Structural characterization of a novel autonomous cohesin from Ruminococcus flavefaciens

The crystal structure of a standalone cohesin protein from the rumen bacterium R. flavefaciens has been determined at 2.44 Å resolution, revealing an elaborate set of secondary-structural elements.

Abstract

Ruminococcus flavefaciens is a cellulolytic bacterium found in the rumen of herbivores and produces one of the most elaborate and variable cellulosome systems. The structure of an R. flavefaciens protein (RfCohG, ZP_06142108), representing a freestanding (non-cellulosomal) type III cohesin module, has been determined. A selenomethionine derivative with a C-terminal histidine tag was crystallized and diffraction data were measured to 2.44 Å resolution. Its structure was determined by single-wavelength anomalous dispersion, revealing eight molecules in the asymmetric unit. RfCohG exhibits the most complex among all known cohesin structures, possessing four α-helical elements and a topographical protuberance on the putative dockerin-binding surface.

1. Introduction  

Cellulosomes are extracellular multi-enzyme complexes designed for efficient degradation of cellulose and related plant cell-wall polysaccharides (Lamed et al., 1983 ▶). The cellulosomal enzymes are assembled on a noncatalytic scaffoldin protein comprised of an array of repeating cohesin modules. The enzymes carry a dockerin module, which interacts with a single cohesin of a specific scaffoldin. The selective and high-affinity binding of the dockerin-bearing enzyme subunits to the scaffoldin-borne cohesins dictates the assembly of the mature cellulosome. The integrity of the entire cellulosome complex is thus maintained by the cohesin–dockerin interaction (Xu et al., 2004 ▶; Bayer et al., 1994 ▶, 1998 ▶).

Ruminococcus flavefaciens is a predominant cellulolytic rumen bacterium, which forms a multi-enzyme cellulosome complex that plays a central role in the ability of this bacterium to degrade plant cell-wall polysaccharides (Antonopoulos et al., 2004 ▶; Julliand et al., 1999 ▶; Krause et al., 1999 ▶; Sijpesteijn, 1951 ▶). R. flavefaciens strain FD-1 possesses the most complex cellulosome system known to date (Fig. 1 ▶), as indicated by its genome sequence (Berg Miller et al., 2009 ▶). Nearly 225 different dockerin-containing coding sequences have been identified in the R. flavefaciens FD-1 genome (Rincon et al., 2010 ▶). Surprisingly, a large number of these ORFs do not encode the types of protein that one would expect for cellulosomes, i.e. carbohydrate-active enzymes, but rather novel and unexpected structural and catalytic modules annotated as putative proteases, trans-glutaminases, LRR proteins, serpins and many small novel scaffoldin sequences (composed of a dockerin and one or two suspected cohesin modules). The role of these proteins has yet to be correlated with the established cellulosome function of enhancing polysaccharide degradation.

Figure 1.

Schematic overview of the modular interactions in the cellulosome system of R. flavefaciens strain FD-1. The primary two-cohesin scaffoldin ScaA binds specifically either to the Cel44A-type dockerins (a) or to the C-terminal dockerin of ScaC (b). The ScaA dockerin module interacts with one of the five type III cohesins (numbered 5–9) (c) carried by the adaptor scaffoldin ScaB. Like the ScaA cohesins, ScaB cohesins 1–4 (d) also bind both the ScaC dockerin (b) and Cel44A-type dockerins (a). The conserved XDoc dyad (in red) of ScaB (e) interacts with the cohesin module of the anchoring scaffoldin ScaE. The ScaE cohesin also interacts with the XDoc dyads of the cellulose-binding proteins CttA and the putative cysteine peptidase RflaF_05439 (Levy-Assaraf et al., 2013 ▶) (f). An additional standalone cohesin, CohG, the topic of this communication, interacts selectively with the XDoc dyads of ScaB and CttA but not with that of RflaF_05439.

Interestingly, several key cellulosomal scaffoldins, including ScaA, ScaB, ScaC, CttA and ScaE, are organized into a single sca gene cluster. The presence of this gene cluster has been documented in at least five different strains of R. flavefaciens (Jindou et al., 2008 ▶). Additionally, one peculiar and very interesting ORF containing only a signal peptide and an autonomous cohesin module (i.e. technically not part of a scaffoldin) was identified by bioinformatics analysis and termed RfCohG. Biochemical evidence including enzyme-linked immunosorbent assays (ELISA) has shown that the RfCohG binds cellulosome-related proteins such as ScaB and CttA (unpublished data). To understand the function and framework of this intriguing cellulosome-free RfCohG and its relation to cellulosomal cohesins, we have undertaken the determination of its three-dimensional structure.

