The Expressed Fxii Rfps Are Soluble Biology Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

SDS-PAGE and western blot analyses have proved that all the FXII-RFPs have been successfully expressed. As it were the supernatants (soluble fraction) of the cell lysates that were purified, these analyses show that the FXII-RFPs are soluble. The homology models of FNII and EGF-1 domain of FXII (Figures 3.14 and 3.18) show that these domains have got two and three disulphide bonds respectively. Thus, post-translational disulphide bonds formation is essential for the stability of the FXII-RFPs and thus their solubility. In situations where disulphide bonds have failed to form, the unfolded proteins formed are unstable and will either be found in inclusion bodies of E.coli or undergo degradation (Peisley and Gooley, 2007). Hence, this study has demonstrated that fused Trx-tag has aided disulphide bonds formation in the FXII-RFPs leading to their solubility.

4.1.2 IMAC: An Analytical and Characterisation Tool

The concept of His-tag purification via immobilised metal affinity chromatography (IMAC) is that the histidine residues will adsorb strongly to the column based on interaction between Ni2+ and the electron-donating groups on histidines (Figure 4.1). The retention of a protein in a metal ion column is influenced by the spatial distributions and accessibilities of histidine residues on its surface (Hemdan et al, 1989; Gaberc-Porekar and Menart, 2001; Ueda et al, 2003). Although histidine is a common amino acid residue in natural proteins, histidines on a protein surface will bind to Ni2+ with slightly lower affinity as compared to a flexible His-tag (Gaberc-Porekar and Menart, 2001; Block et al, 2009).

Figure 4.1 A schematic representation of a His-tagged protein bound to metal-chelated affinity support. The concept of IMAC is based on the interaction between immobilised metal ions, in this case nickel ions, which serve as the electron-pair acceptors and the amino acid residues on the protein surface, in this case histidine residues, which serve as electron donors. (Adapted from Gaberc-Porekar and Menart, 2001).

The FXII-RFPs in this study were eluted at imidazole concentrations ([imidazole]) that range between 440nm to 510nm. These concentrations are rather high as compared to previous report that 20-250mM imidazole is effective for the elution of 6xHis-tagged protein (Terpe, 2003). In the homology models of FXII structural domains (Figure 4.2), it was found that all the histidine residues in these domains are exposed to the solvent accessible surface, the functional surface with respect to the surrounding solvent (Ridgen, 2009; chapter 7). Thus, these surface exposed histidines can bind to Ni2+ and consequently contribute to the high affinity of these FXII-RFPs for Ni2+ column.




Figure 4.2: Cartoon representations of the homology models of the various structural domains of FXII showing the solvent accessible surface. The colour of the surface corresponds to the colour of the underlying molecule. These models show that the histidine residues (blue) are exposed on the surface. (A) Homology model of FXII1-71.The N-terminal region of unknown homology (amino acids 1-22) is shown in purple, the FNII domain (amino acids 23-71) is shown in teal; (B) Homology model of FXII23-71, the FNII domain; (C) Homology model of FXII EGF-1 domain (amino acids 75-112).


Imidazole concentration %/mM (at which proteins were eluted)

Number of histidine residues (excludes His-Trx tag)












11However, it is interesting that the [imidazole] required to elute the FXII-RFPs do not correlate with the number of histidine residues. f-FXII23-112 and f-FXII23-154, which have more histidine residues as compared to f-FXII1-71 and f-FXII23-71, were eluted at lower [imidazole] (Table 4.1) which indicates that f-FXII23-112 and f-FXII23-154 have lower binding affinity to Ni2+ column as compared to f-FXII1-71 and f-FXII23-71. This may suggest that the histidine residues in these larger proteins are less accessible for binding to Ni2+.

Table 4.1: Table shows the imidazole concentration at which the FXII-RFPs were eluted and total number of histidine residues in each FXII-RFPs.

