Structural and Functional Studies of the Fusidic Acid Resistance Protein FusB
Xiaohu Guo
Degree project inapplied biotechnology, Master ofScience (2years), 2010 Examensarbete itillämpad bioteknik 45 hp tillmasterexamen, 2010
Biology Education Centre and Structural biology, Dept. ofCellular and Molecular Biology, Uppsala University
Abstract ...2
Introduction ...3
Fusidic acid ...3
Staphylococcus aureus, a major antibiotic resistant pathogen...3
Staphylococcus aureus antibiotic resistance mechanism...3
The identification of FusB, another fusidic acid resistant determinant ...4
FusB and its homologues...4
Aim of the study and result...5
Material and Method ...5
Plasmid pUB101 extraction...5
TA cloning ...5
Cleavable His-tagged FusB (FusB_LN) TA cloning...5
Non-cleavable His-tagged FusB (FusB_SN) TA cloning...6
Non His-tagged EF-G (EF-G_NoT) TA cloning...6
Transformation and large-scale Expression ...6
His-tagged FusB transformation and large scale expression...6
None His-tagged EF-G (EF-G_NoT) transformation and large-scale expression...7
His-tagged EF-G transformation and large-scale expression...7
Protein purification ...7
His-tagged FusB purification...7
His-tagged EF-G purification...7
TEV protease purification...8
FusB – EF-G complex purification...8
Digested S.a EF-G and FusB binding test...8
Crystallization of FusB and complex ...9
Data collection...9
Results...9
FusB and EF-G plasmid construction ...9
Tev protease His-tag cutting test ...10
FusB binds to S. aureus EF-G ...11
FusB does not bind to E. coli EF-G...11
Complex purification ...12
FusB and S. aureus EF-G affinity test by fluorescence spectroscopy ...13
Digested S. aureus EF-G and FusB binding test...13
Crystallization and crystallography...15
Discussion ...18
Structure determination of FusB and FusB - EF-G complex ...18
The binding target of FusB...19
Hypothetical resistance mechanisms ...21
Acknowledgement ...23
Reference...24
Abstract
Fusidic acid, first derived from the Fusidium coccineum, is an antibiotic used against Staphylococcus aureus. It functions by blocking the release of elongation factor G (EF-G) from the ribosome, thus preventing the binding of a new aminoacyl tRNA to the ribosome and blocking the translation process. One resistance mechanism for S. aureus to fusidic acid involves a single gene fusB, carried by plasmid pUB101. The FusB protein has previously been shown to interact with S. aureus EF-G but not with E. coli EF-G. Further, FusB confers fusidic-acid resistance to an S. aureus in vitro translation system, but fails from protecting an E. coli in vitro translation system from fusidic-acid inhibition.
With the aim of structure determination and biochemical studies of FusB and the FusB-EF-G complex, we have cloned, expressed and purified FusB and S. aureus EF-G. FusB has been crystallized. The crystals diffract to 3.9 Å resolution and belong to space group P21212. Crystals are being further optimized. We also study the in vitro interaction of S. aureus EF-G and FusB.
Our results show that the FusB can not bind to S. aureus EF-G domainⅠtogether with domain Ⅱ, and probably functions by its interaction with domainⅤ.
Introduction
Fusidic acid
Fusidic acid, first derived from the fungus Fusidium coccineum, is an antibiotic used to against gram-positive bacteria such as Staphylococcus and Corynebaterium [1]. Shortly after this finding in the 1960s, sodium salt of fusidic acid was induced into clinical use, especially against Staphylococcus aureus.
Fusidic acid can inhibit protein synthesis by preventing the turnover of elongation factor G (EF-G) [2]. After EF-G has hydrolyzed GTP to GDP and completed the translocation step; it undergoes a large conformational change and falls off from the ribosome, emptying the position for the aa-tRNA. The structure of EF-G band to the ribosome reveals that fusidic acid is located in a pocket surrounded by domainⅠ Ⅱ and domainⅢ of EF-G and there is no direct , interaction with ribosome. This binding maintains the switchⅡconformation of the G domain in GTP binding form, thus preventing the release of EF-G from the ribosome [3].
Staphylococcus aureus , a major antibiotic resistant pathogen
S. aureus is a Gram-positive coccus responsible for many infections in humans and animals through its toxin production or invasion. It’s well known for its super adaptability to antibiotics. In 1947, only after 4 years of penicillin broad usage, the resistance at this strain was reported. Subsequently resistance appeared to other drugs like methicillin, tetracycyline, fusidic acid, erythromycin and vancomycin. In the 1990s, people found Linezolid, a new class of antibiotic and more efficient than vancomycin against S. aureus.
However, merely 1 year after it was approve for use in 2000, a linezolid resistant clinical isolate was found in Israel [4]. That is why understanding the resistance mechanisms are important and that is also the reason that some antibiotics like fusidic acid has been valued again for multiple drug treatment.
Due to the high level development of the resistance, fusidic acid is never recommended as infection treatment on its own [5].
Staphylococcus aureus antibiotic resistance mechanism
The resistance mechanisms have been intensively studied in S. aureus. To date, there are four well known routes for different bacteria to gain resistance against various antibiotics. The first one, drug modification, works by inactiving the antibiotics. This mechanism can for example give resistance against Macrolide, and Lincosamide etc [6]. A second mechanism involves an alteration of metabolic pathways, which means that once a vital metabolic pathway is damaged by antibiotics, some bacteria can somehow compensate this damage
by changing its metabolic pathway to another one. Some sulfonamide-resistant bacteria belong to this kind [7]. The third strategy is to reduce drug accumulation, for example, tetracycline can be pumped out of the cell by a special membrane protein in some of the resistant-strains [8]. The fourth one, which is also the major mechanism used by S. aureus against fusidic acid, works by changing the target site lowering its affinity to the drug [9]. In this case, the changed target is S. aureus EF-G. Several mutations in fusA, the gene coding for EF-G, have been identified both from in vitro generated and clinically isolated fusidic acid resistant strains [10]. Further studies shows three positions of these mutations, V90I, L461K and H457Y/H457Q, could result high-level resistance to fusidic acid [11]. Since these residues are located in domainⅠand
Ⅲ of EF-G, where fusidic acid binds, it strongly indicates that the mutations change the binding affinity of fusidic acid on EF-G.