2. Materials and methods  

2.1. Cloning  

The DNA encoding for the RfCohG module (accession No. ZP_06142108, residues 26–218) was amplified by PCR from R. flavefaciens FD-1 genomic DNA, isolated as described by Murray & Thompson (1980 ▶), using the primers 5′-CATGCCATGGGGAGCAGTTCGGTTACTGCTGATCTG and 5′-CCGCTCGAGTTC­AACTGTTATAGTGCCGCCCTCC designed according to the DNA sequence of the genomic Contig43 (GenBank NZ_ACOK01000043). The PCR product was inserted into pET28a(+) expression vector (Novagen, Madison, Wisconsin, USA) via NcoI and XhoI restriction enzymes to introduce a C-terminal hexahistidine (His) tag.

2.2. Expression of seleno-l-methionine-labelled protein  

Expression was conducted according to the method described earlier (Van Duyne et al., 1993 ▶), with minor modifications. Transformed cells from a culture grown overnight in 1 ml Luria–Bertani broth containing 50 µg ml⁻¹ kanamycin were isolated and resuspended in 1 ml M9 minimal medium (Sigma, St Louis, Missouri, USA), supplemented with glucose (4 mg ml⁻¹), 1 ml 100 mM CaCl₂, 1 ml 1 M MgSO₄, thiamine and vitamin B₁ (5 mg ml⁻¹ each). The resuspended culture was added to 1 l of the same medium pre-incubated at 310 K. Incubation of the culture was continued at 310 K with shaking until the growth culture reached an OD₆₀₀ of 0.6. At this point, seleno-l-methionine was added to a final concentration of 50 µg ml⁻¹. Inhibition of the methionine pathway was achieved with the addition of the following amino acids, which were added as solids: lysine hydrochloride (100 mg), threonine (100 mg), phenylalanine (100 mg), leucine (50 mg), isoleucine (50 mg) and valine (50 mg). After an additional 15 min of shaking, 0.1 mM isopropyl β-d-1-thiogalactopyranoside was added and the culture was grown for an additional 13 h.

Cells were harvested by centrifugation (4000g for 15 min) at 277 K and resuspended in 50 mM NaH₂PO₄ pH 8.0 containing 300 mM NaCl at a ratio of 1 g wet pellet to 4 ml buffer solution. DNase was added prior to the sonication procedure. The suspended pellet was sonicated in a sonicator Ultrasonic Processor X2 (Misonix Inc.) for 20 min in discontinuous mode (0.05 s pulse on and 0.05 s pulse off). The suspension was kept on ice during sonication; it was then centrifuged (20 000g at 277 K for 20 min) and the supernatant fluids were collected.

2.3. Protein purification  

The recombinant His-tagged protein was first isolated by metal-chelate affinity chromatography using Ni–IDA resin (Rimon Biotech, Israel) according to the manufacturer’s recommended protocol. Further purification was accomplished by fast protein liquid chromatography using a Superdex 75 16/60 column and an ÄKTAprime system (GE Healthcare). The purified protein solution consisted of 13 mg ml⁻¹ protein in 25 mM Tris–HCl pH 7.0, 25 mM NaCl, 1 mM DTT, 0.05% sodium azide.

2.4. Crystallization, data collection and structure determination  

Crystals were grown at 293 K by the hanging-drop vapour-diffusion method (McPherson, 1982 ▶). The first crystals appeared after several days in condition No. 37 of Crystal Screen from Hampton Research consisting of 0.1 M sodium acetate pH 4.6, 8%(w/v) polyethylene glycol 4000. The initial crystallization conditions were further optimized, and the best crystals were obtained after 3 d in a 9 µl drop comprised of 2 µl SeMet-RfCohG solution (13 mg ml⁻¹) and 7 µl reservoir solution [0.1 M sodium acetate pH 5, 6%(w/v) polyethylene glycol 4000].

The crystals were harvested from the crystallization drop surrounded by mother liquor using thin-walled glass capillaries (Glas Technik & Konstruktion, Berlin). The capillaries were sealed at the narrow end using a flame and at the funnel end with high vacuum grease to facilitate transfer to the synchrotron. At the beamline, the capillaries were opened by diamond cutter and the crystal was plunged into mother liquor supplemented by 25% ethylene glycol for cryoprotection. Diffraction data were collected on beamline ID29 at the European Synchrotron Radiation Facility (ESRF, Grenoble, France). X-ray radiation of 0.9795 Å wavelength was determined by energy scan and a CCD ADSC detector was used.

