Bachelor thesis, 15 hp Life Science/Chemistry 180 hp
Spring term 2019
Purification and crystallisation of streptococcal collagen
binding proteins
Henri Harry Colyn Bwanika
Supervisor: Karina Persson
1 Table of Contents
Abstract ... 2
Introduction ... 3-4 Materials and methods ... 5
Bacteria strains and plasmids ... 5
Expression of Cbm and Cnm ... 6
Protein Purification ... 6
Size Exclusion Chromatography ... 7
Protein Crystallization and model building ... 8
Cryoprotection of crystals ... 8
Mass Spectroscopy sample preparation ... 8
Results ... 9
Protein Purification of Cbm and Cnm ... 9
Protein Crystallization ... 10
Homology models of Cbm and Cnm ... 11-14 Comparison of the Electrostatic Potential Surfaces of Cbm and Cnm ... 14
Interaction with Collagen ... 15-17 Formation of Intramolecular Isopeptide bonds in Cbm and Cnm ...17-20 Discussion ...21-23 Concluding remarks and acknowledgments ... 23
References ... 24-28
2 Abstract
The Gram-positive bacterium Streptococcus mutans is an oral species which commonly resides in the oral cavity. The organism is mostly harmless but can be opportunistic and can cause life-threatening conditions such as infective endocarditis. S. mutans has been recently characterised into several serotypes including serotypes c, e, f and k. Unlike serotype c and e, S. mutans f and k express two major surface adhesin proteins also known as collagen binding proteins (CBPs). These two adhesin proteins are Cbm and Cnm which are critical components of S. mutans for its ability to bind to collagen and for subsequent evasion and colonisation of host cells. Although Cbm and Cnm proteins have been reported to be expressed in several other streptococci species by previous studies, the structures of the two protein as well as the mechanism they bind collagen is yet to be studied. During this project, I aimed to purify and crystallise Cbm and Cnm proteins and to elucidate their crystal structures. However, due to twinning and pseudo symmetry in the obtained protein crystals, it was not possible to solve the crystal structures of the proteins and homology models were used for this purpose. Cbm and Cnm are closely related to Cna protein from Staphylococcus aureus and it can be primarily assumed that their overall structures as well as folds are similar. Cbm and Cnm both comprise of two main domains N1 and N2 linked together by a flexible linker, each domain comprises of a classic β-sandwich fold. I also report the presence of isopeptide bonds in both Cbm and Cnm.
Intramolecular isopeptide bonds form autocatalytically between Lys-Asn or Lys-Asp residues and result in loss of NH3(17 Da) and H2O (18 Da) respectively. Mass spectroscopy (MS) provides one of the most definitive proof for presence of isopeptide bonds in a protein and here it was used to detect the presence of Lys-Asn isopeptide bond corresponding to a 17 Da deficit in MS data. Possession of such bonds could be of clinical relevance and could explain why Cbm and Cnm expressing S. mutans serotypes are more well adapted to mechanical and thermal stress.
3 Introduction
Streptococcus mutans is a Gram-positive bacterium that normally resides in the mouth on surfaces of the teeth and throughout the oral cavity. [1] The survival of S mutans in the mouth is greatly attributed to its ability to produce large amounts of glucans and acids which counterfeit with saliva buffering properties. [2] Although the organism is usually harmless, at times it can be an opportunistic pathogen of several life threatening conditions which include infective endocarditis, dental caries and cardiovascular diseases. S. mutans is one of the main causes of caries-it is both acidogenic and aciduric. [3] So far, the bacterium is classified into four different serotypes of c, e, f and k. [4, 5] Serotype/serotyping is a way of grouping closely related microorganism depending on the composition and modification of expressed surface proteins or antigens. Van der Beek et. al. have recently shown that serotypes of S. mutans are characterized depending on their composition of cell surface specific Rhamnose-glucose polymers (RGPs). [6] The composition of these four serotypes in the oral cavity can markedly vary however serotype c and e are highly prevalent in the oral cavity of a health person, making up over 90% of S mutans in the mouth as opposed to serotype f and k that make up only 5% . [7, 8] At roughly these percentages, the bacteria poses no serious harm to the host however, problems arise when the minor serotypes f and k increase in percentage. The two minor serotypes f and k have been shown to possess many virulent factors, for example surface adhesin proteins implicated to be key in causing several systemic conditions. [9]
Bacteria surface proteins are essential components in many Gram-positive bacteria due to their involvement in multiple bacterial processes, such as movement of the pathogen, attachment to surfaces and response to external signals. In serotype k and f of S. mutans, two main surface adhesin proteins also known as collagen binding proteins (CBPs) are highly expressed and play a crucial role in binding of type I collagen by S. mutans. [5, 10, 11] The two adhesin proteins, Cbm and Cnm are encoded by cbm and cnm genes respectively. The proteins have been shown to share approximately 80% identity and 90% similarity on a global
4
alignment plus high homology in the collagen binding domain. [12] Since serotypes k and f are they only S. mutans serotypes that express Cbm and Cnm, this prompted the preliminary hypothesis, that possession of Cbm and Cnm proteins could underlie the pathology related to k and f serotypes of S. mutans.
