FEBS Letters 586 (2012) 109–115

journal homepage: www.FEBSLetters .org

Discovery of a novel enzymatic cleavage site for botulinum neurotoxin F5

Suzanne R. Kalb a, Jakub Baudys a, Robert P. Webb b, Patrick Wright b, Theresa J. Smith b,Leonard A. Smith c, Rafael Fernández d, Brian H. Raphael e, Susan E. Maslanka e, James L. Pirkle a,John R. Barr a,⇑a Centers for Disease Control and Prevention, National Center for Environmental Health, Division of Laboratory Sciences, 4770 Buford Hwy, NE, Atlanta, GA 30341, USAb Integrated Toxicology, United States Army Medical Research Institute of Infectious Diseases (USAMRIID), Ft. Detrick, MD 21702, USAc Office of the Chief Scientist, Medical Research and Materiel Command (MRMC), Fort Detrick, MD 21702, USAd Area Microbiologia, Universidad Nacional de Cuyo, Mendoza, Argentinae Centers for Disease Control and Prevention, National Center for Emerging and Zoonotic Infectious Diseases, Enteric Diseases Laboratory Branch, Atlanta, GA 30329, USA

a r t i c l e i n f o

Article history:Received 20 October 2011Revised 23 November 2011Accepted 28 November 2011Available online 9 December 2011

Edited by Maurice Montal

Keywords:Botulinum neurotoxinBoNT/FSynaptobrevin-2Mass spectrometry

0014-5793/$36.00 Published by Elsevier B.V. on behadoi:10.1016/j.febslet.2011.11.033

Abbreviations: BoNT, botulinum neurotoxin; SNAPprotein; PBS, phosphate buffered saline; TPGY, trypextract; CHCA, alpha-cyano-4-hydroxy cinnamic acMALDI, matrix-assisted laser desorption/ionization; F⇑ Corresponding author. Fax: +1 770 488 0509.

E-mail address: jbarr@cdc.gov (J.R. Barr).

a b s t r a c t

Botulinum neurotoxins (BoNTs) cause botulism by cleaving proteins necessary for nerve transmis-sion. There are seven serotypes of BoNT, A–G, characterized by their response to antisera. Many ser-otypes are further distinguished into differing subtypes based on amino acid sequence, some ofwhich result in functional differences. Our laboratory previously reported that all tested subtypeswithin each serotype have the same site of enzymatic activity. Recently, three new subtypes ofBoNT/F; /F3, /F4, and /F5, were reported. Here, we report that BoNT/F5 cleaves substrate synaptobre-vin-2 in a different location than the other BoNT/F subtypes, between 54L and 55E. This is the firstreport of cleavage of synaptobrevin-2 in this location.

Structured summary of protein interactions:BoNT/F5 cleaves Synaptobrevin-2 by protease assay (View interaction: 1, 2)BoNT/F1 cleaves Synaptobrevin-2 by protease assay (View interaction: 1, 2)

Published by Elsevier B.V. on behalf of the Federation of European Biochemical Societies.

1. Introduction each serotype, and this has led to the designation of subtypes,

Botulism is a disease which can be fatal if untreated and iscaused by exposure to any one of the highly toxic proteins knownas botulinum neurotoxins (BoNTs). BoNT are composed of a heavychain, which binds to receptors on the neuron, and a light chainthat is a protease. In vivo, the BoNT light chain cleaves proteinsnecessary for nerve signal transmission. This enzymatic cleavageleads to flaccid paralysis. Botulinum neurotoxins are currentlyclassified into seven serotypes, labeled A–G. BoNT/A, /C, and /Ecleave SNAP-25 (synaptosomal-associated protein) [1–6]whereasBoNT/B, /D, /F, and /G cleave synaptobrevin-2 (also known asVAMP-2) [7–11].

Our laboratory previously reported on BoNT’s ability to cleave apeptide substrate that mimics its natural target, with peptidecleavage detected by mass spectrometry [12,13]. BoNT/A through/F are known to exhibit genetic and amino acid variance within

lf of the Federation of European Bi

-25, synaptosomal-associatedticase-peptone-glucose-yeastid; TFA, trifluoroacetic acid;A, formic acid

some with identified antigenic differences [14–16]. However, itwas uncertain if the amino acid variance would also result in dif-ferential enzymatic activity. Therefore, previous work in our labo-ratory focused on determining the cleavage location of all availablesubtypes of BoNT/A, /B, /E, and /F, the serotypes known to causebotulism in humans. Previously, 15 such subtypes were availablein sufficient quantities for testing, and all BoNT subtypes withinthe same serotype cleaved their respective substrate in the samelocation [17]. At the time that this work was reported, only foursubtypes of BoNT/F were known.

