncated FedF15–165 led to higher yields of soluble protein (Fig. 2 ▶). It was purified and set up for crystallization.

Figure 2.

Figure 2

Open in a new tab

FedF1–165 (lane 2) showing two bands on 12% SDS–PAGE, the N-terminal truncate FedF15–165 (lane 3) and PageRuler protein ladder (Fermentas SM0661; lane 1).

Crystals were obtained in one of the 500 conditions screened (Fig. 3 ▶). Although these crystals were smaller (300 × 40 × 40 µm) than those later reproduced in-house (1 × 0.2 × 0.1 mm), they diffracted considerably better (Table 1 ▶). It is most likely that the variable nonthermostatic conditions during setup are partly responsible for these differences, as the protein droplets tend to precipitate rapidly upon mixing at temperatures above 298 K in the laboratory. On the other hand, the temperature is well controlled in the crystallization robot and storage room.

Figure 3.

Figure 3

Open in a new tab

Crystals of FedF15–165 grown at the EMBL high-throughput crystallization facility.

Sugars that bind specifically to FedF may serve as a tool to solve its structure, as exemplified by the use of an Se derivative of GlcNAc (Buts, Loris et al., 2003 ▶), or to obtain crystals in different space groups. The glycan array pointed out the recognition of N-acetyllactosamine units, Gal(β1-4)GlcNAc, that were either sulfated at position 3 and/or position 6 or fucosylated at position 2 of the galactose (Fig. 4 ▶). α1-Acid glycoprotein, which carries Gal(β1-4)­[Fuc(α1-3)]GlcNAc, was the best binding glycoconjugate. Other recognized carbohydrates were α1-2-fucosylated, namely Gal(α1-3)[Fuc(α1-2)]Gal and the blood group H type II trisaccharide Fuc(α1-2)­Gal(β1-4)GlcNAc (Fig. 4 ▶). Blood group H type II tri­saccharide has previously been characterized as an F18 receptor (Snoeck et al., 2004 ▶). (α1-2)-Fucosyltransferases have been implicated as being of importance for the presence of F18 receptors in piglets (Meijerink et al., 1997 ▶, 2000 ▶). Trials to cocrystallize FedF in complex with several of these carbohydrates have so far remained unsuccessful.

Figure 4.

Figure 4

Open in a new tab

Glycoconjugates recognized by FedF1–165 on a glycan array by the Consortium for Functional Glycomics.

Natural C-terminal truncates have been observed when overproducing recombinant tip adhesins, for example FimH and F17G. This possibly occurs as noncomplemented pilin domain of the adhesin becomes degraded. To date, N-terminal cleavage of tip adhesins has not been observed. It is therefore currently not understood what the mechanism and role of the cleavage of the 14-residue N-terminal truncation of FedF1–165 is (Fig. 1 ▶). The N-terminal 14-residue sequence is very susceptible to proteolysis (http://www.expasy.org/tools/peptidecutter), whereas the amino-acid sequence just after the N-terminal truncation site (Asn15) is predicted not to be susceptible. Its sensitivity to proteolysis or hydrolysis could indicate that the structure of the N-terminal 14 residues behaves independently of the globular receptor-binding domain. Its length is reminiscent of the typical N-terminal extension that serves as a donor strand for the complementation of the incomplete Ig-fold of pilins. Moreover, its amino-acid sequence has an alternating hydrophobic–hydrophilic pattern and is conserved among variants (Fig. 1 ▶). Both those features are not present for the new N-terminus of the N-terminal truncate. If the 14 residues were a donor strand, this would have important implications. It would allow FedF to intercalate in the middle of F18 fimbriae by donating this N-terminal extension to a previous pilin in the pilus. In the absence of the 14-residue N-terminal sequence, FedF would be destined to reside only at the tip of F18 fimbriae. Solving the FedF crystal structure could shed light on this mechanism.

Acknowledgments

We thank Liu Qiang from the Research Center of Centre Hospitalier de l’Université de Montreal, Canada for his help during the initiation of this project. We wish to thank Richard A. Alvarez, Director of Core H, for performing the glycan-array analysis, conducted by the Protein-Glycan Interaction Core H of the Consortium for Functional Glycomics funded by NIGMS grant GM62116.  We are grateful to Dr Manfred Weiss, EMBL beamline scientist at X12 at the DESY synchrotron, Hamburg. JB is a postdoctoral fellow of the Fonds voor Wetenschappelijk Onderzoek–Vlaanderen, which also provided financial support (FWO-G.0513.04) and the DNA-sequencing equipment (FWO-AL215).