What’s more, previous studies have reported the presence of Isopeptide bonds in closely related collagen binding proteins like Cna, a collagen binding protein of Staphylococcus aureus and adhesin proteins of Corynobacterim diphtheriea. [13, 14, 15] Isopeptide bonds
(from the Greek isos, meaning equal), are amide bonds and poses the same structure as the classic peptide bond however they are formed between two protein groups, and at least one of the groups cannot be α-amino or α-carboxy group. [16] One variant of isopeptide bonds is the intramolecular isopeptide bonds that form between Lys and Asn side chains within the hydrophobic cores of the respective domains of the protein and have been reported to be well conserved in pili and fimbriae organelles of many gram-positive bacteria as well. [17]
Intramolecular isopeptide bonds are the main reason for the remarkable thermal stability and resistance to mechanical stress as well as proteolysis in many collagen and fibrinogen binding proteins. [18] Given the stability of Cbm and Cnm proteins, this prompted the second hypothesis of this project, that Cbm and Cnm proteins could possibly be stabilized by isopeptide bonds. Interestingly, preliminary results during this current project confirmed the presence of isopeptide bonds in both Cbm and Cnm proteins. These results were obtained from combining database search in the Protparam ExPASy server, [19] and comparing data from Mass-spectroscopy.
Although the proteins (Cbm and Cnm) have been well identified and characterised, their crystal structures and their collagen binding model are yet to be elucidated. During this project, I report the expression and purification of the collagen binding domains of Cbm and Cnm proteins. Furthermore, I aimed to crystallize Cbm and Cnm proteins however, the obtained crystals from the crystallisation trials were twinned and had pseudo-symmetry which made it
5
impossible to solve the structures of Cbm and Cnm proteins. Henceforth, comparative modelling was used to generate homology models which were used during this project to elucidate the probable structures and collagen binding models of Cbm and Cnm proteins.
Subsequently, it should be noted that all the structural results, as well as the hydrogen bond distances discussed in Fig.7 are entirely based on homology models and not actual diffraction data.
This study reports the probable structural models of Cbm and Cnm streptococcal adhesins as well as elucidating the interaction nature of these proteins with collagen with help of homology models. Henceforth, I also propose a probable explanation to why the two minor serotypes of Streptococcus mutans, k and f have a higher prevalence of causing life threatening conditions like infective endocarditis and dental caries.
Materials and methods
Bacterial strains and plasmids.
All bacterial strains and plasmids used during this thesis project are summarised in Table 1.
Escherichia coli BL21(DE3) is a laboratory strain which was obtained from the Structural Biology department at Umeå University and was modified to be chemically competent. The two plasmids that were used during the transformation assays of the cells were cnm and cbm plasmids both derivatives of pET-His1a cloning vector obtained from Gunter Stier, EMBL.
Table 1. Bacterial strains and plamsids used during the thesis project.
Strains/plasmids Relevant
phenotype/genotype
Source
BL21 Laboratory strain of
E. coli
This lab
Cnm plasmid pET-cnm km Gunter Stier,
EMBL
Cbm plasmid pET-cbmkm Gunter Stier,
EMBL
6 Expression of Cbm and Cnm proteins.
Expression of proteins were as previously described by Furuyama et. al, with minor changes.
[20] The genes encoding respective proteins of cbm and cnm including a His-tag in the upstream of the sequence but without the signal peptide were inserted into the pET-His1a cloning vector using NdeI and BamHI sites. 3 ul of the respective plasmids was added to 100 µl E. coli BL21 cells. Cells were briefly incubated on ice for 30 minutes. Taking up of the plasmid was induced by a 20 second heat shock procedure at 42 °C. 1 ml of S.O.C medium;
ThermoFicher Scientific (2% tryptone, 0.5% yeast extract, 10 mM NaCl, 2.5 mM KCl, 10 mM MgCl2, 10 mM MgSO4, and 20 mM glucose) was added to the cells. The cells were incubated at 37 °C for 1 hour with shaking and spread on a kanamycin selective Luria Bertani broth (LB) plate and incubated at 37 °C for overnight. Single colony of transformed BL21 cells harbouring respective plasmids were inoculated into 50 ml of 1×LB broth medium (2×Yeast extract and Tryptone (2×YT) medium with 20 mgml-1 kanamycin) and grown overnight at 37 °C with shaking. The OD600 of the overnight cultures were constantly measured until the OD600 nm was approximately 4. Approximately 24 ml of overnight cultures was added to 2000 ml of LB with kanamycin to a final OD600 nm of 0.05. The new cultures were incubated at 37 °C with shaking for 3 approximately 3 hours. When the absorbance of the cell culture at OD600 nm reached approximately 0.6, IPTG was added to a final concentration of 0.5 mM for the induction of protein expression. Expression of the protein was induced for at least 3 hours at 30 °C.
Protein purification and concentration
The purification procedure used during this project was adapted from Yang et. al, with minor changes. [21] In the earlier stages of purification, Immobilized metal-affinity chromatography (IMAC) was used. IMAC is based on the interaction between a metal ion (Ni2+ often) immobilized on the column matrix and specific amino acid residue. [22] For this purpose, the protein of interest was fused with a poly-histidine tag, as histidine amino acid also exhibits the
7
strongest interaction with immobilized metal ion matrices thus ensuring that the protein is retained on the column. Elution of the protein is done in high concentration of imidazole which displaces the histidine imidazole ring from the metal ion.