The identification of FusB, another fusidic acid resistant determinant
Studies in 1970s identified another fusidic acid resistance determinant is carried by a 22kb plasmid pUB101, which is also the host for penicillinase and cadmium resistance genes. In 2000, a gene named fusb was for the first time recognized causing the resistance [12]. Interestingly, fusb-caused resistance does not involve to enzymatic inactivation or modification [9]. Sequence analysis showed that the protein doesn’t have any membrane location signals or secretion signals, indicating that it is unlikely to work by reducing drug accumulation. So, since more than 30 years, the resistance mechanism is still unclear. However, in 2006, the fusidic acid resistance protein FusB was identified and successfully purified. Importantly, FusB was found to interact specially with S. aureus EF-G in vitro [13]. These results suggest that FusB works by preventing fusidic acid binding to EF-G or by somehow helping to unlock the EF-G – fusidic acid ribosome complex. Unraveling the mechanism behind FusB resistance will undoubtably enrich our understanding of antibiotics resistance.
FusB and its homologues
FusB is a small protein with a molecular weight of 25kD. Its homologues could be found in some Gram-positive genera, for example, Enterococci, Lactococci and Lactobacilli. The majority of these species were found to be inherently insusceptible to fusidic acid [14] [15] [16]. Only one of these homologues, located in the chromosome of Listeria monocytogenes, has been studied previously.
However, it was identified as a fibronectin-binding protein (Fbp) and its function is to bind to a eukaryotic cell’s fibronectin facilitating infection of bacteria [17]. Although Fusb and Fbp have a high sequence identity (43%), they bind to different target and engage in very different missions. It is noticeable
that Listeria monocytogenes is one of the strains inherently insusceptible to fusidic acid and there is no study showing whether Fbp can only interact with fibronectin or also cause fusidic acid resistance.
Aim of the study and result
With the aim of structure determination and biochemical studies of FusB and the FusB-EF-G complex, we have successfully cloned FusB and S. aureus EF-G into different constructs. FusB, EF-G and the FusB - EF-G complex were purified by immobilized metal ion affinity chromatography (IMAC) and size exclusion chromatography. The purified proteins were subjected to screens for crystallization. FusB with a short polyhistidine-tag (His-tag) has been crystallized. The crystals diffract to 3.9 Å resolution and belong to space group P21212 with cell dimension 129, 187, 93; 90, 90, 90. The crystals are being further optimized. We also study the in vitro interaction between FusB and EF-G.
Due to the inner filter effect of FusB, binding affinity could only been poorly determine to 1μM using the fluorescence spectroscopy method. Surprisingly, further study of the biniding shows that FusB can not bind to the digested fragments of EF-G domainⅠtogether with domain Ⅱ. Together with the sequence alignment analysis and structure analysis, domain Ⅴ is believed the key domain interacting with FusB.
Material and Method
Plasmid pUB101 extraction
A Single S. aureus colony carrying the FusB resistance determinant was picked and grown in 15ml LB medium containing 0.5% glycine at 37℃ overnight without shaking. The cells were spun down at 4000rpm for 15mins and re-suspended in lysis buffer (100ul TE pH7.0, 100μl 300μg/ml lysostaphin, Sigma; 100μl 100mg/ml lysozyme, Sigma; 45μl 20mg/ml proteinase K, Sigma) incubating at 37℃ 60mins. Lysis buffer from QIAprep® Spin Miniprep Kit (Qiagen) was added into the cell lyse and then followed the kit protocol to extract the plasmid.
TA cloning
Cleavable His-tagged FusB (FusB_LN) TA cloning
Primers for cloning were designed to be 24bp (fusB_f1) for the forward direction, and 28bp (fusB_b1) for reverse direction (Table1). Because of the low GC content of the gene (25%), a silent mutation A6G was directed by the forward primer to increase the GC content in order to prevent any formation of
secondary structure. PfuUltraTM High-Fidelity DNA Polymerase (Stratagene) was used for amplification. The PCR reaction was programmed as following;
double-stranded DNA was denatured at 95℃, for 5mins, followed by 35cycles of amplification. Each cycle contains, 1min denaturing at 95℃, 1min annealing at 56.8℃, 1min extension at 72℃. The program was ended with extra 10min at 72℃.The amplification product was detected on 1% agarose gel and purified by QIAquide® Gel Extraction Kit (Qiagen). A 3’ adenine overhang was added by Taq polymerase (Invitrogen) at 72℃ for 10mins. Ligation of the gene to the plasmid was conducted by pEXP5-NT/TOPO TA Expression Kit (Invitrogen). The plasmids carrying the gene were transformed into TOP10 competent cells following kit protocol.
Non-cleavable His-tagged FusB (FusB_SN) TA cloning
The same protocol was followed as above. The Amplification was using forward primer fusB_f2 and backward primer fusB_b1 (Table1). The amplified gene was cloned into the pEXP5-CT (Invitrogen) vector.
Non His-tagged EF-G (EF-G_NoT) TA cloning
Plasmid pET-30 was constructed carrying S. aureus EF-G from collaborator Suparna Chandra Sanyal. The template was then used as the template for further cloning. The same protocol was used as above except, forward primer EF-G_f1 and backward primer EF-G_b1 (Table1). The annealing temperature was optimized to be 59℃.