It has been reported that IMAC can be exploited as an analytical tool to probe the topography of histidine residues of a protein molecule (Hamden, 1989). Thus further investigations of the behaviour of these FXII-RFPs in IMAC may reveal important information in the spatial distribution of the histidine residues. As it has been reported that histidines play a major role in the surface binding activity of FXII (Samuel et al, 1993), information about their distribution on the surface of FXII is crucial for the investigation of FXII binding to negatively-charged surfaces.

The chromatogram for f-FXII23-71 (Figure 3.2B) shows that its elution produced a single sharp peak. This indicates that the f-FXII23-71 proteins have similar affinity for Ni2+ which may suggest that they have adopted a similar conformation. This is the same for f-FXII1-71, f-FXII23-112 and f-FXII23-154 (Figure 3.2A, C, D). The implication of this observation will be discussed further in section 4.2.3.


4.2.1 Analysis of Protein Elution via GFC

The principle of protein separation in GFC is based on the hydrodynamic volume of a protein, the volume which it takes up when it is in the solution. The pores in the gel filtration column serve as molecular sieves where the retardation time of a protein in a column will depend on its shape as well as molecular weight (Rosenberg, 2002). Thus, two proteins of similar MW which behave differently in solution may have different hydrodynamic volumes and thus different retention time in a column (Rodríguez-Díaz et al, 2005).

In the chromatogram for f-FXII23-71 (Figure 3.5B), a small leading peak is visible right before the sharp peak produced by protein elution from the gel filtration column S75 GL. When the second half of f-FXII23-71 was purified using a different column, S75 PG, the presence of a small leading peak can again be noticed in the chromatogram (Figure 3.7A). The difference between the above two mentioned columns is their bed volume. S75 GL has a bed volume of 24ml whereas S75 PG has a bed volume of 120ml; column with a larger bed volume has a higher separation resolution. Thus, it is unlikely that this small leading peak is caused by a contaminant of different MW or f-FXII23-71 of a different conformation because S75 PG, which has a higher separation resolution, would be able to separate that contaminant resulting in a distinct peak.

The chromatograms of f-FXII1-71, f-FXII23-71 and f-FXII23-112 (Figure 3.6 and Figure 3.7) show that the elution of FXII-RFPs from gel filtration column produced single sharp A280nm peak. These show that the eluted FXII-RFPs were highly homogeneous.

4.2.2 GFC as an Analytical Tool

GFC has been widely used to study conformation changes of protein folding and unfolding (Martensen, 1978; Bates et al, 1997; Uversky et al, 2001) because proteins of the same that are folded and unfolded will behave differently in GFC due to their different hydrodynamic volumes. In this study, the estimated MW of f-FXII1-71 and f-FXII23-71 were quite close to their theoretical MW (Table 3.3). As globular proteins of known MW were used for calibration of the S75 GL, this MW estimation shows that f-FXII1-71 and f-FXII23-71 have got a globular shape. In addition, f-FXII1-71 was eluted with sharp single peak, suggesting that they are of the same species with similar conformation. This is also the same for f-FXII23-71.

These indicate that the fused Trx was sufficient for disulphide bonds formation in the FXII-RPs resulting in all of them having a similar globular conformation. The basis of this suggestion is that if the proteins were not folded, their estimated MW will be larger than what were obtained because unfolded proteins will adopt a random coil configuration and consequently have a larger hydrodynamic volume as oppose to compact globular proteins.

Besides this, the estimated MW also shows that f-FXII1-71 and f-FXII23-71 exist as monomers. The other column used in this study, S75 PG, was not calibrated, thus it is not possible to estimate the MW of f-FXII23-112. However, based on (1) the chromatogram (Figure 3.7B) which shows a sharp single peak; (2) the difference in the Ve of f-FXII23-71 and f-FXII23-112 was not huge, it is unlikely that f-FXII23-112 have formed aggregates.


4.2.3 Role of Trx in Disulphide Bonds Formation

Observations from both IMAC and GFC have provided strong evidence that the Trx-tag alone is sufficient for post-translational disulphide bond formation in the FXII-RFPs.