In 2010, three new subtypes of BoNT/F; /F3, /F4, and /F5, werereported based on phylogenetic analysis of botulinum neurotoxintype F genes [18]. All three subtypes were defined as BoNT/F basedon their ability to cause signs of botulism in mice which could beprevented by preincubation of toxin with F antitoxins, althoughit was reported that the BoNT/F5 toxin required 20 times moretype F antiserum to neutralize an equivalent amount of BoNT/F1toxin [19]. Additionally, it was noted that BoNT/F5 was highly dif-ferent in its gene and amino acid sequences compared to the otherBoNT/F subtypes [18]. For instance, Table 1 shows that the wholetoxin (holotoxin) of BoNT/F1 and BoNT/F2 have similar (83.5%identical) amino acid sequences. However, BoNT/F1 and /F5

ochemical Societies.

Table 1Percent amino acid identity between BoNT/F [18].

HolotoxinF1 F2 F3 F4 F5 F6 F7

F1 – 83.5 83.9 92.3 69.9 87.7 73.7F2 81.8 – 97.1 83.5 74.2 90.2 68.9F3 82.9 97.9 – 83.8 74.0 90.1 69.2F4 96.4 82.7 83.8 – 69.5 87.1 72.2F5 47.3 46.3 46.3 47.6 – 74.0 63.9F6 94.3 81.5 82.9 93.6 48.3 – 69.9F7 63.3 59.5 60.8 64.2 46.9 63.3 –

F1 F2 F3 F4 F5 F6 F7

Enzymatic domain (light chain).

110 S.R. Kalb et al. / FEBS Letters 586 (2012) 109–115

holotoxins are dissimilar (69.9% identical). The amino acid dissim-ilarity (47.3% identity) observed in the light chain (the enzymaticdomain) between BoNT/F5 and BoNT/F1 is even more pronounced.Until the identification of BoNT/F5, this level of variation in thelight chain amino acid sequences within a serotype had not beenreported. Indeed, an amino acid identity of only 47% is much lessthan the light chain identity between differing serotypes whichhave different sites of enzymatic activity, such as 56% betweenBoNT/E and /F and 61% between BoNT/B and /G, based on aminoacid comparison.

The second-most divergent BoNT/F subtype, F7, was reported tohave different peptide substrate recognition requirements [20].BoNT/F7, produced by Clostridium baratii, is 73.7% identical toBoNT/F1 at the holotoxin level and 63.3% identical at the enzymaticdomain level. BoNT/F7 was reported to cleave both synaptobrevin-2and peptides based on synaptobrevin-2 in the same location as theother BoNT/F subtypes, between 58Q and 59K. However, BoNT/F7did not cleave a peptide substrate cleaved by other availableBoNT/F subtypes based on the sequence of synaptobrevin-2, andwould only cleave TSNRRLQQTQAQVDEVVDIMRVNVDKVLERDQKL-SELDDRADAL, a lengthened peptide substrate [20]. A similar phe-nomenon was reported with two neurotoxins which also cleavesynpatobrevin-2, BoNT/B and tetanus toxin [21]. This was due todifferential binding of BoNT/F7 with synaptobrevin-2 as comparedto other BoNT/F. Because the amino acid sequence of the BoNT/F5light chain is even more divergent than BoNT/F7, it was theorizedthat BoNT/F5 may have a different molecular reaction with its sub-strate. In this work, we describe the discovery of a novel enzymaticcleavage site on synaptobrevin-2 by BoNT/F5.

2. Materials and methods

2.1. Materials

Synaptobrevin-2, SNAP-25, and synatxin recombinant proteinswere purchased from GenWay Biotech, Inc. (San Diego, CA). Mono-clonal antibody 6F5 was obtained from Dr. James Marks at theUniversity of California at San Francisco. Dynabeads� (M-280/Strep-tavidin) were purchased from Invitrogen (Carlsbad, CA.) at 1.3 g/cm3

in phosphate buffered saline (PBS), pH 7.4, containing 0.1% Tween�-20 and 0.02% sodium azide. All chemicals were from Sigma-Aldrich(St. Louis, MO) except where indicated. Peptide substrates LQQTQAQVDEVVDIMRVNVDKVLERDQKLSELDDRADAL and TSNRRLQQTQAQVDEVVDIMRVNVDKVLERDQKLSELDDRADAL were synthesized byBiopeptide (San Diego, CA). Commercially purified BoNT/F1 (strainLangeland) was purchased from Metabiologics (Madison, WI).