Table 1. Primers
Name Sequence
fusB_f1 ATGAAGACAATGATTTATCCTCAC
fusB_f2 ATGGCTCATCATCATCATCATCATGGTATGAAGACAATGATTTATCCTCAC
fusB_b1 CACAAACATAGTTAATTCCTTAATCTAG EF-G_f1 ATGGCTAGAGAA TTTTCATTAGAAAAAACT EF-G_b1 AAGCCCGGTTATTCACCTTTATTTTTC
Transformation and large-scale Expression
His-tagged FusB transformation and large scale expression
100μl BL21(DE3) competent cells were taken directly from a -70℃ refrigerator and thawed on ice, 10μl 20ng/μl plasmid was added. After incubation on ice for 10mins, cells were heat shocked for 90 seconds at 42℃ and immediately placed back on ice for additional 2mins. 150μl of pre-warmed SOC medium (Invitrogen) was added and incubated for 1 hour at 37℃ with shaking at 225rpm. 200μl medium containing transformed cells was plated on a LA plate with 100mg/ml ampicillin (Sigma) and incubated overnight at 37℃. A single colony was picked and inoculated to 10ml 2×TY medium (1.6% tryptone, 1%
yeast extract, 0.5% NaCl) containing 50mg/ml ampicillin at 37℃ overnight. The small culture was then inoculated into 1L medium and incubated at 37℃. IPTG (Sigma) was added to a final concentration of 1mg/ml when the OD600 of the culture reached 0.6 and the culture was then continued at 30℃ for overnight.
The culture was shaked at 100rpm. Cells were harvested by centrifugation at 4000rpm for 30mins and stored at -20℃.
None His-tagged EF-G (EF-G_NoT) transformation and large-scale expression The same protocol was used as above except the following: competent cells were BL21-AITM One Shot® Chemically Competent cells (Invitrogen); when OD600 of the 1L cell culture reached 0.5, the culture was immediately moved to 4℃ for 30mins with a shaker speed of 70rpm. The culture was induced by IPTG (Sigma) and L-arabinose (Sigma) at final concentration of 1mg/ml and 0.2%
respectively and incubated at 16℃ overnight.
His-tagged EF-G transformation and large-scale expression
Same protocol used as for EF-G_NoT purification except that kanamycine (Sigma) was used for LA plates and 2×TY medium was used at a concentration of 50mg/ml.
Protein purification
His-tagged FusB purificationCell pellet was re-suspend in 10ml IMAC A buffer (Table2) with 1/2 pill Complete Protease Inhibitor Cocktail Tablets (Roche). Cells were lysed by sonication and debris was spun down at 18,000rpm for 30mins. Supernatant was transferred into a filter column together with 2ml Ni SepharoseTM (GE healthcare) and equilibrated for at 4℃ for 1 hour. The matrix was washed with 30ml IMAC A buffer, followed by 150ml IMAC A (high salt) buffer. His-tagged FusB was then eluted with 16ml IMAC B buffer. Eluted fractions were poured together and concentrated to 5ml in a VIVASPIN 6 (Sartorius Stedim) with membrane cut-off 10,000 at 4000rpm. Concentrated fractions were further purified by size exclusion column HiLoad 16/60 Superdex75 TM (GE healthcare) with gel filtration buffer (Table2) using AKTA Purifier (GE healthcare). Protein concentration was measured spectrophotometrically using extinction coefficient ε280 = 26610 by NanoDrop ®, and concentrated to 12mg/ml.
His-tagged EF-G purification
The same procedure as with His-tagged FusB was uesd except; IMCA A/B buffer (Table2), extinction coefficient ε280 = 52510; and size exclusion column was HiLoad 16/60 Superdex200 TM (GE healthcare).
TEV protease purification
His-tagged TEV protease was purified using Ni SepharoseTM using a previously published protocol [18].
Table2 Protein purification buffer
Buffer FusB EF-G
IMAC A 50mM Na-phosphate, 300mM NaCl, 20mM Imidazole, 5mM BME pH7.0
50mM Tris-HCl, 200mM NaCl, 20mM Imidazole, 5mM BME, pH7.5
IMAC A (high salt)
50mM Na-phosphate, 600mM NaCl, 20mM Imidazole, 5mM BME, pH7.0
50mM Tris-HCl, 600mM NaCl, 20mM Imidazole, 5mM BME, pH7.5
IMAC B 50mM Na-phosphate, 300mM NaCl, 500mM Imidazole, 5mM BME, pH7.0
50mM Na-phosphate, 200mM NaCl, 500mM Imidazole, 5mM BME, pH7.5
Gel filtration buffer
20mM Tris-HCl, 300mM NaCl, 5mM BME, pH8.3
20mM Tris-HCl, 200mM NaCl, 5mM BME, pH7.5
FusB – EF-G complex purification
The same protocol was used as for FusB purification until the Ni SepharoseTM was equilibrated with the supernatant of FusB cell lyses. The Sepharose was washed by 20ml FusB IMAC A buffer, and then equilibrated with EF-G cell lyses supernatant (The debris was spun down at 18,000rpm 30mins) for 1 hour. Then followed the same procedure for FusB purification to wash, elute and further purify by size exclusion column HiLoad 16/60 Superdex200 TM.