Consequently, after inducing protein synthesis, the cells containing the proteins were spin down at 4000 ×g for 20 mins. The resultant cell pellet was resuspended in a 35 ml Lysis buffer containing 50 mM Na-phosphate buffer pH 7.5, 0.2 M NaCl, 1o mM Imidazole. 500 ul of Triton was also added to mobilise the proteins from the cell membrane. The cells were sonicated for 3 minutes at intervals of 15 seconds. After sonication, the lysate was centrifuged at 23000 ×g for 20 mins in a JA25:50 Beckman Centrifuge to pellet the cell debris and membranes. After centrifugation, the soluble protein contained in the supernatant was purified by IMAC on a nickel–nitrilotriacetic acid resin column that was preequilibrated with the lysis buffer. 1 ml of the nickel–nitrilotriacetic acid resin beads was added to the supernatant and the mixture was incubated for 1 hour at 8 °C. The mixture of the beads and the supernatant was centrifuged for 8 minutes at 1500 RPM followed by 2 washes in 35 ml of wash buffer (50 mM Na-phosphate buffer pH 7.5, 0.2 M NaCl, 30 mM Imidazole) Elution of the protein from the resin column was performed with 20 ml elution buffer (50 mM Na-phosphate buffer pH 7.5, 0.2 M NaCl, 300 mM Imidazole). The protein was primarily concentrated with Amicon Ultra- centrifuge filters [10000] NMWL concentrators. (Merck KGaA, Darmstadt, Germany).
Size exclusion chromatography
The proteins were further purified by size exclusion chromatography (SEC) (Superdex 200 16/60; GE Healthcare Sweden). [23] Size exclusion chromatography or gel filtration is a protein purification technique that separates molecules on basis of size. Molecules that are larger than the matrix pores cannot diffuse onto the matrix pores and thus elute earlier during elution. Smaller molecules elute later since they go through the whole matrix length. During this step, a gel filtration buffer consisting of 20 mM Tris pH 7.5, 200 mM NaCl was first filtered with a cellulose acetate filter (0.2 µl pore size) and degassed in a Branson 3210 degassing
8
apparatus. Based on the curve generated using standard proteins, both Cbm and Cnm showed apparent molecular mass of ~ 35 kDa according to results from SEC chromatogram. The standard proteins used to generate calibration curve were based on GE healthcare life Sciences protein standards. [23] The SDS gel for Cnm also indicated that Cnm may exists as a dimer (Fig.1B). The purified proteins were collected and ultra filtered to final concentrations of 92 mgml-1 for Cnm and 86 mgml-1 for Cnm.
Protein crystallization and model building
Preliminary screening of crystallization conditions for Cbm and Cnm proteins was done using Structure and MIDAS screens respectively, [24] both from Molecular Dimensions Suffolk, CB8 7SQ, UK. MosquitoR crystal robot (TTP Labtech Ltd. Melbourn, Hertfordshire SG8 6EE United Kingdom) was used to pipette 1μl sitting drops of pure proteins onto a 96 well titter plate.
Crystals were grown at room temperature. All models and molecular simulations were done in CCP4MG software version 2.10.10. [25]
Cryoprotection and mounting of crystals onto loops.
Three KH2PO4 and PEG based cryosolutions were tested and the best solution was used for cryoprotection. Solution one comprised of 50 mm KH2PO4, 20% PEG 8000 and 20% glycerol.
The second solution comprised of 50 mm KH2PO4, 20% PEG 8000 and 20% PEG 400 and the last solution was made up of 50 mm KH2PO4, 20% PEG 8000 and 20% ethylene glycerol. The solutions were added directly onto the wells containing the crystals with initial reservoir solutions. Crystals were scooped from the cryosolution using a CryoLoop and froze with liquid nitrogen.
Sample preparation and analysis in Mass Spectrophotometry.
10 μl of the pure protein was added to a small Mass spectrophotometer vial and placed inside the chamber of the Mass spectrophotometer. The protein sample was washed in 10% 2- propanol and dissolved in a Formic acid based mobile phase. Innitial mobile phase consisted
9
of MQ water and 0.1% Formic acid. The sample was eluted in a mobile phase consisting of 100% Acetonitrile and 0.1% Formic acid.
Results
Protein purification of Cbm and Cnm
During the purification of Cbm and Cnm, the supernatant containing the expressed protein was applied to an affinity chromatography column, nickel–nitrilotriacetic acid resin column (Takara) preequilibrated with the same lysis buffer. The resultant eluate containing the respective protein was concentrated by ultrafiltration in Amicon cell to a final volume of 1ml followed by size exclusion chromatography on a Superdex 200 16/60 column. [23] Both Cbm (Fig.1A ) and Cnm (Fig.1B) featured a single main peak on the SEC chromatograms at elution volumes of 80-90 ml. For this elution volume, the average estimated molecular weight ranges between 29 kDa to 44 kDa which was deduced from known protein standards (data not shown). The standard proteins included Ribonuclease A (13.7 kDa), Carbonic anhydrase (29 kDa), Ovalbumin (44 kDa), Conalbumin (75 kDa) and Aldolase (156.8 kDa). To validate the observations, SDS-PAGE analysis of the fractions from the main peak was performed (fractions in lanes 4-7 on the respective gels) Cbm (Fig1.A gel) resulted in a single band at ~35 kDa on the resultant gel while Cnm (Fig.1B gel) featured a double band between ~37 to ~33 kDa. The double band on the Cnm gel could be because of an isopeptide bond, which does not break during the denaturing of the protein and may affect how the protein migrates on the gel.
Highly purified fractions (4-7 on the gel) of Cbm and Cnm were subsequently pooled and concentrated in Amicon cell tube to a final concentration of 92 mgml-1 and 86 mgml-1 respectively which were later used for crystallisation.