2.2. Gene construction of BoNT/F1 and F5 LC-GST

An existing BoNT/F1 LC ORF was PCR amplified using primers toadd 5’ BamHI and 3’ EcoRI sites. The PCR reaction product was runon a 1% TAE gel and the amplified band at approximately 1400 bp

was excised, purified, ligated into the homologous restriction en-zymes sites of pGEX6P-1, and transformed into the DH5-a strainof E. coli. A single recombinant colony was used to inoculate 5 mlof LB-50 mg/L ampicillin and grown overnight at 37 �C. PlasmidDNA was recovered, completely sequenced to insure accuracy,mobilized into E. coli strain BL21 DE3, and spread onto LB plateswith 50 mg/L ampicillin. A single recombinant colony was usedto inoculate a LB-50 mg/L ampicillin, grown overnight at 37 �C.The overnight culture was used to inoculate a fresh flask of 1L ofLB-50 mg/L ampicillin and grown to an OD of 0.7, cooled to18 �C, IPTG was added to a final concentration of 1 mM, and theinduction was allowed to proceed at 18 �C overnight. The sameprocedure was followed for BoNT/F5, with the exception that theBoNT/F5 light chain ORF (representing nucleotides 1-438 fromGU213211.1) was synthesized de novo using a generalized E. coliK12 codon bias and ligated into the BamHI/EcoRI sites of bacterialexpression vector pGEX6P-1.

2.3. Purification of fusion proteins

Both proteins bound to glutathione sepharose resin and wereeluted in the presence of 10 mM reduced glutathione; however,the protein was not purified at this stage and required furtherchromatography steps. Both proteins were further separated withcation exchange chromatography after a pH change from �8.0 to�5.7. After cation exchange chromatography the purified BoNT/F5 LC-GST was approximately 100 lg/ml, while the BoNT/F1 LC-GST was approximately 40 lg /ml.

2.4. Production and characterization of BoNT/F5 culture supernatant

CDC54075, originally isolated in March 1978 from soil in a corn-field in Mendoza, Argentina at latitude 34� 98’ and longitude 67�59’, was inoculated into 10 mls of Trypticase-Peptone-Glucose-Yeast Extract (TPGY) broth (Remel, Inc. Lexana, KS) at 35 �C for5 days in an anaerobe chamber system (Coy Laboratory Products,Inc. Grass Lake, MI) utilizing 10% Hydrogen, 5% CO2, and 85% Nitro-gen. The culture was tested by PCR for the presence of neurotoxingenes as previously reported [18]. The culture was centrifuged at4000�g for 10 min and filtered using a 0.45 lm syringe filter(Whatman #6876-2504) to remove viable cells. The presence ofbotulinum toxin in the culture supernatant was confirmed by ELI-SA and mouse bioassay [22].

2.5. Reaction of BoNT fusion protein with a peptide or protein target

The BoNT fusion proteins were serially diluted in water toachieve three different concentrations of each protein. For F1, con-centrations of 40 lg/ml (52.7 pmol/ml), 4 lg/ml (5.27 pmol/ml),and 2 lg/ml (2.64 pmol/ml) were used; whereas concentrationsof 100 lg/ml (1.32 nmol/ml), 10 lg/uL (132 pmol/ml), and 5 lg/ml (66 pmol/ml) were used for F5. 2 ll of each concentration offusion protein was added to 16 ll of reaction buffer consisting of0.05 M Hepes (pH 7.3), 25 mM dithiothreitol, and 20 lM ZnCl2.Finally, 2 ll of recombinant synaptobrevin-2 or peptide sub-strate (LQQTQAQVDEVVDIMRVNVDKVLERDQKLSELDDRADAL orTSNRRLQQTQAQVDEVVDIMRVNVDKVLERDQKLSELDDRADAL) wasadded to achieve a final concentration of 250 ng/ll (18 pmol/ll)for synaptobrevin-2 and 50 pmol/ll for the peptide reactions. Allsamples then were incubated at 37 �C for 4 hrs with no agitation.