Screening numerous SeMet-RfCohG crystals finally yielded a crystal that gave a reasonable diffraction pattern extending to a resolution of 2.44 Å with well separated diffraction spots. The data set was collected in 1° oscillations, and a total of 360 images were collected, indexed, processed and scaled using DENZO and SCALEPACK as implemented in HKL-2000 (Otwinowski & Minor, 1997 ▶). The crystal belonged to the monoclinic space group C2. The calculated Matthews coefficient (Matthews, 1968 ▶) was 3.18 ų Da⁻¹, corresponding to the presence of approximately eight molecules in the asymmetric unit. The statistics of the diffraction data are shown in Table 1 ▶.

Table 1. Crystal parameters and data-collection and refinement statistics.

Values in parentheses are for the highest resolution shell.

X-ray source	ESRF ID29
Space group	C2
No. of crystals	1
Total rotation angle ()	360
Unit-cell parameters
a ()	116.24
b ()	67.73
c ()	258.91
()	93.08
V (3)	2038247
No. of molecules in asymmetric unit	8
Resolution range ()	502.44 (2.472.44)
Total No. of reflections	554681
Unique reflections	74822 (3735)
Mosaicity ()	0.48-1.14
Multiplicity	7.34 (6.0)
Completeness (%)	98.8 (97.3)
Mean I/(I)	14.4 (2.0)
R merge †	0.135
Overall average B factor (2)	45.8
No. of protein residues	1544
No. of solvent atoms	167
No. of ions	1
R cryst/R free	0.20542/0.22941
Geometry
R.m.s.d., bonds ()	0.011
R.m.s.d., angles ()	1.976
MolProbity validation
Ramachandran favoured (%) (goal >98%)	99.33
Ramachandran outliers (%) (goal 0.2%)	0.0
Rotamer outliers (goal 1%)	1.5
C outliers (goal 0)	0
Clashscore‡	4.43

^† R _merge = , where denotes the sum over all reflections and the sum over all equivalent and symmetry-related reflections (Stout Jensen, 1968 ▶).

^‡Clashscore is the number of serious steric overlaps (>0.4) per 1000 atoms.

2.5. Structure determination and refinement  

The structure of SeMet-RfCohG was determined by the single-wavelength anomalous dispersion (SAD) method. The programs SHELXC/D/E (Sheldrick, 2010 ▶) as implemented in HKL2MAP (Pape & Schneider, 2004 ▶) were used in heavy-atom substructure determination, phase determination and polyalanine tracing. The default decision of the programs regarding the starting parameters to be used in substructure determination proved unsuccessful. Manual forcing of the complete resolution and increasing the number of trials to 1000 led to the determination of the substructure. Subsequently, the protein structure was phased, phases were modified, and the initial polyalanine skeleton was built, exhibiting eight independent molecules in the asymmetric unit. ARP/wARP was used for building the structure (Langer et al., 2008 ▶). REFMAC5 (Murshudov et al., 2011 ▶) and Coot (Emsley et al., 2010 ▶) were used for structure refinement and rebuilding. The final coordinates and structure factors were deposited in the PDB under accession code 4n2o.

3. Results and discussion  

3.1. Overall structure of RfCohG reveals jelly-roll topology decorated by four α-helical elements  

RfCohG forms an elongated nine-stranded β-sandwich in a classical jelly-roll topology with approximate overall dimensions of 51 × 35 × 28 Å as measured on the C^α skeleton using PyMOL (DeLano, 2002 ▶). The ‘front face’ of the molecule is formed by β-strands 8, 3, 6, 5 and α-helices 1, 2 and 4, and the ‘back face’ is formed by β-strands 9, 1, 2, 7, 4 and α-helix 3. As shown in Fig. 2 ▶, strands 1 and 9 are aligned parallel to each other, while the other β-strands are anti-parallel. β-Strands 4 and 8 are disrupted by β-flaps (Fig. 2 ▶, coloured green). There are four distinct α-helices in RfCohG. These include a seven-residue α-helix (H1) in the N-terminal region, a three-residue α-helix (H2) located on the β-flap between β-strands 3 and 4, a five-residue α-helix (H3) located between β-strands 5 and 6, and a ten-residue α-helix (H4) at the C-terminus (see Figs. 2 ▶ and 3 ▶). There are eight molecules per asymmetric unit, organized as two tetramers, each of which is a dimer of dimers (Fig. 4 ▶ a). These independent molecules are very similar with r.m.s.d.s of between 0.03 and 0.13 Å on C^α positions. Upon operation of symmetry elements of space group C2, the structure packs into helical fibre-like formations (Fig. 4 ▶ b).