10
Figure 1. Purified CBM (A) and CNM (B) eluted during size exclusion chromatography. The purest fractions (lanes 4-7 on the gels) indicated by asterics were analysed on SDS gel using Simplyblue. The MW indicates the molecular mass marker in kDa. Lane 1 was the lysis sample. Lane 2 was the flow through sample from affinity chromatography step and lane 3 was eluate sample on affinity chromatography.
Protein crystallisation
Screening of crystallization conditions for Cbm and Cnm proteins was done with Structure and MIDAS screens by sitting drop. Crystals were grown in PEG-based conditions, with reservoir solution containing 20% PEG 8000 and 50 mM KH2PO4 at room temperature for Cnm only (Fig.2 A) after a period of 5 weeks. The crystallisation conditions for Cnm were optimized by
11
varying a wide range of the initial reservoir conditions. Figure 2.B shows the resultant diffraction image obtained from the crystals at Synchrotron Max IV in Lund, with the edge of the detector corresponding to 2.1Å
Figure 2. Image of CNM crystals obtained after 5 weeks in PEG based reservoir conditions (A) and the X-ray diffraction image of the crystal
Homology models of Cbm and Cnm
Except for having high resolution data- to better than 2 Å, the structure of Cnm could not be solved using molecular replacement method. The data indicated twinning and pseudo- translational symmetry which made structure determination and refinement very difficult.
Growing more crystals would have helped to overcome this problem.
A pairwise alignment performed between cbm and cnm using EMBOSS needle software (BLOSUM62 matrix, Gap penalty 10.0, extend penalty 0.5) showed that the proteins share a significant sequence identity of 79.1% and similarity of 89.2%. [26] Subsequently, it could be assumed that Cbm and Cnm also have very similar folds. Model building was carried out using the SWISS-MODEL building tool-ProMod3 1.3.0. [27] The collagen adhesin protein Cna from Staphylococcus aureus (PDB ID: 2f6a) [28] was used as template both for cbm and cnm since
12
2f6a had the highest identity/similarity and query coverage for both proteins, thus 52.36% and 52.38% identity for cbm and cnm respectively. [27] Query coverage was >90% in both cases. A multiple sequence alignment between Cna with Cbm and Cnm (Fig.3) shows the most likely model of the secondary structure of Cbm and Cnm illustrated using arrows (for beta strands) and coils (for alpha helices) as well as a consensus sequence that represents the three proteins.
[29]
Figure 3. Multiple sequence alignment of the Cna, Cbm and Cnm proteins. Possible model of the secondary structure of the three proteins is illustrated by use of arrows(beta-sheets) and coils (alpha helices). Cna was used as template during model building for cbm and cnm.
The homology models of Cbm and Cnm are shown in Fig. 4A and 4B respectively which depicts very identical folds and similar quaternary structure of Cbm and Cnm. As previously been noted in Cna protein by Zong et. al, Cbm and Cnm proteins also exhibit two distinct domains (Fig. 4), that’s N1 and N2 domains. [28] Superposition of the 2 proteins in Fig. 4C (blue-Cbm and pink-Cnm) resulted in an RMSD 0.25 Å for 294 residues which implies very close relatedness between the two proteins which also goes to show that the proteins carry out same or similar functions for the cell. Fig. 4D shows a sketch representation of the main secondary structures of Cbm and Cnm which were generated by comparative modelling in CCP4MG software. Superposition of the Cbm and Cnm to the known template structure of Cna resulted
13
in RMSD values of 0.21 Å and 0.32 Å respectively both values were generated from superposing 294 residues. This shows that all the three proteins are very similar at the structural level, also from judging by the RMSD value for Cbm and Cna, it can be concluded that Cbm and Cna are even more similar to each other.
In Fig. 4D, the main β-strands and α-helices are numbered from 1 to 26 and coloured in a rainbow colours to easily visualise where the protein starts and ends. 24 β-strands and 2 α- helices have been illustrated in the sketch. The flexible linker which connects the two distinct domains of the proteins is labelled with an arrow. This linker is crucial for binding of collagen as it forms the enclosure that stabilizes the protein-collagen complex. Secondary structures less than 3 residues are not shown in Fig 4.D but a detailed 2D sketch of the protein was shown in Fig.3. Both proteins also contain long stretches of putative B-repeats and an LPXTG motif in the C-terminal (not included in the sequences) which has been earlier reported by Nomura et. al. These repeats, as well as the LPXTG motif are very characteristic of most Gram-positive adhesin proteins [10] The LPXTG acts to anchor the new synthesised protein onto the cell wall of the bacteria. In both Cbm and Cnm, the repeats are full of Ser and/or Thr implying that they are probably glycosylated when expressed on the surface of the bacteria.
14
Figure 4. The homology models of (A) cbm and (B) cnm. (C) Superposition comparison between the two models, cbm is labeled in blue while cnm is pink. (D) An overall sketch that shows the main secondary structures of the proteins. The flexible linker that connects the two domains of the protein is shown using an arrow in all models.
Comparison of the electrostatic potential surface for Cbm and Cnm
However-much cbm and cnm exhibit very high similarity and identity in their sequences as well as structural folding, there are a few features that distinguish these two closely related proteins. One of the main differences can be spotted by observing the electrostatic potential surfaces of the two collagen binding proteins. (Fig.5) The great diversity in electrostatic properties between the two proteins can be seen clearly from the fact that Cbm in Fig. 5A is very electropositive (blue) around the collagen binding pocket also known as the interdomain
15
hole compared to Cnm (Fig. 5B) which shows slightly stronger electronegative (red) properties in its interdomain hole as well as all over its surface. Another observed difference between the two proteins exists in the nature of the B-repeats expressed at the C-terminal of the protein.