2.6. Reaction of BoNT/F5 holotoxin with a peptide or protein target

Monoclonal antibody 6F5, a clonal relative of mAb E17.1 [23]engineered toward higher affinity for BoNT/F subtypes, was immobi-lized to streptavidin Dynabeads� after being rinsed three times with

S.R. Kalb et al. / FEBS Letters 586 (2012) 109–115 111

HBS-EP buffer (GE Healthcare; Piscataway, NJ.). A 2 lg aliquot ofantibody was used per 100 ll of beads. A standard orbital shakerwas used to bind antibody onto the beads for 1.5 h. An aliquot of20 ll of antibody-coated beads was added to a solution of 400 llof HBS-EP buffer and 100 ll of culture supernatant containingBoNT/F5. After mixing for 1 h with constant agitation at room tem-perature, the beads were washed twice in 1 ml each of HBS-EP bufferand then once in 100 ll of water. Negative controls consisted ofblank culture supernatant medium with no toxin, with the remain-der of the extraction protocol as above. Positive controls consisted ofblank culture supernatant medium spiked with 20 mLD50 of BoNT/F1complex. The level of BoNT/F5 toxin in the culture used was notknown.

The beads were reconstituted in 18 ll of reaction buffer consist-ing of 0.05 M Hepes (pH 7.3), 25 mM dithiothreitol, and 20 lM ZnCl2

and 2 ll of recombinant synaptobrevin-2 or peptide substrate (LQQTQAQVDEVVDIMRVNVDKVLERDQKLSELDDRADAL or TSNRRLQQTQAQVDEVVDIMRVNVDKVLERDQKLSELDDRADAL). The final con-centration of substrate was 250 ng/ll (18 pmol/ll) for synaptobre-vin-2 and 50 pmol/ll for the peptide reactions. All samples thenwere incubated at 37 �C for 4 h with no agitation.

2.7. Peptide mass spectrometric detection

A 2-ll aliquot of each reaction supernatant was mixed with18 ll of matrix solution consisting of alpha-cyano-4-hydroxy cin-namic acid (CHCA) at 5 mg/ml in 50% acetonitrile, 0.1% trifluoro-acetic acid (TFA), and 1 mM ammonium citrate. A 0.5-ll aliquotof this mixture was pipeted onto each spot of a 192-spot matrix-assisted laser desorption/ionization (MALDI) plate (Applied Biosys-tems, Framingham, MA). Mass spectra of each spot were obtainedby scanning from 1100 to 4800 m/z in MS-positive ion reflectormode on an Applied Biosystems 4800 Proteomics Analyzer (Fra-mingham, MA). The instrument uses an Nd-YAG laser at 355 nm,and each spectrum is an average of 2400 laser shots.

2.8. Protein mass spectrometric detection

All reactions were first separated by using a nano-ACQUITYUPLCTM System (Waters, Milford, MA). Mobile phases were 0.04%TFA with 0.06% formic acid (FA) in water (mobile phase A) and0.04% TFA and 0.06% FA in acetonitrile (mobile phase B). Synaptob-revin-2 and cleavage products were trapped at 500 ng on a Pep-swift PS-DVB monolithic trapping column, 200 lm x 5 mm(Dionex, Sunnyvale, CA), and washed for 4 min at a flow rate of7.5 ll/min with 99% mobile phase A. Intact synaptobrevin-2 andcleavage products were eluted and separated by using a 70 minRP gradient at 750 nl/min (1–50% mobile phase B over 35 min)on a Pepswift PS-DVB monolithic 100 lm � 5 cm nanoscale LC col-umn (Dionex). The column temperature was set to 60 �C.

A NanoMate TriVersa (Ithaca, NY) was used for infusion and on-line LC coupling analysis of the samples at a capillary spray voltageof 1.82 kV. The mass spectral data were acquired on a Synapt HDMSQTOF (Waters); the instrument was calibrated for a mass range of550–4550m/z with Cesium Iodide through direct infusion. The sam-pling and extraction cone voltage were optimized at 40 V and 4 V,respectively, for maximum intact synaptobrevin-2 sensitivity bycomparing on-column injections. Source temperature was set to150 �C. A quadrupole RF transmission profile was defined to trans-mit masses from 800 to 3500 Da. Trap and transfer collision energieswere set to 6 V and 2 V, respectively, for maximum transmission ofthe most abundant synaptobrevin-2 charge states. The data were ac-quired in TOF V-mode at a mass range of 800-3000 m/z and a 2 scan/sacquisition time. All data were processed by using the Waters Mass-Lynx MaxEnt 1 software to obtain the deconvoluted mass at a rangeof 700–15000 Da with a mass resolution of 0.5 Da. All spectra were

processed with a uniform Gaussian damage model with an iterate toconvergence option selected.