Digested S.a EF-G and FusB binding test
100μl 10μg/ml trypsin (Jena Bioscience) was added into 200μl 12mg/ml S.a EF-G and incubated at room temperature for 1 hour. Meanwhile, 100μl 16mg/ml FusB was equilibrated with 100μl Ni Sepharose in a filter tube at room temperature. The digestion reaction was quenched by adding 1/20 pill Complete Protease Inhibitor Cocktail Tablets (Roche). Digested EF-G was poured into the filter tube and incubated together with FusB and Ni Sepharose for 5min. Flow though was collected after centrifugation 4000rpm for 1min. The digested fragments that couldn’t bind to nickel or FusB were further washed with 500μl FusB IMAC A buffer for 6 times. Everything binding to Ni Sepharose was eluted by 200μl FusB IMAC B buffer at 4000rpm for 1 min. Imidazole in the elution was diluted by a concentrator with FusB gel filtration buffer. Samples in different stages were checked by both SDS and native PAGE.
Crystallization of FusB and complex
FusB was screened by Easy Xtal JCSG+ (Molecular Dimensions), MorpheusTM (Molecular Dimensions) and Structures Screen (Molecular Dimensions). Drops mixed with 0.3μl FusB (12mg/ml) and 0.3μl reservoir buffer (80μl) were prepared as a sitting-drop vapor diffusion experiment using a crystallization robot and incubated both at 277K and 293K. Hits from the Morpheus screen were repeated manually with 2μl hanging drops (50% FusB) against 400μl reservoir buffer. Morpheus screen C8, the only successfully repeated condition was optimized by grid-screening with different combination of pH (from 7.1 to 7.7) and precipitant concentration (from 32% to 39%). Further optimization was initially screened by ADDitTM Additive screen (Emerald BioSystems) in sitting drops each contains 0.3μl 12mg/ml FusB, 0.3μl buffer (Morpheus C8, pH7.1, 32% precipitant) and 0.1μl Additive. Some promising conditions with less precipitate were re-checked by 2μl hanging drops with additive concentration 2.5×10-4M, 5×10-3M and 2×10-3M.
Easy Xtal JCSG+, MorpheusTM, Structures screen and Protein Complex Suite (Qiagen) were used for complex screening by sitting drops 0.3μl 15mg/ml complex mixed with 0.3μl reservoir buffer (80μl).
Data collection
Before vitrified by liquid nitrogen, FusB crystals were soaked with different concentrations of glycerol. Some of them were dehydrated by changing the reservoir buffer with 50% precipitant and incubate for two days. Crystals were checked at ID23-1 station of European Synchrotron Facility in Grenoble, France.
Two data sets were collected with resolution beyond 5 Å. One of them was collected by using 10% intensity of the beam. 725 images were collected with oscillation range 0.4 degree, each image 0.79s. Another one was using the same beam intensity but collected 340 images with oscillation range 0.25 degree, each image 0.581s.
Results
FusB and EF-G plasmid construction
The fusB gene (661bp) was successfully amplified (Figure.1) and cloned into the pEXP-NT vector. fusB gene with added His-tag sequence (688bp) and fusA which encodes EF-G, (2090bp) were cloned into the pEXP-CT vectors. Plasmids with insert of correct site were then sent for sequencing. All plasmids were confirmed to have the right sequence except EF-G which contained a point mutation E242G. This mutation was shown to come from the template and
since it does not affect the activity, we decided to keep going with it.
Figure.1. Amplified PCR products used for TA cloning run on 1% agarose gel. 200bp ladder (O’ranger) was used for reference.
Tev protease His-tag cutting test
In order to increase the possibility of crystallization, the presumably unstructured His-tag was cut off after protein purification. We purified TEV protease and FusB_LN separately. To test the home made TEV protease activity, a series of FusB mixed with different concentration of TEV were incubated for different time at 4℃ and 25℃. The digestion was then checked by SDS-PAGE.
(Figure.2). From the SDS-PAGE we can see that after digestion of His-tagged FusB (28.8kD) by TEV protease, a smaller band corresponding to FusB without tag (26kD) appears. Later large scale purification shows that after TEV digest, FusB could not bind to the nickel column, indicating that the His-tag has been cut off. The experiment shows that even at protein concentration 1:200 (TEV:
FusB), most His-tag has been successfully chopped off at 25℃ overnight. Later experiment, the radio 1:10 mg/ml (TEV:FusB) was used to make sure the reaction went to completeness.
Figure.2. TEV protease cutting test. 1mg/ml FusB was incubated overnight with different concentration of TEV protease at both 4℃ and 25℃ and then examined by SDS-PAGE (20%
PhastGelTM Homogeneous, GE Healthcare) using PhastSystemTM (Pharmacia). The “n/a” mean no TEV protease was added.
FusB binds to S. aureus EF-G
To test if FusB could bind to S.a EF-G, purified FusB and S.a EF-G were mixed at different molar ratios in FusB gel filtration buffer (20mM Tris-HCl, 300mM NaCl, 5mM 2-Mercaptoethanol) and incubated at room temperature for 10mins. The complexes were examined by native-PAGE. (Figure.3 B) From the SDS-PAGE (Figure.3 A), we can see that the molar ratio of these two protein roughly agree with the calculation. Since FusB has a very high pI (8.9, calculated by ExPASy Proteomics Server), it did not run into the native-PAGE. S.a EF-G alone shows two bands with similar contribution. This indicates S.a EF-G could have two different conformations in the native-PAGE buffer. After mixing with FusB, the S.a EF-G bands disappear. Instead, a single band at a higher position shows up.
Mass spectrometry analysis later showed that it contained both EF-G and FusB.
With molar ratio (FusB:EF-G) 1:2, excess EF-G was found together with the complex band, suggesting that the complex is formed by these two proteins in a 1:1 molar ratio.
Figure.3. Binding test of FusB and S.a EF-G. the two proteins were mixed at different molar ratio (FusB:EF-G; 1:1, 1:2, 2:1) and then run on both (A) 20% Homogeneous, SDS-PAGE (GE healthcare) and (B) 8-25% gradient native-PAGE (GE healthcare) by PhastSystemTM (Pharmacia).