References

  1. Bertschinger, H. U., Bachmann, M., Mettler, C., Pospischil, A., Schraner, E. M., Stamm, M., Sydler, T. & Wild, P. (1990). Vet. Microbiol.25, 267–281. [DOI] [PubMed] [Google Scholar]
  2. Blixt, O. et al. (2004). Proc. Natl Acad. Sci. USA, 101, 17033–17038. [Google Scholar]
  3. Bouckaert, J., Berglund, J., Schembri, M., De Genst, E., Cools, L., Wuhrer, M., Hung, C. S., Pinkner, J., Slättegård, R., Zavialov, A., Choudhury, D., Langermann, S., Hultgren, S., Wyns, L., Klemm, P., Oscarson, S., Knight, S. & De Greve, H. (2005). Mol. Microbiol.55, 441–455. [DOI] [PubMed] [Google Scholar]
  4. Buts, L., Bouckaert, J., De Genst, E., Loris, R., Oscarson, S., Lahmann, M., Messens, J., Brosens, E., Wyns, L. & De Greve, H. (2003). Mol. Microbiol.49, 705–715. [DOI] [PubMed] [Google Scholar]
  5. Buts, L., Loris, R., De Genst, E., Oscarson, S., Lahmann, M., Messens, J., Brosens, E., Wyns, L., De Greve, H. & Bouckaert, J. (2003). Acta Cryst. D59, 1012–1015. [DOI] [PubMed] [Google Scholar]
  6. Buts, L., Wellens, A., Van Molle, I., Wyns, L., Loris, R., Lahmann, M., Oscarson, S., De Greve, H. & Bouckaert, J. (2005). Acta Cryst. D>61, 1149–1159. [DOI] [PubMed] [Google Scholar]
  7. Choudhury, D., Thompson, A., Stojanoff, V., Langermann, S., Pinkner, J., Hultgren, S. J. & Knight, D. S. (1999). Science, 285, 1061–1065. [DOI] [PubMed] [Google Scholar]
  8. Coligan, J. E., Dunn, B. M., Ploegh, H. L., Speicher, D. W. & Wingfield, P. T. (2002). Editors. Current Protocols in Protein Science. New York: Wiley.
  9. Collaborative Computational Project, Number 4. (1994). Acta Cryst. D50, 760–­763. [Google Scholar]
  10. Dodson, K. W., Pinkner, J. S., Rose, T., Magnusson, G., Hultgren, S. J. & Waksman, G. (2001). Cell, 105, 733–743. [DOI] [PubMed] [Google Scholar]
  11. Gyles, C., Johnson, R., Gao, A. L., Ziebell, K., Pierard, D., Aleksic, S. & Boerlin, P. (1998). Appl. Environ. Microbiol.64, 4134–4141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Hahn, E., Wild, P., Schraner, E. M., Bertschinger, H. U., Häner, M., Müller, S. A. & Aebi, U. (2000). J. Struct. Biol.132, 241–250. [DOI] [PubMed] [Google Scholar]
  13. Imberechts, H., De Greve, H., Schlicker, C., Bouchet, H., Pohl, P., Charlier, G., Bertschinger, H., Wild, P., Vandekerckhove, J., Van Damme, J., Van Montagu, M. & Lintermans, P. (1992). Infect. Immun.61, 1963–1971. [DOI] [PMC free article] [PubMed]
  14. Imberechts, H., Wild, P., Charlier, G., De Greve, H., Lintermans, P. & Pohl, P. (1996). Microb. Pathog.21, 183–192. [DOI] [PubMed] [Google Scholar]
  15. Jędrzejczak, R., Dauter, Z., Dauter, M., Piatek, R., Zalewska, B., Mroz, M., Bury, K., Nowicki, B. & Kur, J. (2006). Acta Cryst. D62, 157–164. [DOI] [PubMed] [Google Scholar]
  16. Kapitany, R. A., Forsyth, G. W., Scoot, A., Mckenzie, S. F. & Worthington, R. W. (1979). Infect. Immun.26, 173–177. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Korotkova, N., Le Trong, I., Samudrala, R., Korotkov, K., Van Loy, C. P., Bui, A. L., Moseley, S. L. & Stenkamp, R. E. (2006). J. Biol. Chem.281, 22367–22377. [DOI] [PubMed] [Google Scholar]
  18. Meijerink, E., Fries, R., Vogeli, P., Masabanda, J., Wigger, G., Stricker, C., Neuenschwander, S., Bertschinger, H. U. & Stranzinger, G. (1997). Mamm. Genome, 8, 736–741. [DOI] [PubMed] [Google Scholar]
  19. Meijerink, E., Neuenschwander, S., Fries, R., Dinter, A., Bertschinger, H. U., Stranzinger, G. & Vogeli, P. (2000). Immunogenetics, 52, 129–136. [DOI] [PubMed] [Google Scholar]
  20. Miroux, B. & Walker, J. E. (1996). J. Mol. Biol.260, 289–298. [DOI] [PubMed] [Google Scholar]
  21. Newman, J., Egan, D., Walter, T. S., Meged, R., Berry, I., Ben Jelloul, M., Sussman, J. L., Stuart, D. I. & Perrakis, A. (2005). Acta Cryst. D61, 1426–1431. [DOI] [PubMed] [Google Scholar]
  22. Otwinowski, Z. & Minor, W. (1997). Methods Enzymol.276, 307–326. [DOI] [PubMed]
  23. Pettigrew, D., Anderson, K. L., Billington, J., Cota, E., Simpson, P., Urvil, P., Rabuzin, F., Roversi, P., Nowicki, B., du Merle, L., Le Bouguenec, C., Matthews, S. & Lea, S. M. (2004). J. Biol. Chem.279, 46851–46857. [DOI] [PubMed] [Google Scholar]
  24. Rost, B. & Liu, J. (2003). Nucleic Acids Res.31, 3300–3304. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Sauer, F. G., Fütterer, K., Pinkner, J. S., Dodson, K. W., Hultgren, J. S. & Waksman, G. (1999). Science, 285, 1058–1061. [DOI] [PubMed] [Google Scholar]
  26. Smeds, A., Hemmann, K., Jakava-Viljanen, M., Pelkonen, S., Imberechts, H. & Palva, A. (2001). Infect. Immun.69, 7941–7945. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Snoeck, V., Verdonck, F., Cox, E. & Goddeeris, B. M. (2004). Vet. Microbiol.100, 241–246. [DOI] [PubMed] [Google Scholar]

Articles from Acta Crystallographica Section F: Structural Biology and Crystallization Communications are provided here courtesy of International Union of Crystallography

RESOURCES