3. Results

3.1. GST-BoNT/F1 cleaves synaptobrevin-2 as BoNT/F1 holotoxin

The sequence of recombinant synaptobrevin-2 is MRGSHHHHHHGMASMTGGQQMGRDLYDDDDKDRWGSHMSATAATAPPAAPAGEGGPPAPPPNLTSNRRLQQTQAQVDEVVDIMRVNVDKVLERDQKLSELD-DRADALQAGASQFETSAAKLKRKYW, and BoNT/F1, /F2, /F6, and /F7holotoxins have been reported to cleave it in the underlined location,between 58Q and 59K [20]. The GST-BoNT/F1 light chain fusion pro-tein was reacted with recombinant synaptobrevin-2, and Fig. 1shows that this resulted in two peaks in the mass spectrometerwhich correspond to cleavage of recombinant synaptobrevin-2 bythe F1 fusion protein. The peak at mass 13824.0 in Fig. 1A acquiredby electrospray ionization mass spectrometry corresponds to intactrecombinant synaptobrevin-2, whereas the peak at mass 10344.0corresponds to the N-terminal cleavage product. The peak at m/z3496.8 in Fig. 1B acquired by MALDI-TOF/MS corresponds to the C-terminal cleavage product, and the peak at 1749.4 corresponds tothe doubly-charged C-terminal cleavage product. Both cleavageproducts demonstrate that the GST-BoNT/F1 light chain fusionprotein cleaves recombinant synaptobrevin-2 between 58Q and59K, exactly where the BoNT/F1 holotoxin cleaves recombinant syn-aptobrevin-2.

3.2. GST-BoNT/F1 cleaves peptides which mimic synaptobrevin-2 asBoNT/F1

Additionally, this fusion protein was tested with two additionalpeptide substrates, LQQTQAQVDEVVDIMRVNVDKVLERDQKLSELD-DRADAL and TSNRRLQQTQAQVDEVVDIMRVNVDKVLERDQKLSEL D-DRADAL, whose sequences are based upon the sequence ofsynaptobrevin-2, and those results are in Figs. 1C and 1D. The peakat m/z 1345.7 in Figs. 1C and 1D corresponds to the C-terminal cleav-age product of the substrates. Note that the C-terminal cleavageproduct is the same in both cases, as the two peptides are identicalat the C-terminus. The peak at m/z 3168.6 in Fig. 1C corresponds tothe N-terminal cleavage product of the first peptide substrate, andthe peak at m/z 3782.9 in Fig. 1D corresponds to the N-terminalcleavage product of the second peptide substrate. In both cases,the GST-BoNT/F1 light chain fusion protein cleaved the peptide sub-strates in the above-underlined location, corresponding to a locationbetween 58Q and 59K of synaptobrevin-2. These data demonstratethat the GST-BoNT/F1 light chain fusion protein has the same activ-ity upon its substrates as BoNT/F1 holotoxin.

3.3. GST-BoNT/F5 cleaves synaptobrevin-2 and peptides in a differentlocation

Next, GST-BoNT/F5 light chain fusion protein was reacted withrecombinant synaptobrevin-2. Similar to the reaction with F1, thisresulted in peaks in the mass spectrum corresponding to N-termi-nal and C-terminal cleavage products, as seen in Fig. 2. However,the masses of these peaks did not correspond with cleavage be-tween 58Q and 59K. The peaks at mass 9815.0 (Fig. 2A) and m/z4025.2 (Fig. 2B) correspond to cleavage between 54L and 55E of syn-aptobrevin-2, with the peak at m/z 2013.6 corresponding to thedoubly-charged C-terminal cleavage product. The GST-BoNT/F5light chain fusion protein was also tested with the two peptidesubstrates used with BoNT/F1, and those results are in Figs. 2Cand 2D. The peak at m/z 1872.9 in Figs. 2C and 2D correspondsto the C-terminal cleavage product of each of the peptides. The