The two proteins have slightly different repeats (not included in sequence)
Figure 5. Electrostatic potential surfaces of CBM(A) and CNM(B). The flexible cross-linker is facing up on the right and facing down on the left after flipping the models 180°
Interaction with collagen
The homology model of only Cbm in complex with the collagen triple helix is shown in Fig. 6 for simplification since there is little difference compared to that of Cnm. The anterior view of the complex in Fig. 6A shows the collagen triple helix penetrating the hole between the two main domains of Cbm while the lateral view in Fig. 6B shows the collagen triple helix in the
16
plane angle running from right to left, green is the leading chain, blue is middle chain and yellow is the trailing chain. A closer view of the interaction between Cbm and
collagen is shown in Fig.7 with the residues that are most critical during this interaction and stabilization of the complex. The identified direct interactions between collagen and Cbm were only restricted to the middle chain of collagen blue. The polar residue Asn192 which is part of the N2 domain forms a H-bond with the middle chain of collagen (Fig. 7A). This residue lies within the electropositive core of the interdomain hole in the protein and may be important in stabilizing the complex. The other residue implied in the stability of the protein is Thr165 which is also polar in nature (Fig. 7B). This residue marks the start of the flexible linker which was marked before in Fig. 4, a closer examination of this linker (Fig. 7C) that its composed of 10 residues in total running from Lys160 to Tyr170. This linker plays a crucial role in holding the ligand in place via interaction with Thr165. The topological positions of the critical residues is also shown in Fig. 7D.
Figure 6. The homology model of cbm-collagen peptide complex. (A) with collagen triple helix running in and out the interdomain hole which is critical for collagen binding and (B) with collagen in the plane angle.
17
Figure 6. Identified critical residues for interaction of Cbm with collagen triple helix. (A) H-bond formed between the polar residue Asn192 with 3Å. (B) The other interaction between collagen and cbm is the H-bond with Thr165 measuring 2.7Å. (C) The flexible linker (Thr165 -tyr170) is critical for holding collagen in place. (D) Overview of the positions of the critical residues interacting with the middle chain of collagen.
Formation of intramolecular iso-peptide bonds in Cbm and Cnm
Isopeptide bonds have been extensively documented to be present in several Gram-positive bacteria, unlike Gram negative bacteria that mostly boast disulphide linkages. Isopeptide bonds cross-link the protein model hence stabilizing the protein against stress. Fig.8 outlines the main steps involved in formation of a typical isopeptide bond. When an iso-peptide bond is formed, there is release of either NH3 or H2O. (Fig.8 circle in red) This depends on whether
18
the reaction is between Lys residue and Asn (NH3) or between Lys and Asp (H2O). The reaction results in a strong interaction that does not depend on redox reactions. A stabilizing acid, either Glu or Asp in proximity of the formed isopeptide bond makes the interaction even stronger by forming H-bonds -final product (see Fig.8)
Figure 7. Outline of the formation of isopeptide bonds. (Top) The reaction between Lys and Asn results in release of NH3. (Bottom) The reaction between Lys and Asp results in release of H2O. The stabilizing acid (Glu/Asp) position on top of the reaction forms H-bonds after formation of the isopeptide bond-final product.
Isopetide bonds were identified in both Cbm and Cnm homology models during this study (Fig.9). In Cbm, the isopeptide bond is formed between Lys172 and Asn286 which implies that NH3 is released, in the process Asp205 acts as stabilizing acid in this reaction (Fig. 9A). This isopeptide bond crosslinks three different β-sheets in the N2 domain, that is 13, 22 and 15.
(Fig. 9B) for the numbering of the β-strands refer to Fig. 4D. This crosslinking highlights the importance of isopeptide bonds in stabilzing the protein via interactions between distant
19
secondary protein structures. Fig. 9C shows the position of the bond in the model, thus sandwiched by parallel β-sheets and burried within the hydrophobic core implying that its important in holding the N2 domain together.
Figure 8. Intramolecular isopeptide formation. (A) Lys172 and Asn286 are involved in isopeptide bond formation and stabilized by Asp205. (B) Distant β.sheets crosslinked by the bond are 13,22 and 15. (D) Position of the formed bond in N2 domain of the protein
Presence of the isopeptide bonds has been previously comfimed by means of several biophysical methods for example NMR and Mass spectroscopy. Mass spectrophotometry offers one of the best definitive characteristion of type of isopeptide bonds preesent in a protein. The presence of Lys-Asn isopeptide is comfirmed if the mass obtained by MS is 17 Da lower than the expected or theoretical value and this difference can be explained by the loss of NH3. For Lys-Asp, mass obtained from MS should be 18 Da lower than expected value to
20
comfirm the presence of this particular isopeptide bond and this loss of mass corresponds to a H2O molecule released. During this project, Electrospray ionization time-of-flight (ESI-TOF) was used for the purpose of comfirming the presecence of isopeptide bonds in Cbm and Cnm.
(Fig.10)
Figure 9.Mass spectroscopy. ESI-TOF comfimed the presence of Lys-Asn isopeptide bonds in both Cnm and Cbm.