Figure 2.

Cartoon representation of the overall structure of RfCohG. The major secondary-structural elements are numbered according to Fig. 3 ▶. α-Helices (H1–H4) are coloured yellow, β-strands are coloured blue and β-flaps are coloured green.

Figure 3.

Structure-based alignment of RfCohG versus type I, type II and type III cellulosomal cohesins. RfCohG was superimposed with type I (PDB entry 1anu), type II (PDB entry 1tyj) and type III (PDB entry 2zf9) cohesin structures, and the sequences were aligned accordingly. The residues of the β-strands at homologous positions are indicated in blue. The residues of the β-flaps are indicated in green. The helical elements are highlighted in yellow. Secondary-structural elements are marked according to RfCohG. The three tandem tyrosine residues (160–162) of RfCohG are highlighted in cyan.

Figure 4.

Stereoview (cross-eye) of (a) eight independent molecules of the asymmetric unit of RfCohG and (b) two adjacent fibre-like formations formed by applying the symmetry elements of space group C2 to the asymmetric unit. Note that head-to-head contacts (A–A and G–G) between asymmetric units are formed by twofold axes.

3.2. Comparison of RfCohG with other cohesin structures  

A structural similarity search, using the PDBeFold server (European Bioinformatics Institute; http://www.ebi.ac.uk/msd-srv/ssm; Krissinel & Henrick, 2004 ▶) revealed that despite the relatively low sequence identity (9–25%), RfCohG can be structurally aligned to cohesin modules from all of the known types (I–III). RfCohG is the most similar to the type IIIe cohesin from R. flavefaciens ScaE (PDB entry 2zf9; Alber et al., 2009 ▶; Salama-Alber et al., 2013 ▶) with an r.m.s.d. of 1.36 Å, compared with an r.m.s.d. of 2.29 Å with the type I cohesin from the scaffoldin subunit of Clostridium thermocellum (PDB entry 1anu; Shimon et al., 1997 ▶) and an r.m.s.d. of 2.8 Å with the type II cohesin from the scaffoldin subunit of Bacteroides cellulosolvens (PDB entry 1tyj; Noach et al., 2005 ▶). The positions of the RfCohG β-flaps are consistent with those of the type II cohesins, i.e. in the midst of β-strands 4 and 8 (Noach et al., 2003 ▶, 2005 ▶; Carvalho et al., 2005 ▶).

The addition of the secondary-structural elements is clearly observed upon comparison between the different cohesin types (Figs. 3 ▶ and 5 ▶). In this context, type I cohesins are constrained to form only β-sheets or extremely short α-helix motifs (Fig. 5 ▶ a), type I cohesins possess longer α-helices (Fig. 5 ▶ b) in addition to two β-flaps, and type IIIe cohesins possess one or more dominant α-helix elements in addition to these two β-flaps (Figs. 5 ▶ c and 5 ▶ d).

Figure 5.

Structural comparison among different cellulosomal cohesin types. (a) Type I cohesin (PDB entry 1anu), (b) type II cohesin (PDB entry 1tyj), (c) the type IIIe cohesin RfCohE (PDB entry 2zf9), (d) the type III cohesin RfCohG. α-Helical elements in each molecule are coloured yellow, β-strands are coloured blue and β-flaps are coloured green.

Moreover, RfCohG possesses a unique structural feature that distinguishes it from the other known cohesin structures. While the type II and type III cohesins (RfCohE from ScaE) possess a single α-helix located at different positions in the cohesin molecule (Noach et al., 2005 ▶; Alber et al., 2009 ▶), RfCohG possesses four α-helical elements (indicated in yellow in Fig. 5 ▶). The position of the last helix (H4) is equivalent to the position of the dominant helix in the RfCohE type IIIe molecule (Fig. 3 ▶), although the helix of RfCohG is four residues shorter. All four α-helices are separated from the major aligned portion of the molecule. α-Helix 2 (marked H2 in Fig. 2 ▶) and the short β-strand 4 are both located within the lengthy β-flap connecting strands 3 and 5. This β-flap protrudes 11 Å from the planar face of the 8–3–6–5 β-sheet, which together with β-flap 4 forms the protuberance in the middle of the molecule (Fig. 6 ▶).