Low molecular maker is from Amersham Bioscieces.
FusB does not bind to E. coli EF-G
To check if FusB and E. coli EF-G could also form a complex, FusB was mixed with E. coli EF-G and checked by native-PAGE. From Figure.4 we can see that FusB and S. aureus EF-G could form a complex. E. coli EF-G alone runs as a single bind slightly higher than S. aureus EF-G. Since both proteins have similar
molecular weight and shape, this is probably because E.coli EF-G has higher PI (5.24) compared to S. aureus EF-G (4.93). No change is observed when FusB is added, indicating that FusB does not bind to E. coli EF-G. This means the FusB interact specifically with S. aureus EF-G. Because E. coli EF-G and S. aureus EF-G’s sequence identity is quite high (59%), and they presumably have very similar over all structure, most likely the binding determinant is based on some surface-exposed residues.
Figure.4. Binding test of FusB and E. coli EF-G compared to FusB and S.a EF-G. FusB was incubated with the two EF-Gs separately in FusB gel filtration buffer (20mM Tris-HCl, 300mM NaCl, 5mM 2-Mercaptoethanol) for 10mins and then run on a 8-25% gradient native-PAGE (GE healthcare).
Complex purification
To increase the chance of complex crystallization, an EF-G construct without His-tag was designed for complex purification. FusB_SN was used to fish out EF-G. By doing so, the complex with just one short His-tag can be purified directly and there is no need to calculate the two protein concentrations separately for mixing. Here we compare the size exclusion (HiLoad 16/60 Superdex200 TM, GE healthcare) results from the EF-G purification and complex purification (Figure.5 A). There are two peaks for complex purification. One comes earlier (69.81ml) than EF-G monomer (74.37ml) peak, which is conformed by SDS-PAGE to be the complex (Figure.5 B). Another peak that comes later than EF-G is excess FusB and there is no peak around the EF-G monomer elution volume. All these proteins have larger molecular weight calculated from the elution volume than expected theoretically, possibly because they have elongated shapes.
Figure.5. (A) Complex purification comparing with EF-G purification by size exclusion column HiLoad 16/60 Superdex200 TM (GE healthcare) The first peak form EF-G purification is EF-G dimmer or precipitation (B) SDS-PAGE results from different peaks
FusB and S.a EF-G affinity test by fluorescence spectroscopy
The fluorescence spectroscopy method was used to determine the binding affinity between FusB and S.a EF-G. However, due to the inner filter effect [19]
of FusB, protein-protein interaction could only be poorly determined to 1μM (data not shown).
Digested S.a EF-G and FusB binding test
To investigate which part of S.a EF-G in interacting with FusB, S.a EF-G was first digested by chymotrypsin, trypsin, subtilisin and papain (Floppy-Choppy, Jena Bioscience) separately and then checked by SDS-PAGE. Among these four proteases, trypsin and papain could cut EF-G into two distinguishable bands with molecular weight around 60kD and 25kD (Figure6 A). Based on this result, His-tagged FusB was incubated with the trypsin-digested S.a EF-G in Ni SepharoseTM. After washing out the fragments which could bind neither nickel
nor FusB, everything was eluted by 500mM imidazole. The samples were run on both SDS and native-PAGE and checked by mass spectrometry.
Digested S.a EF-G had 5 distinctive bands on SDS-PAGE (Figure.6 B). The top band is full-length EF-G not digested by trypsin. The same band can also be found in previous test-experiments (Figure.6 A). Three bands with molecular weight around 55kD, 50kD and 45 kD were found both in digested EF-G and in the flow-through of the Ni SepharoseTM, meaning that they bind neither to FusB nor to nickel. Mass spectrometry analysis showed that they all belong to S.a EF-G domainⅠand domain Ⅱ (Figure.6 B). The larger band (Figure.6 B, bind
“1”) contains fragments “a” (Figure.6 D) at the N-terminus of the protein.
Taking into account the molecular weight and the fact it could not bind to nickel, this suggests it contains the main part of domainⅠand domain Ⅱ but not the His-tag from the N-terminus. Since band “1” was not further digested into smaller pieces as suggested by PeptideCutter (ExPASy Proteomics Server) and it ran as a clear band on the native-PAGE (Figure.6 C), probably its structure was not destroyed by the digestion. Therefore, we conclude that domainⅠand domain Ⅱof S.a EF-G can not bind to FusB just by themselves.
Band “4” (Figure.6 B) with a molecular weight similar with FusB (26kD), became a very pale band in flow though meaning it could have some interaction with nickel or FusB. Mass spectrometry analysis shows that it containing peptides fragments of EF-G domainⅢ (only C-terminal part), domainⅣand domainⅤ (Figure.6 D, with blue label). Because this part of EF-G only has 3 histidines spread out in the sequence, it is unlikely to bind the Ni SepharoseTM, indicating that it probably interacts with FusB. According its size from SDS-PAGE, most likely it contains the whole part of domainⅣ, domainⅤand few residues from the C-terminus of domain Ⅲ. Another evidence is that, although bind “5” (Figure.6 B) is identified as FusB, two peptide peaks, 1837.79 (Figure.6 D, “b”) and 1979.82 (Figure.6 D, “c”) which do not belong to FusB, were detected and recognized parts of domainⅣ and domainⅤ. These results indicate that the FusB mainly interact with S.a EF-G with the domains that doesn’t have any interaction with Fusidic acid.
Figure.6 (A) S.a EF-G was digested by chymotrypsin, trypsin, subtilisin and papain at room temperature for 30mins and 60mins. (B) Different samples from the fragment binding test run on a 20% Homogeneous SDS-PAGE (GE healthcare). (C) Different samples from the fragment binding test run on 8-25% gradient gel (GE healthcare). (D)The mass spectroscopy results of different bands on SDS and native PAGE. S.a EF-G sequence is colored by different domains. Fragments labeled yellow are the ones found in bands 1, 2 and 3 except fragment “a” just found in band 1.