(A) Main peak represnts molecular weight for cnm and (B) main peak represents molecular weight for cbm
Theoretical values for the mass of the proteins were obtained from ProtParam ExPASy server.
[19] For Cnm, the following paraeters were computed (No. aa: 321, Molecular weight: 35179.25 and theoretical I.p: 4.86). For Cbm, (No. aa: 324, Molecular weight: 35432.78.68 and theoretical I.p: 5.46). Molecular weight values obatined from MS are shown in Fig. 10A for Cnm and Fig. 10B for Cbm. After doing the simple math, Cnm resulted in a 16.77 Da deficit while Cbm had a deificit of 16.72. Both values are ~17 Da henceforth comfirming the presence of a Lys-Asn isopeptide bond.
21 Discussion
Surface adhesin proteins are very indispensable components of almost all bacterial cells and critical features for attachment of bacteria onto the extracellular matrix (ECM) of the host cells.
During this thesis project, I report the expression and purification of Cbm and Cnm proteins.
I also report my crystallisation trials in attempt to elucidate the structure of the two recently characterised S. mutans surface adhesin proteins; Cbm and Cnm. The two proteins Cbm and Cnm fall into a large family of adhesin proteins collectively called Microbial surface components recognizing adhesive matrix molecules (MSCRAMMs) family which are categorized on basis of their structural similarities and common mechanism for ligand binding.
[30] The MSCRAMM family of proteins aid bacteria in adhering to a range of host cell ECM, such as collagen that has been studied during this project, other ECMs include fibrinogen, fibronectin and laminin. [31, 32, 33, 34] There are several proposed mechanisms by which MSCRAMMs archive attachment to ECM, but the most common ones and extensively studied mechanisms are the ‘Dock, Lock, Latch’ (DLL) and the collagen hug model. [30, 35]
Analysis of the homology models of Cbm and Cnm shows that these two proteins employ the
‘collagen hug’ model which has been elucidated by several previous studies. Especially, Cbm and Cnm are very similar, both on sequence and structural level to Cna which is collagen binding MSCRAMM of S. aureus. This is even more true for Cbm which had as low as 0.21 Å for RMSD value when superposed to Cna protein. This implies that the proteins have very similar folds and could most likely be homologs. That is, they could share a common ancestor and carry out similar functions for the cell. The RMSD of 0.25 Å obtained from superposing Cbm and Cnm also has similar implications. What is not know is whether expression of Cbm and Cnm at once, by the cell may have synergetic activity or is simply redundancy in which case the proteins carry out exact the same function. This could be an opener for future research questions regarding the functions of the two proteins, Cbm and Cnm.
22
Subsequently, from comparison analysis of Cbm and Cnm to Cna, Cbm and Cnm fulfil the criteria of collagen hug model described by Zong et. al which include, (a) collagen binding being initiated through hydrophobic interactions between residues in shallow trench, (b) a flexible linker connecting the two domains N1 and N2, (c) ability of proteins to shrink its interdomain hole to tightly secure the ligand and (d) the insertion of the latch in form of N2 domain extension which folds back onto N1 domain hence securing the bound ligand or peptide. [28, 36] Binding specificity to collagen in Cbm and Cnm is highly dependent on hydrophobic interaction and formation of H-bonds between specific residues and collagen.
Specifically, in case of Cbm, residues Thr165 and Asn192 among others could be critical for effective collagen binding and securing the ligand in place. The flexible linker that connects the two domains N1 and N2 is also important for securing the ligand in place and stabilizing the protein-collagen complex since it contributes highly to the interaction of the protein with collagen. There are possibly several other interactions that have not been reported in this study due to poor geometries in the homology models that were generated and this could be further improved for a better analysis of these interactions.
Interestingly, during this thesis project I identified the presence Iso-peptide bonds in both Cbm and Cnm proteins. Analogous to sulphide bonds found in several Gram-negative bacteria, iso-peptide bonds have been reported to be adapted and widely spread in Gram-positive bacteria. [17, 15, 37] This could be an evolutionary advantage for Gram-positive bacteria since iso-peptide bonds are much stronger and much stabilizing forces than the conventional disulphide bonds. The biological implications of this is that, most Gram-positive bacteria with isopeptide bonds embedded in their adhesin proteins can be well versed in harsh environments. For instance, certain serotypes of S. mutans readily and easily colonise niches like the oral cavity which is rather a hostile environment for most Gram-negative as well as Gram-positive bacteria due to high mechanical and thermal stress in the mouth. [5] Some studies have gone further to illustrate that indeed isopeptide bonds are critical even in other
23
bacteria appendages like in pili proteolytic stability. Eliminating isopeptide bonds either in BcpA protein of B. cereus or in major pili protein of Streptococci pyogenes caused dramatic destabilization of the proteins. [38, 39]
What’s more, in terms of clinical relevance, previous studies have shown that certain serotypes of S. mutans are more implicated to be involved in various life-threatening conditions like dental caries and infective endocarditis. For example, the minor serotypes of S. mutans, k and f are prevalent in patients with infective endocarditis. [7] This prevalence could be attributed to possession of Cbm and Cnm which are key factors during invasion and colonisation of f and k S. mutans.
In conclusion, this research project elucidates and report models of S. mutans adhesin proteins Cbm and Cnm based on comparative modeling, the nature of iso-peptide bonds that stabilize Cbm and Cnm adhesin proteins making them important virulence factor. The analysis of how Cbm and Cnm interact with collagen can pave a way on how this interaction can be attenuated in patients with Cbm and Cnm positive S. mutans strains.