In the cytoplasm of E.coli, thioredoxin reductase (the product of the trxB gene) maintains thioredoxin 1 and 2 (encoded by trxA and trxB respectively) in their reduced state. These thioredoxins will in turn reduce thiol groups of proteins in the cytoplasm (Stewart et al, 1998; Jurado et al, 2006). The reducing environment is the main reason why it is not possible for formation of disulphide bond in the E.coli cytoplasm (Peisley and Gooley, 2007). It has been proposed that when thioredoxin reductase is absent, the thioredoxins will serve as oxidant instead of their usual role as reductants. Thus, the use of Trx fusion protein together with trxB deficient E.coli cells will enhance oxidation of disulphide bonds (Basette et al, 1999). In a previous study by Peisley and Goole (2007), the FNII domain of MMP-2 was successfully expressed in BL21trxB (DE3) cells, the mutant strain of BL21 which is deficient in thioredoxin reductase, as recombinant proteins with a Trx-tag. There are also various examples where the use of thioredoxin fusion protein in BL21trxB (DE3) were successful in the expression of soluble proteins with multiple disulphide bonds (Wilkinson et al, 2004; Jurado et al, 2006).

In this experiment, the FXII-RFPs were expressed in BL21 (DE3) cells, not the mutant strain which is deficient in thioredoxin reductase. Yet, as mentioned above, the formation of disulphide bonds were successful. There are also many other studies where the use of Trx-fusion tag in BL21 (DE3) have successfully expressed the fusion proteins as soluble proteins (Ribas et al, 2000; Simmons et al, 2006; Wang et al, 2008). So how did Trx-tag mediate oxidation of disulphide bonds in the presence of thioredoxin reductase in BL21 (DE3)? Trx has been demonstrated to aid protein folding in a mechanism that is independent of their redox activity in a study by Berndt et al (2008). In their study, Trx with mutations at the 'active site' which is important for catalysis of oxidative protein folding was as efficient as the wild type Trx fusion partner in the induction of correct folding. It has been suggested that Trx which is fused to a protein of interest may act as molecular chaperone to prevent the precipitation and aggregation of the fused nascent partner proteins until they have reach a stable folding state. Trx is a compact, highly soluble protein with robust folding characteristics (LaVallie et al, 2000). Its high solubility may prevent the aggregation of the fused proteins while its ability to reach native conformation rapidly may promote the downstream fusion partner to adopt their correct structure as well.


Although MALDI-TOF can determine protein mass with an accuracy of at least 0.1 per thousand, it works best with proteins from 30-40kDa (Rehm, 2006). The MW of His-Trx tag, FXII1-71 and FXII23-71 are approximately 13.8kDa, 8.45kDa and 5.86kDa respectively. Peptide fragments generated from these proteins with such low MW will not be sufficient to match against protein databases which explain why results obtained with MALDI-TOF analysis in this study were inconclusive.

For LC-MS/MS analysis, it is important to note that not all fragment ions produced during MS can be present at a level which is detectable; hence several peptides from the target protein should allow us to identify the target protein with relative certainty. It was suggested that a stretch of seven amino acids is often unique to a specific protein in the human genome (Rosenberg, 2004).

4.3.1 Analysis of Tag Cleavage from f-FXII1-71 and f-FXII23-71 Identification of His-Trx Tag

In this study, LC-MS/MS analysis of the in-gel digestion products of band 1 and 3 (Figure 4.3) identified at least four ion masses that matched the sequence of E.coli Trx, with each of the fragments having more than seven amino acids. In addition, none of the peptide fragments were matched to FXII protein of any species. These confirm that the Trx-tag has been successfully cleaved using 3C protease.

Figure 4.3: The 18% SDS-PAGE gel (from Figure 3.4B) used to analyse the proteolytic tag removal by 3C protease. Analysis of His-Trx tag Cleavage from f-FXII1-71

De novo peptide sequencing identified three intact fragment ion masses from band 4 that matched that of FXII with each fragment consisting of a stretch of at least 13 amino acids. Furthermore, 65 of the 71 amino acids of FXII1-71 were identified which suggest that the whole FXII1-71 protein is intact in band 4. Based on these results, we can confidently conclude that band 4 corresponds to FXII1-71 and confirm that the active site of 3C protease is located between His-Trx tag and FXII1-71 as demonstrated by Walker et al (1994).