Fragments labeled gray were detected from band 4. Two peptides marked “b”and “c”also recognized belonging to S.a EF-G from band 5, which is mainly FusB.
Crystallization and crystallography
FusB_LN was first purified and screened with three commercial crystallization kits, Easy Xtal JCSG+ (Molecular Dimensions), MorpheusTM (Molecular Dimensions) and Structures Screen (Molecular Dimensions) using a crystallization robot. No protein crystal was found even after two months.
The second batch of FusB_SN was purified and screened using the same screens. In MorpheusTM, 6 conditions with crystals were found after 1 month (Table.3). The crystals from condition C8 were successfully reproduced in manually prepared hanging drops. The crystals were tested for diffraction at a synchrotron beamline. However, the X-ray results showed that these crystals diffracted to low resolution (7Å). To improve the crystals, the crystallization condition was first optimized by grid-screening with different combinations of pH and precipitant concentrations. Combinations of pH7.1, 32% precipitant and pH7.7, 36% precipitant were used for further tests because they gave crystals in different shapes (Figure.7 A, B).
ADDit screen was used for further optimization. Crystal could grow in the present of low concentration hexamine cobalt trichloride, magnesium chloride hexhydrate, guanidine HCl and N-acetyl-L-cysteine (Figure.7 C, D, E, F), but these additives failed to improve the resolution.
Selenomethionyl protein was produced using the methionine pathway inhibition method [20] for multi-wavelength anomalous diffraction (MAD). After 1 month, some tiny crystals were found but they were not good enough for analysis.
The complex with FusB_SN and EF-G without His-tag was subjected to the JCSG+, MorpheusTM, Structure and ProComplex screens but no crystals appeared.
Table3 A. Hits formulation from MorpheusTM screen
Well Buffer system Mix of additives Mix of precipitants
B8 0.1M MOP/HEPES-Na pH7.5 0.03M of each halide C8 0.1M MOP/HEPES-Na pH7.5 0.03M of each NPS G8 0.1M MOP/HEPES-Na pH7.5 0.03M of each carboxylic acid A12 0.1M bicine/Trizma base pH8.5 0.03M of each divalent cation C12 0.1M bicine/Trizma base pH8.5 0.03M of each NPS G12 0.1M bicine/Trizma base pH8.5 0.03M of each carboxylic acid
12.5% w/v PEG 1000, 12.5%
w/v PEG 3350, 12.5 v/v MPD
Table3 B. Component of MorpheusTM additives mix
Name Component
Halides Sodium fluoride, Sodium bromide, sodium iodide NPS Sodium nitrate, Sodium hydrogen phosphate, ammonium sulfate Divalent cation Magnesium chloride, Calcium chloride
Carboxylic acid Sodium formate, Ammonium acetate, trisodium citrate, Sodium potassium L-tartrate, Sodium oxamate
Figure.7 (A) Crystals in MorpheusTM C8, pH7.1, 32% precipitant (B) Crystals in MorpheusTM C8, pH7.7, 36% precipitant. (C) FusB co-crystallized with hexamine cobalt trichloride. The yellow crystal in the central is the cobalt crystal with the colorless protein crystal covering it. (D) FusB crystal grew with the present of 5×10-3M guanidine HCl. (E) FusB crystal grown with the present of 5×10-3M magnesium chloride hexhydrate. (F) FusB crystal grew with the present of 5×10-3M N-acetyl-L-cysteine.
The obtained data sets were processed by XDS. Systematic absences from data set A indicates that the most probable space group is P21212 with cell dimensions a=187.516, b=99.537, c=129.703, α=β=γ=90°. Self-rotation analysis with polarrfn shows a 4-fold axis 3° away from a crystallographic 2-fold.
The crystal has a severe anisotropy problem (Figure.8) which means the resolution is direction related. In this case, dimension b and c are diffracted to 4.5Å; dimension a can diffract beyond 4 Å and data completeness will dramatically fall at high resolution shell (Table.4) thus, further optimization of the crystal is needed to solve this problem.
Table4. Data collection statistics of FusB crystals
Data set A Data set B
Resolution 50-3.69 Å 50-3.8 Å
Highest space group
determined P21212 P2122
Cell parameters a=187.5, b=99.5, c=129.7, α=β=γ=90°
a=185.7, b=97.9, c=127.8, α=β=γ=90°
Unique reflections 20840 18626
Completeness
96.8% (5.0 Å) 70.0% (4.17Å) 53.2% (3.68 Å) 77.5% (total)
90.8% (5.11 Å) 68.0% (4.05Å) 44.3% (3.81 Å)
78.7% (total)
1/σ 9.11 10.14
Figure.8 Data set B was analysised by diffraction anisotropy server:
http://www.doe-mbi.ucla.edu/~sawaya/anisoscale.
Discussion
Structure determination of FusB and FusB - EF-G complex
Further optimization of the FusB crystal is needed to get higher resolution and to minimize the anisotropy problem. First thing to try is to crystallize FusB without His-tag. Since long-tagged FusB can not be crystallized, it is a hint that the His-tag might disturb the crystal packing. Second, we know that FusB has a high pI (8.9) and high lysine content (13%), proteins of this kind are usually not easy to crystallize, because high charge and lysine on protein surface is not good for protein packing, which often requires hydrophobic interaction and low surface entropy [21]. FusB crystals could also suffer from this. The easiest thing to try is to change buffer pH. Theoretically, when the pH is near pI, the net charge on the surface of protein is minimized. Buffer screen showed that FusB is stable in different buffers from pH 6.0 to 10.0, thus, a pH screen from 8.5 to 9.5 is worth trying. Lysine methylation is often used as a standard rescue method when the lysine content is high. After methylation, the lysines on the protein surface are “neutralized” and often cause a drop of pI and facilitate the crystallization. Based on the same principle, mutation of the lysine on protein surface could also help. Since no homologue structure is known for FusB, it is hard to decide which lysine to mutate; leaving lysine methylation the best way to try. We noticed that FusB tend to bind nucleic acids during purification. This means that it probably prefer an environment with the presence of nucleic acid,
thus by adding some ATP or ADP could help to stabilize the protein.