Acknowledgements.
My sincere gratitude to Professor Karina Persson and the whole of Structural Molecular biology department for granting me the opportunity of doing my thesis work in her well- equipped laboratory and the privilege to work on such a thrilling project.
I am also very grateful to Thomas Heildler who was so helpful to me during my labs to find certain equipment and reagents, even more so for helping me with Mass Spectroscopy. I would also like to thank Ragha who was always rendering a hand when I needed assistance with performing certain assays.
24 References.
1. Zero DT, Fontana M, Martínez-Mier EA, Ferreira-Zandoná A, Ando M, González- Cabezas C, Bayne S. The biology, prevention, diagnosis and treatment of dental caries:
scientific advances in the United States. J Am Dent Assoc. 2009;140 Suppl 1:25S-34S.
2. Falsetta ML, Klein MI, Lemos JA, Silva BB, Agidi S, Scott-Anne KK, Koo H. Novel antibiofilm chemotherapy targets exopolysaccharide synthesis and stress tolerance in Streptococcus mutans to modulate virulence expression in vivo. Antimicrob Agents Chemother. 2012;56(12):6201-11.
3. Gross EL, Beall CJ, Kutsch SR, Firestone ND, Leys EJ, Griffen AL. Beyond Streptococcus mutans: dental caries onset linked to multiple species by 16S rRNA community analysis.
PLoS One. 2012;7(10):e47722.
4. Nakano K, Nomura R, Nakagawa I, Hamada S, Ooshima T. Demonstration of Streptococcus mutans with a cell wall polysaccharide specific to a new serotype, k, in the human oral cavity. J Clin Microbiol. 2004;42(1):198-202.
5. Lapirattanakul J, Nomura R, Nemoto H, Naka S, Ooshima T, Nakano K. Multilocus sequence typing of Streptococcus mutans strains with the cbm gene encoding a novel collagen-binding protein. Arch Oral Biol. 2013;58(8):989-96
6. Van der Beek SL, Zorzoli A, Çanak E, Chapman RN, Lucas K, Meyer BH, Evangelopoulos D, de Carvalho LPS, Boons GJ, Dorfmueller HC, van Sorge NM. Streptococcal dTDP-L- rhamnose biosynthesis enzymes: functional characterization and lead compound identification. Mol Microbiol. 2019;111(4):951-964
7. Nakano K, Nomura R, Nemoto H, Mukai T, Yoshioka H, Shudo Y, Hata H, Toda K,Taniguchi K, Amano A, Ooshima T. Detection of novel serotype k Streptococcus mutans in infective endocarditis patients. J Med Microbiol. 2007;56 (Pt10):1413-5 8. Kojima A, Nakano K, Wada K, Takahashi H, Katayama K, Yoneda M, Higurashi
T,Nomura R, Hokamura K, Muranaka Y, Matsuhashi N, Umemura K, Kamisaki Y,
25
Nakajima A, Ooshima T. Infection of specific strains of Streptococcus mutans, oral bacteria, confers a risk of ulcerative colitis. Sci Rep. 2012;2:332.
9. Nakano K, Nemoto H, Nomura R, Homma H, Yoshioka H, Shudo Y, Hata H, Toda K,Taniguchi K, Amano A, Ooshima T. Serotype distribution of Streptococcus mutans a pathogen of dental caries in cardiovascular specimens from Japanese patients. J Med Microbiol. 2007;56(Pt 4):551-6
10. Nomura R, Nakano K, Taniguchi N, Lapirattanakul J, Nemoto H, Grönroos L,Alaluusua S, Ooshima T. Molecular and clinical analyses of the gene encoding the collagen-binding adhesin of Streptococcus mutans. J Med Microbiol. 2009;58(Pt4):469-75.
11. Sato Y, Okamoto K, Kagami A, Yamamoto Y, Igarashi T, Kizaki H. Streptococcus mutans strains harboring collagen-binding adhesin. J Dent Res. 2004;83(7):534-9.
12. Nomura R, Nakano K, Naka S, Nemoto H, Masuda K, Lapirattanakul J, Alaluusua S,Matsumoto M, Kawabata S, Ooshima T. Identification and characterization of a collagen-binding protein, Cbm, in Streptococcus mutans. Mol Oral Microbiol.
2012;27(4):308-23.
13. Herman-Bausier, Philippe et al. “Mechanical Strength and Inhibition of the Staphylococcus aureus Collagen-Binding Protein Cna.” mBio vol. 7,5 2016; e01529-16.
14. Kang HJ, Paterson NG, Gaspar AH, Ton-That H, Baker EN. The Corynebacterium diphtheriae shaft pilin SpaA is built of tandem Ig-like modules with stabilizing isopeptide and disulfide bonds. Proc Natl Acad Sci U S A. 2009;106(40):16967-71.
15. Kang HJ, Coulibaly F, Clow F, Proft T, Baker EN. Stabilizing isopeptide bonds revealed in gram-positive bacterial pilus structure. Science. 2007;318(5856):1625-8.
16. Kang HJ, Baker EN. Intramolecular isopeptide bonds: protein crosslinks built for stress?
Trends Biochem Sci. 2011;36(4):229-37.