Yet, two peptide ion masses that were identified to Trx were also found in in-gel digestion products of band 4. It is unlikely that this is a result of incomplete tag cleavage from f-FXII1-71 because f-FXII1-71 has a higher MW (~22kDa) and thus should be located higher up in the SDS-PAGE gel. These two fragment ion masses generated in band 4 were identical to those found in band 3. Note that the in-gel digestion products of band 4 were analysed after band 3. It was also found that the peptide ion masses of band 3 produced very high signals which also did appear during the blank run between runs of band 3 and 4 during LC-MS/MS. Thus, there is a possibility that the Trx signals identified in band 4 were contaminations carried-over from the run of band 3. On top of that, the presence of Trx fragments in band 4 could be due to the poorly resolved band 3 and 4 on the 18% SDS-PAGE gel. As seen in Figure 4.3, band 3 and band 4 were located very near to each other and there is not a distinct border between the two bands. Thus, there is a possibility that when the bands were incised from the gel, a small portion of band 3 remained intact in band 4. To further elucidate this, we can repeat the proteolysis trial and use other novel 2D gel techniques which have better separation resolution. This will be discussed in more detail in section 4.3.2. Analysis of His-Trx Tag Cleavage from f-FXII23-71

De novo peptide sequencing identified two intact peptide fragments in band 2 that correspond to FXII which consist of more than seven consecutive amino acids and covered 34 of the 49 amino acids of FXII23-71. Unlike band 4, Trx was not detected in band 2 which suggests that all the fusion protein f-FXII23-71 have been cleaved leaving just the FXII23-71. This further suggests that the traces of Trx detected in band 4 were due to contaminations rather than uncleaved fusion protein. At this point of the discussion, it can be confirmed that 3C protease can be used to remove the His-Trx tag from f-FXII1-71 and f-FXII23-71. Stability of FXII-RPs after tag removal

In the 18% SDS-PAGE gel used for analysis of tag cleavage (Figure 4.3), the bands that correspond to FXII23-71 and FXII1-71 (band 2 and 4) have lower intensities than the bands that correspond to His-Trx tag (band 1 and 3). The principle of gel staining using coomassie dye is that the dye will bind to groups of amino acids. As the FXII-RPs have lower MWs than the His-Trx tag, naturally less dye will bind to them as compared to His-Trx tag. Thus, it is not unusual for the FXII-RPs to be stained less intensely than the His-Trx tag.

However, this difference in intensity may be contributed by another factor. In this study, although we have proven that the Trx has aided disulphide bond formation, we do not know whether the disulphide bonds formed were permanent, i.e. will the FXII recombinant proteins still remain soluble upon removal of the His-Trx tag. It is important to note that there were cases where once the fusion partner was removed, the fused recombinant protein does not remain soluble (Waugh, 2005; Esposito and Chatterjee, 2006). A hypothesised explanation for this observation is that in the presence of the solubility-enhancing tag, the seemingly soluble fused proteins are held in solution as 'soluble aggregates' by interactions with the fusion partner (Esposito and Chatterjee, 2006). Thus, once the solubility-enhancing tag is removed, the fused proteins return to their aggregated state.

If the FXII-RPs did behave this way, they will precipitate out of the solution after tag cleavage resulting in their lower concentration in the solution as compared to the His-Trx tag; consequently, when analysed on SDS-PAGE gel, they will be less intensely stained. As the volume of reaction mixture used in this study is very small, it is not possible to notice any precipitates if any were formed. To verify this, we can scale up this trial using more FXII-RFPs and protease to give a larger reaction volume so that it can be observed whether any precipitates are being formed. In addition, after separation of the His-Trx tag and FXII fusion proteins via GFC, these proteins can be analysed quantitatively to determine their concentration and subsequently compare their relative yield.