For the FusB – EF-G complex, lysine methylation is also worth trying because it can change protein surface property. In order to increase the crystallization probability and to conform which domain of EF-G can bind to FusB, we are now constructing hybrid EF-Gs with different domains from E. coli and S. aureus EF-G. If a certain domain is found to bind to FusB, a complex of FusB and that single domain could also be tested for crystalization.
The binding target of FusB
The binding experiment of FusB with protease digested EF-G was aiming to find out where FusB binds. This will help us to understand the resistance mechanism and maybe facilitate the complex crystallization by designing some small binding partner. The overall structure of EF-G is conserved among different species. EF-G can be divided into 5 different domains (Figure.9 A), where domainⅠand domain Ⅱ are quite similar to EF-Tu (another bacterial elongation factor) [22]. Domain Ⅲ acts as a linker of domainⅠ,Ⅱ and domainⅣ,Ⅴ. The structure of ribosome with EF-G trapped by fusidic acid reveals the binding of fusidic acid involves domainⅠ,Ⅱ and domain Ⅲ.
When EF-G is locked on the ribosome, domains Ⅲ and Ⅳ are relatively buried inside; therefore they are not easy for FusB to access (Figure.9 B, C). Since our result showed that domainⅠand Ⅱ together can not bind to FusB, leaving only domainⅤ. There are some other evidences also supporting that domainⅤ is responsible for the binding. First, the mass spectroscopy identified two peptide fragments belong to domainⅣ and domainⅤ eluted together with FusB.
Second, the multiple alignment between S. aureus, L. monocytogenes, and E. faecalis EF-G (all have FusB homologues and intrinsically resistance to fusidic acid) and E. coli EF-G. shows that if we split domainⅤ into two parts, the one facing to the ribosome is very conserved, however, the other part exposed to the environment is only conserved in S. aureus, L.monocytogenes and E.faecalis (Figure.9 G, H). Actually, there are only four places where E. coli EF-G has low similarly with the other three species and two of them belong to domainⅤ. This indicates the face of domainⅤ exposed to the environment could possibly be the FusB binding determinant.
If this assumption is right, how can the binding of domainⅤ cause the resistance when the fusidic acid binds to other parts of the protein? A previous study observed a movement between domain Ⅲ,Ⅳ,Ⅴ and domainⅠ,Ⅱ [23]
suggesting that domainⅤ and domainⅢ are functionally closely related. For instance, a movement of domainⅤ could change the conformation of domainⅢ.
Interestingly, some mutations, F652S, Y654N, A655V, located on domainⅤ and were found to be related to fusidic acid resistance [10]. Furthermore, they all
belong to the conserved part of domainⅤ and facing towards the ribosome (Figure.9 A, G). This means the changes on domainⅤ itself could cause fusidic acid resistance.
Figure.9 (A) S.a EF-G domains were depicted in different colors. (PDB: 2XEX) the mutations on domainⅢ and domainⅤ are highlighted (B) Side view of EF-G binding on the ribosome, showing that domains Ⅰ, Ⅱ and Ⅴ are more exposed. (PDB: 2WRI, 2WRL) (C) Top view of EF-G binding on the ribosome (D) EF-G is extracted from the ribosome, showing the binding side of fusidic acid (E) fusidic acid (F) Multiple alignment showing that domain Ⅲ is very conserved among S. aureus, L.monocytogenes, E.faecalis and E. coli. Mutations that cause fusidic acid resistance are labled. (G) Multiple alignment of domianⅤ, two pieces of sequences are only conserved in S. aureus, L.monocytogenes and E.faecalis. (H) S.a EF-G with highlighted sequences which only conserved among S. aureus, L.monocytogenes and E.faecalis,
Hypothetical resistance mechanisms
As discussed previously, the mechanism of FusB mediated fusidic acid resistance doesn’t fall into any of the four classic strategies bacteria used to against antibiotics. Since FusB can bind to S. aureus EF-G and rescue the translation but not E. coli, this binding is probably the key for the resistance.
Assuming the FusB binds to EF-G domainⅤ, there are two possible FusB induced resistance mechanisms. The first one is due to changes in the fusidic acid binding environment. This mechanism is like some point mutations in domainⅠ and domainⅢ (P88A, V90I, L461K and H457Y/H457Q) that may affect the fusidic acid binding [3] [24]. Another possible mechanism is that FusB could facilitate the release of the EF-G from the ribosome and fusidic aicd drop off afterwards.
As mentioned before, we found that FusB tend to bind nuclear acids in vitro (data not shown). A binding experiment of FusB with ribosome in presence and in absence of EF-G and fusidic acid will clarify if FusB can interact with the ribosome. Since we don’t know if FusB will constantly bind to EF-G in vivo or it needs to be re-cycled by other factors, the ribosome could be a potential candidate competing with EF-G.
We also plan to test FusB in an in vitro translation assay to see how it will affect the translation in presence and in absence of the fusidic acid. It will tell whether FusB can rescue the in vitro translation and at what cost. Because FusB may interfere the normal functions of EF-G or the ribosome, by comparing the translation rate with and without FusB, we can investigate if there is any side effect of this resistance.