100

0

%

9000 10000 11000 12000 13000 14000 15000 16000 17000 18000 19000mass

5000 6000 7000 8000

13824.010344.0 A

1500 2300 3100 3900 4700 5500m/z

0

100%

Inte

nsity

1749.4

3496.8

B

1000 1900 2800 3700 4600 5500m/z

0

100

% In

tens

ity 1345.7 3168.6

1585.3 4495.3

C

1000 1900 2800 3700 4600 5500m/z

0

100

% In

tens

ity

1345.7

3782.9

D

Fig. 1. Mass spectra for the reaction of GST-BoNT/F1 light chain fusion protein with recombinant synaptobrevin-2 (1A, 1B), LQQTQAQVDEVVDIMRVNVDKVLERDQKLSELDDRADAL (1C), and TSNRRLQQTQAQVDEVVDIMRVNVDKVLERDQKLSELDDRADAL (1D). Peaks at mass 10344.0, m/z 3496.8, 1345.7, 3168.6, and 3782.9 indicate cleavageof the substrates by the GST-BoNT/F1 light chain fusion protein. Fig. 1A represents protein (ESI) data and Figs. 1B-D represent peptide (MALDI) data.

13824.0

9815.0

100

0 %

8000 8500 9000 9500 10000 10500 11000 11500 12000 12500 13000 13500 14000mass 100

A

1500 2300 3100 3900 4700 5500m/z0

% In

tens

ity 2013.6

4025.2

B

1000 1900 2800 3700 4600 5500m/z0

100

% In

tens

ity

2640.41872.9

4495.3

C

1000 1900 2800 3700 4600 5500m/z0

100

% In

tens

ity

1872.9

3254.6

D

Fig. 2. Mass spectra for the reaction of GST-BoNT/F5 light chain fusion protein with recombinant synaptobrevin-2 (2A, 2B), LQQTQAQVDEVVDIMRVNVDKVLERDQKLSELDDRADAL (2C), and TSNRRLQQTQAQVDEVVDIMRVNVDKVLERDQKLSELDDRADAL (2D). Peaks at mass 9815.0, m/z 4025.2, 1872.9, 2640.4, and 3254.6 indicate cleavage ofthe substrates by the GST-BoNT/F5 light chain fusion protein. Fig. 2A represents protein (ESI) data and Figs. 2B-D represent peptide (MALDI) data.

112 S.R. Kalb et al. / FEBS Letters 586 (2012) 109–115

peak at m/z 2640.4 in Fig. 2C corresponds to the N-terminal cleav-age product of the first peptide substrate, and the peak at m/z

3254.6 in Fig. 2D corresponds to the N-terminal cleavage productof the second peptide substrate. In all cases, the GST-BoNT/F5 light

13824.0

9815.0

100

0

%

8000 8500 9000 9500 10000 10500 11000 11500 12000 12500 13000 13500 14000mass

A

1100 1980 2860 3740 4620 5500m/z0

100

% In

tens

ity

2013.9

4025.3

B

1100 1980 2860 3740 4620 5500m/z0

100

% In

tens

ity

1873.0

2640.5

4495.6

C

1100 1980 2860 3740 4620 5500m/z0

100

% In

tens

ity

1873.0

3254.8

D

Fig. 3. Mass spectra for the reactions of BoNT/F5 holotoxin with recombinant synaptobrevin-2 (S-1A,S-1B), LQQTQAQVDEVVDIMRVNVDKVLERDQKLSELDDRADAL(S-1C), andTSNRRLQQTQAQVDEVVDIMRVNVDKVLERDQKLSELDDRADAL (S-1D). Peaks at mass 9815.0, m/z 4025.3, 1873.0, 2640.5, and 3254.8 indicate cleavage of the substrates byBoNT/F5 holotoxin. Fig. 3A represents protein (ESI) data and Figs. 3B-D represent peptide (MALDI) data.

S.R. Kalb et al. / FEBS Letters 586 (2012) 109–115 113

chain fusion protein cleaved the peptide substrates in a locationwhich corresponds to between 54L and 55E of synaptobrevin-2.