4.3.2 Analysis of Tag Cleavage from f-FXII23-112

On the 18% SDS-PAGE gel (Figure 4.4), we can see that there is only one visible band in the lanes (3, 7, 10) which contain cleavage products of f-FXII23-112. These bands are on the same horizontal line as the His-Trx tags whose identity have been proved in MS analysis as mentioned above. This proved that the His-Trx tag was cleaved from f-FXII23-112 but it is rather odd that there are no other bands seen in these lanes because His-Trx tag cleavage from f-FXII23-112 should yield both His-Trx and FXII23-112. An explanation for this observation is that the FXII23-112 was retained in the same band as His-Trx in the gel. This suggestion is based on the poor separation of the bands seen in lane 8. The proteins in these bands have been identified as His-Trx (~13.8kDa) and FXII1-71 (~8.45kDa) respectively via MS. In lane 8, we can see that the band which corresponds to FXII1-71 is located just below the His-Trx band. If a protein of ~8.45kDa MW is so poorly resolved from a protein of ~13.8kDa MW in the gel, it is unlikely that FXII23-112, which has a MW of ~10.27kDa, will form a distinct band on the gel.

Figure 4.4: The 18% SDS-PAGE gel (from Figure 3.4B) used for analysis of His-Trx tag cleavage. Lane 3, 7 and 10 contains proteolytic products of 4hr, 6hr and overnight incubation of f-FXII23-112 with 3C protease.

To elucidate this, further experiments can be done using separation techniques with better resolving power in order to separate the FXII23-112 and His-Trx tag distinctly. For better separation, a 20% SDS-PAGE gel can be used. Alternatively, we can use the Tricine gel system which a has high resolving power in the 5-20kDa range (Schägger and von Jagow, 1987).


4.4.1 Identification of Protein Crystals

Precipitates were observed in the crystallisation condition 0.1 M sodium acetate pH 4.6 and 1.0 M sodium chloride for f-FXII1-71 and their birefringence properties confirmed that these are crystals. It is important to note that besides protein crystals, many other organic and inorganic materials that may be present within crystallisation screen are birefringent as well, salt crystals being the most common one. Thus, it is important to identify protein crystals from others. There are several ways to differentiate a protein crystal from a salt one. Typically, the birefringence properties of salt crystals are more pronounced than that of protein crystals (Bergfors, 1999). Besides this, salt crystals tend to have sharp edges and are larger in size. They are also able to grow within a short period of time. The crystals as seen in Figure 3.8B were only slightly birefringent and relatively small. Furthermore, they started to grow only after three weeks upon setting up the screen. All these strongly suggest that these are proteins crystals rather than salt. Although there are various other tests which can more definitely prove that these are protein crystals, these tests are undesirable as they involve disruption of the crystals.

4.4.2 Has Trx Aided Crystallisation?

The reservoir solution in which crystals of f-FXII1-71 were formed has a pH of 4.6. Ideally, the pH of the reservoir solution should be as close to the protein's isolectric point (IEP) as possible because protein carries a net charge of zero at its IEP. However, the estimated IEP of f-FXII1-71 is 6.09. Interestingly, the estimated IEP of the thioredoxin sequence has an IEP of 4.8 which is fairly similar to the pH of the reservoir. The IEP of FXII1-71 on the other hand was found to be 7.06. There are several reports where Trx-tag was demonstrated to aid crystallisation of its fusion partner either by acting as a carrier or a stabilising agent. In the former, the Trx-tag remained fused to the partner protein (Stoll et al, 1998; Corsini et al, 2008) whereas in the latter the tag was cleaved but not removed (Cura et al, 2007). Thus, there is a possibility that in this study, the crystallisation condition is suitable for the crystallisation of Trx and that its crystallisation has aided the FXII-RPs to crystallise as well.

4.4.3 Optimisation of Crystallisation Screen

In Figure 3.8, we can see that overlapping needle-shaped micro-crystals were formed from two nucleation sites. These crystals are not ideal for analysis as they are too small to be characterised by diffraction (Bodenstaff et al, 2002). Also, individual crystals are required for X-ray diffraction analysis (Judge et al, 2005).