26
17. Hu X, Hu H, Melvin JA, Clancy KW, McCafferty DG, Yang W. Autocatalytic intramolecular isopeptide bond formation in gram-positive bacterial pili: a QM/MM simulation. J Am Chem Soc. 2011;133(3):478-85
18. Sridharan U, Ponnuraj K. Isopeptide bond in collagen- and fibrinogen binding MSCRAMMs. Biophys Rev. 2016;8(1):75-83.
19. Gasteiger E., Hoogland C., Gattiker A., Duvaud S., Wilkins M.R., Appel R.D., Bairoch A.;
Protein Identification and analysis Tools on the ExPASy Server; (In) John M. Walker
(ed): The Proteomics Protocol Handbook, Humana Press (2005) pp. 571-607
20. Furuyama T, Ishii Y, Ohya N, Tateda K, Hanson ND, Shimizu-Ibuka A. Purification, crystallization and preliminary X-ray analysis of IMP-18, a class B carbapenemase from Pseudomonas aeruginosa. Acta Crystallogr Sect F Struct Biol Cryst Commun. 2013;69(Pt 12):1397-400.
21. Yang XY, Li ZQ, She Z, Geng Z, Xu JH, Gao ZQ, Dong YH. Structural analysis of Pseudomonas aeruginosa H3-T6SS immunity proteins. FEBS Lett. 2016;590(16):2787- 96.
22. Bornhorst JA, Falke JJ. Purification of proteins using polyhistidine affinity tags. Methods Enzymol. 2000;326:245-54.
23. GE Healthcare Life Sciences Sweden. 2018. HiLoad Superdex 200 pg preparative size exclusion chromatography columns. A handbook
24. Molecular dimensions. Jancarik, J & Kim, S.H.J. (1991), J. Appl.Cryst.2003; 24, 409- 411. Wooh et al, (2003), Acta Cryst , D59, 769 – 772.
25. McNicholas S, Potterton E, Wilson KS, Noble ME. Presenting your structures:the CCP4mg molecular-graphics software. Acta Crystallogr D Biol Crystallogr. 2011;67(Pt 4):386-94
26. Needleman S.B. and Wunsch C.D. A general method applicable to the search for similarities in amino acid sequences of two proteins. J.Mol.Biol. 1970;48(3):443-53
27
27. Waterhouse, A., Bertoni, M., Bienert, S., Studer, G., Tauriello, G., Gumienny, R., Heer, F.T., de Beer, T.A.P., Rempfer, C., Bordoli, L., Lepore, R., Schwede, T. SWISS-MODEL:
homology modelling of protein structures and complexes. Nucleic Acids Res. 2018;
46(W1), W296-W303
28. Zong Y, Xu Y, Liang X, Keene DR, Höök A, Gurusiddappa S, Höök M, Narayana SV. A 'Collagen Hug' model for Staphylococcus aureus CNA binding to collagen. EMBO J.2005;24(24):4224-36.
29. Notredame C., Higgins D.G. and Heringa J. T-Coffee: A novel method for fast and accurate multiple sequence alignment. Journal of Molecular Biology 2000; 302(1):205- 17
30. Berne C, Ducret A, Hardy GG, Brun YV. Adhesins Involved in Attachment toAbiotic Surfaces by Gram-Negative Bacteria. Microbiol Spectr. 2015;3(4).
31. Konkel ME, Larson CL, Flanagan RC. Campylobacter jejuni FlpA binds fibronectin and is required for maximal host cell adherence. J Bacteriol. 2010;192:68–76
32. Hienz SA, Schennings T, Heimdahl A, Flock JI. Collagen binding of Staphylococcus aureus is a virulence factor in experimental endocarditis. J Infect Dis. 1996;174:83–88.
33. Sridharan U, Ponnuraj K. Isopeptide bond in collagen- and fibrinogen-binding MSCRAMMs. Biophys Rev. 2016;8(1):75-83.
34. Foster TJ, Geoghegan JA, Ganesh VK, Höök M. Adhesion, invasion and evasion:the many functions of the surface proteins of Staphylococcus aureus. Nat Rev Microbiol.
2014;12(1):49-62.
35. Ponnuraj K, Bowden MG, Davis S, Gurusiddappa S, Moore D, Choe D, Xu Y, Hook M,Narayana SV. A "dock, lock, and latch" structural model for a staphylococcaladhesin binding to fibrinogen. Cell. 2003;115(2):217-28.
28
36. Liu Q, Ponnuraj K, Xu Y, Ganesh VK, Sillanpää J, Murray BE, Narayana SV, Höök M.
The Enterococcus faecalis MSCRAMM ACE binds its ligand by the Collagen Hug model.
J Biol Chem. 2007;282(27):19629-37.
37. Hagan RM, Björnsson R, McMahon SA, Schomburg B, Braithwaite V, Bühl M, Naismith JH, Schwarz-Linek U. NMR spectroscopic and theoretical analysis of a spontaneously formed Lys-Asp isopeptide bond. Angew Chem Int Ed Engl. 2010;49(45):8421-5.
38. Kang HJ, Baker EN. Intramolecular isopeptide bonds give thermodynamic and proteolytic stability to the major pilin protein of Streptococcus pyogenes. J Biol Chem.
2009;284(31):20729-37.
39. Budzik JM, Poor CB, Faull KF, Whitelegge JP, He C, Schneewind O. Intramolecular amide bonds stabilize pili on the surface of bacilli. Proc Natl Acad Sci U S A.
2009;106(47):19992-7.
Department of Chemistry Umeå university
901 87 Umeå, Sweden Telephone +46 90 785 28 69 www.kemi.umu.se