Figure 6.

Structural comparison of RfCohG with RfCohE. (a) Cartoon representation of the superposition of C^α atoms of RfCohG (blue) and RfCohE (red). (b) Surface and cartoon representations highlighting the protrusion of RfCohG (coloured cyan), composed of β-flaps 4 and 8 and α-helix 2, and the protrusion of RfCohE (colored magenta) composed of β-flap 8. (c) Molecules in (b) rotated 90° about the y axis, emphasizing the height of the protrusion.

3.3. Three tandem tyrosine residues anchor the helices to the molecule core  

Similar to other cohesin modules (Shimon et al., 1997 ▶; Alber et al., 2009 ▶), the nine β-strands of RfCohG are assembled around a hydrophobic core of aromatic and aliphatic residues with numerous stabilizing aromatic interactions (Burley & Petsko, 1985 ▶).

Interestingly, the core includes a distinctive stretch of three repeated tyrosines (Tyr160, Tyr161 and Tyr162), located in the linker region between β-strand 8a and β-strand 8b (Fig. 3 ▶, highlighted in cyan, Fig. 7 ▶). Tyr 162 connects α-helices 1 and 3 via hydrophobic interactions with Val14, Ala23, Phe24 and Phe182. Tyr161 connects α-helix 2 by hydrophobic interactions with Ile15, Ile71, Trp107, Pro112, Val114 and Leu115. Tyr 160 consolidates the α-helices to the rest of the molecule via hydrophobic interactions with Ala13, Val14, Ile37, Leu39, Ile68 and Ile70. This composite feature may serve as an example whereby an extensive hydrophobic core in a molecule connects and melds distinct secondary-structural elements together. In the related RfScaE structure only one of these three tyrosines is conserved but the adjacent β-sheets are close enough (5.2 Å) to be stabilized by main-chain hydrogen bonding.

Figure 7.

Stabilization of the spatial position of the three α-helices by packing around three tandem tyrosine residues. The three tandem tyrosine residues and their hydrophobically interacting partners are presented as sticks coloured as follows: Tyr160 in green, Tyr161 in red and Tyr162 in purple.

Two hydrophobic clusters are observed in the structure of the RfCohG molecule. The first large hydrophobic cluster at one end of the cohesin comprises aromatic residues (Phe24, Phe129, Phe168, Tyr177, Tyr180 and Phe182) and aliphatic residues (Val27, Ile37 and Val56). The second hydrophobic cluster at the centre of the molecule includes aromatic residues (Tyr72, Phe90, Trp107, Phe124, Phe142 and Phe144), connecting the front and the back of the molecule and the loops, together with numerous aliphatic residues (Val14, Ile15, Leu39, Ile68, Ile70, Ile71, Ile72, Val88, Leu96, Ile102, Leu111, Val114, Leu123, Leu125, Leu158 and Ile190).

An additional observation that supports the stabilizing role of the α-helices in the RfCohG structure is evident from interface calculations between the two β-sheets of the molecule. Using the PISA server (http://www.ebi.ac.uk/msd-srv/prot_int/pistart.html; Krissinel & Henrick, 2007 ▶), analysis of the hydrophobic contacts in the core of the RfCohG and RfCohE structures was performed by calculating the contact area between the two β-sheets that form the β-sandwich with the α-helices included and between the same β-sheets only (eliminating the other secondary structures). Using this approach, it was found that RfCohG possesses a larger interface area of 912 Ų compared with 802 Ų for RfCohE. Moreover, when the calculation was performed without helix regions, the interface was found to be 412 and 681 Ų for RfCohG and RfCohE, respectively. The overall dimensions of β-sheet regions from RfCohG and RfCohE are very similar (r.m.s.d. of 2.8 Å). It thus follows that α-helical regions contribute to more than half of the interface area in the RfCohG molecule.

The additional helical elements revealed in the RfCohG structure show a continuum of progressive cohesin complexity (Fig. 5 ▶). The helices and the hydrophobic core may impart robustness to RfCohG and may play a role in the as yet unknown function of this unusual, freestanding cohesin.

Supplementary Material

PDB reference: cohesin from R. flavefaciens, 4n2o