The crystallisation process can be divided into two phases: nucleation and crystal growth. Ironically, the conditions required for these two phases are often different. For nucleation, a supersaturated solution is essential whereas growth tend to occur in a solution where the concentration is just slightly below that of the supersaturated. In the crystallisation phase diagram, this is known as the metabstable zone which lies between supersaturation and solubility. The metastable zone supports crystal growth but no further nucleation will occur. Thus, to obtain good quality well-ordered crystals, it is often necessary to separate nucleation from growth. One of the techniques which can be used to achieve this is seeding where crystals nuclei formed are introduced directly into a metastable protein solution (Chayen, 2005; Benvenuti and Mangani, 2007).

Crystallisation is usually the rate-limiting in protein structure determination via X-ray crystallography as there are a large number of variable factors that can affect crystals growth (Benvenuti and Mangani, 2007). Important parameters that may influence crystal formation include temperature, protein concentration, pH, buffer type, ionic strength. To increase success rate, one must constantly evaluate the feedback obtained from each screen (Bodenstaff et al, 2003; Benvenuti and Mangani, 2007). Now that we know the conditions that support nucleation of f-FXII1-71, this is a good starting point for subsequent optimisation screening.


4.5.1 The Rationale of Functional Annotations Based on Sequence Homology

Through evolution, functionally and structurally important residues tend to be more conserved as compared to other positions in the sequence, thus sequence homology has been suggested to be a useful tool in functional assignment of proteins (Goldsmith-Fischman and Honig, 2003; Petsko and Ringe, 2004). A 40% rule (Figure 4.5) has been put forward based on the analysis of the biochemical function of proteins with sequence similarity; if a protein shares more than 40% sequence identity with a target protein whose biochemical function is known, it is reasonable to assume that the two proteins may have a common biochemical function with the condition that the functionally important residues are conserved between the two sequences (Petsko and Ringe, 2004).

Figure 4.5: The relationship between sequence similarity and functional similarity. When the sequence identity of two proteins is above 40%, it is very likely that these two proteins have similar biochemical function. This graph is obtained by Mark Gerstein by plotting the percentage of proteins pair with the exact biochemical function against the percentage of sequence identity (Petsko and Ringe, 2004). (Adapted from

4.5.2 Multiple Sequence Alignment of FNII domain

Sequence alignment of FXII FNII sequence to that of fibronectin and MMP-9 both show sequence similarity of more than 40% (Figure 4.6), thus fibronectin and MMP-9 are suitable targets for predicting the function of FXII FNII domain. The FNII domains of both fibronectin and MMP-9 have been implicated in binding to collagen (Pickford et al, 1997; Visse and Nagase, 2003).

In a study of the solution structure of the first type II module of fibronectin, it was demonstrated that the type II module is involved in binding to collagen and that hydrophobic interaction is important. It was proposed that in this FNII domain, the large solvent-accessible hydrophobic pocket formed by the residues, Tyr21, Trp40, Tyr47, Tyr53 and Phe55, is responsible for collagen binding (Pickford et al, 1997). Multiple sequences alignment (Figure 4.6) shows that the positions in the FNII sequence of FXII and MMP-9 that correspond to the positions of Tyr21, Trp40, Tyr47, Tyr53 and Phe55 in fibronectin consist of hydrophobic aromatic residues tyrosine, tryptophan or phenylalanine.

Using the homology model of FXII FNII domain, it was found that these conserved hydrophobic positions do form a solvent accessible pocket as well (Figure 4.7A). Furthermore, the surface around this pocket has got a net electropositive charge as shown in Figure 4.7B. Also, the side chain of arginine which is positively-charged is located near to the hydrophobic pocket and this arginine residue is conserved in other FNII domains as well.