Acknowledgement
First and foremost I’d like to thank my supervisor Maria Selmer. For the chance you gave me and your patience to guide me into structure biology. Everyday I had in your group is fantastic adventure. I remember although sometimes the experiment seems going no where, you always have faith in me and encourage me until the break though comes.
I’ll also thanks to our group member Cha San Koh, Kristina Bäckbro, Chen Yang and Avinash Punekar, for all your kindly help and suggestions. Especially Cha San, She always came first when I needed help.
Thanks to Alwyn Jones, Lars Liljas and Terese Bergfors, who taught me the course crystallography.
Thanks to Celestine Chi, and Per Jemth for the help of fluorescence spectroscopy test.
Thanks to Suparna Chandra Sanyal for the S.a EF-G vector.
Thanks to Diarmaid Hughes for the extraction of pUB101 plasmid.
Reference
[1] W., Roholt, K. and Tybring, L. 1962. Fucidin, A new orally active antibiotic. Godtfredsen.
Lancetⅰ:928-931
[2] James W. Bodley, Franklin J. Zieve, Lillian Lin and Sandra T. Zieve. 1969. Formation of the ribosome-G factor-GDP complex in the presence of fusidic acid. Biochemical and Biophysical Research Communications, 37 (3): 437-443.
[3] Yong-Gui Gao, Maria Selmer,Christine M. Dunham, Albert Weixlbaumer, Ann C. Kelley, V.
Ramakrishnan. 2009. The structure of the ribosome with elongation factor G trapped in the posttranslocational state. Science. 326:694–699.
[4] Sotirios Tsiodras, Howard S Gold, George Sakoulas et al. 2001. Linezolid resistance in a clinical isolate of Staphylococcus aureus. Sotirios Tsiodras. Lancet 358: 207-208
[5] Mason BW, Howard AJ, Magee JT. 2003. Fusidic acid resistance in community isolates of methicillin-susceptible Staphylococcus aureus and fusidic acid prescribing. J Antimicrob Chemother. 51:1033–6.
[6] Leclercq, R., and P. Courvalin. 1991. Bacterial resistance to macrolide, lincosamide, and streptogramin antibiotics by target modification. Antimicrob. Agents Chemother.
35:1267-1272
[7] Wise, E.M., Jr, and Abou-Donia, M.M. 1975. Sulfonamide resistance mechanism in
Escherichia coli: R plasmids can determine sulfonamide-resistant dihydropteroate synthases.
Proc Natl Acad Sci USA 72: 2621–2625.
[8] Li, X., and H. Nikaido. 2004. Efflux mediated drug resistance in bacteria. Drugs 64:159-204 [9] O'Neill AJ, Larsen AR, Henriksen AS, Chopra I. A fusidic acid-resistant epidemic strain of
Staphylococcus aureus carries the fusB determinant, whereas fusA mutations are prevalent in other resistant isolates. Antimicrob Agents Chemother 2004; 48: 3594–3597.
[10] Nagaev I, Björkman J, Andersson DI, Hughes D. 2001. Biological cost and compensatory evolution in fusidic acid-resistant Staphylococcus aureus. Mol Microbiol . 40: 433–9.
[11] Besier S, Ludwig A, Bade V, Wichelhaus T. 2003. Molecular analysis of fusidic acid resistance in Staphylococcus aureus. Mol Microbiol. 47:463–469
[12] Projan, S.J. 2000. Antibiotic resistance in the staphylococci, in Gram-positive Pathogens.
American Society for Microbiology: 463-470
[13] Chopra, I. 1976. Mechanisms of resistance to fusidic acid in Staphylococcus aureus. J. Gen.
Microbiol. 96:229-238.
[14] O'Neill, A. J., and I. Chopra. 2006. Molecular basis of fusB-mediated resistance to fusidic acid in Staphylococcus aureus. Mol. Microbiol. 59:664-676.
[15] Traub, W.H. 1998. Antibiotic susceptibility testing of clinical isolates of Enterococcus faecalis and E.faecium. Chemotherapy 44: 217-229
[16] Troxler, R., von Graevenitz, A., Funke, G., Wiedemann, B.,and Stock, I. 2000. Natural antibiotic susceptibility of List-eria species: L. grayi, L. innocua, L. ivanovii, L.
monocyto-genes, L. seeligeri and L. welshimeri strains. Clin Microbiol Infect. 6: 525–535 [17] P. Gilot, Y. Jossin and J. Content. 2000. Cloning, sequencing and characterization of a
Listeria monocytogenes gene encoding a fibronectin-binding protein. J Med Microbiol.49:
887–896.
[18] S. van den Berg, P. Lofdahl. 2005. T. Hard and H. Berglund, Improved solubility of TEV protease by directed evolution, J Biotechnol 121: 291–298.
[19] Lakowicz JR. 1983. Principles of Fluorescence Spectroscopy. New York: Plenum
[20] S. Doublie. 1997. Preparation of selenomethionyl proteins for phase determination. Methods Enzymol. 276: 523–530.
[21] Walter, T. S., Meier, C., Assenberg, R., Au, K. F. et al., Lysine methylation as a routine rescue strategy for protein crystallization. Structure 2006, 14, 1617-1622.
[22] Czworkowski J, Wang J, Steitz TA, Moore PB. 1994. The crystal structure of elongation factor G complexed with GDP, at 2.7 Å resolution. The EMBO journal. 13:3661-3668 [23] Connell, S.R. et al. 2007. Structural basis for interaction of the ribosome with the switch
regions of GTP-bound elongation factors. Mol. Cell 25, 751–764
[24] Lannergard J, Norstrom T, Hughes D. 2009.Genetic determinants of resistance to fusidic acid among clinical bacteremia isolates of Staphylococcus aureus. Antimicrob Agents Chemother 53:2059–65.