FN 1st nlngepcvlpftyngrtfyscttegrqdghlwcsttsnyeqdqkysfc



Figure 4.6: The multiple sequences alignment of the first FNII domain of fibronectin (FN 1st), FNII domain of FXII and the first FNII domain of MMP-9. The five residues highlighted (cyan) in the FNII domain of fibronectin (Tyr21, Trp40, Tyr47, Tyr53 and Phe55) form a hydrophobic pocket which is indicated in collagen binding. Hydrophobic residues(red) were found to be conserved in the positions of FXII and MMP-9 sequences that correspond to Tyr21, Trp40, Tyr47, Tyr53 and Phe55. The conserved arginine residues that are located near to the hydrophobic residues are highlighted (green).



Figure 4.7: A ribbon representation of the FNII domain of FXII. The residues in red are the conserved hydrophobic residues. The side-chain of arginine (purple) which is positively-charged is located near to the pocket. (A) The solvent accessible surface is generated and the hydrophobic pocket is found to be accessible to solvent; (B) The electrostatic surface is generated and the hydrophobic pocket is found to be located near to a net electropositive surface.

As mentioned before, the sequence consisting of amino acids 39-47, which lies within the FNII domain of FXII, is involved in binding to negatively-charged surfaces (Citarella et al, 2000). Also, it was demonstrated in various literatures that FXII activation is initiated upon binding to negatively-charged surfaces (Cochrane et al, 1973; Griffin, 1978; Samuel et al, 1992). The electropositive surface and the positively-charged side-chain fulfil the characteristics of a binding site for negatively-charged surfaces. As collagen (negatively-charged) is also known to activate FXII, based on the high sequence identity of the FNII domain of FXII to that of fibronectin and MMP-9, there is a possibility that the conserved hydrophobic pocket seen in Figure 4.7 could be the putative binding site for negatively-charged surfaces.

Yet, this is just an assumption as there are many unknown variables. For example, we do not know if this hydrophobic pocket is accessible within the whole FXII structure as the pocket may be sterically hindered by other domains; the conformation or the exposed residues of the FNII domain could be different in reality when this domain exists as part of the FXII chain. In addition, the precise mechanism of FXII activation is not known yet.


In conclusion, we have successfully expressed the FXII-RFPs, f-FXII1-71, f-FXII23-71, f-FXII23-112 and f-FXII23-154, in BL21 (DE3) cells with His-Trx tag at the N-terminus. These were found to be soluble. Several analyses have proved that f-FXII1-71 and f-FXII23-71 have adopted the correct globular conformation which suggests that the Trx fusion tag has aided disulphide bonds formation. We believe that this would also be the case for f-FXII23-112 and f-FXII23-154 if time has allowed for their further characterisations. Step-wise purifications of the FXII-RFPs via IMAC and GFC resulted in highly homogeneous sample. A crystallisation screen for f-FXII1-71 has produced crystals which we believe are protein crystals. The pH condition of the reservoir was found to be very close to the IEP of Trx suggesting the possibility that the Trx might have an assisting role in the crystallisation of f-FXII1-71. A trial to cleave the His-Trx tag in f-FXII1-71 and f-FXII23-71 using 3C protease was successful but further investigations are required to determine the stability of the FXII-RPs after tag cleavage. Homology model of the FNII domain of FXII shows the existence of a conserved hydrophobic pocket that is solvent-accessible; this hydrophobic pocket may be functionally significant.


Due to time constraint, there were many things that we did not manage to validate in this study. Now that we know the E.coli expression system used is suitable for the expression of FXII-RFPs, the next step would be to scale up the production of these fusion proteins. Further characterisations of these fusion proteins via IMAC/GFC may reveal more information on their tertiary structure. In order to elucidate the role that Trx has in the solubility of the fusion proteins, the cleavage of His-Trx tag should be repeated at a larger scale to determine if the FXII proteins still remain soluble in the absence of Trx. Also, more crystallisation screens of the proteins with/without tag can be set up to obtain a good quality crystal for X-ray diffraction analysis. As this study is part of an on-going project, large-scale production of f-FXII1-71 is going on as the crystallisation screen in this study demonstrated that f-FXII1-71 has successfully yielded protein crystals. If good quality crystals can be obtained, diffraction analysis will reveal the tertiary structure the FNII domain.