3.4. BoNT/F5 holotoxin cleaves synaptobrevin-2 and peptides as GST-BoNT/F5

Although the light chain of BoNT/F5 exhibits a different cleav-age pattern than that of the other BoNT/F, it was unclear that thispattern would be repeated with the holotoxin or progenitor formof the toxin produced by BoNT/F5-encoding C. botulinum. There-fore, holotoxin BoNT/F5 was reacted with recombinant synaptob-revin-2 and the two peptide substrates listed above, and thoseresults are in Fig. 3. The peaks at mass 9815.0 (Fig. 3A) and m/z4025.2 (Fig. 3B) correspond to cleavage between 54L and 55E of syn-aptobrevin-2, with the peak at m/z 2013.9 corresponding to thedoubly-charged C-terminal cleavage product. BoNT/F5 was alsotested with the two peptide substrates used with GST-BoNT/F5,and those results are in Figs. 3C and 3D. The peak at m/z 1872.9in Figs. 3C and 3D corresponds to the C-terminal cleavage productof each of the peptides. The peak at m/z 2640.4 in Fig. 3C corre-sponds to the N-terminal cleavage product of the first peptide sub-strate, and the peak at m/z 3254.6 in Fig. 3D corresponds to the N-terminal cleavage product of the second peptide substrate. The

Fig. 4. Amino acid sequences of all BoNT/F subtypes in the region of the active site. The avariant.

results are identical to that of Fig. 2, indicating that BoNT/F5 holo-toxin cleaves synaptobrevin-2 between 54L and 55E, in the samelocation as the GST-BoNT/F5 light chain fusion protein.

4. Discussion

As noted above, BoNT/F5 is the most divergent of the botulinumneurotoxins. Previously BoNT/F7, the second-most divergent of thebotulinum neurotoxins, was shown to interact with its substrate,synaptobrevin-2, differently from most of the BoNT/F subtypessince F7 required a longer peptide substrate for cleavage by thetoxin[20]. However the actual cleavage site was identical to theother known BoNT/F subtypes In contrast, the previously unknownF5 subtype cleaves the substrate at a different site than all otherBoNT/Fs tested. Of particular interest are several amino acid substi-tutions near the active site of BoNT/F5 as seen in Fig. 4. The activesite of the toxin, 228E, is conserved among all BoNT/F, and the re-gion immediately surrounding the active site is also conservedamong all BoNT/F, except for BoNT/F5. With BoNT/F5, three ofthe five amino acids immediately upstream of the active site differfrom all other BoNT/F. This degree of variance so close to the activesite is an important factor in the ability of this toxin to cleave syn-aptobrevin-2 in a different location.

ctive site (228E) is in red. Residues in blue indicate regions where BoNT/F5 is the only

Table 2Amino acid contacts between residues 25 and 57 of synaptobrevin-2 and BoNT/F1 [24]. Sequence alignments with all other BoNT/F are also included, with mutations shown in redand conserved mutations shown in green.

114 S.R. Kalb et al. / FEBS Letters 586 (2012) 109–115

BoNT/F5 has a large degree of amino acid variance, and this ac-counts for its ability to cleave synaptobrevin-2 in a different

location from all other BoNT/F. These amino acid differences are lo-cated not only at the interaction of the active site of BoNT/F5 with

S.R. Kalb et al. / FEBS Letters 586 (2012) 109–115 115

synaptobrevin-2 as discussed above, but also at other interactionsof BoNT/F5 with synaptobrevin-2. Of the amino acids in BoNT/F1which are reported to be important for interactions with synaptob-revin-2, there are many which are mutated in BoNT/F5. Table 2lists the residues of synaptobrevin-2 from 25N to 57D and their re-ported contact with corresponding residues of BoNT/F1 [24], withadded sequence alignments of all other BoNT/F. Of the 28 aminoacids in BoNT/F5 expected to be contacted by residues 25N to 57Dof synaptobrevin-2, approximately three-quarters are significantlyaltered from that of BoNT/F1, with 20 of the 28 residues differingsignificantly. This is in contrast to the other BoNT/F, in which 0to 14 residues differ. Six of the amino acids from 25N to 57D of syn-aptobrevin-2—33Q, 39V, 41E, 43V, 54L, and 57D—reduce the cleavabil-ity of synaptobrevin-2 by more than 66% when mutated,demonstrating that these residues are critical for binding of syna-ptobrevin-2 to BoNT/F1 [21]. Four of those six residues contact mu-tated residues on BoNT/F5, which could explain why residues 25Nto 70L of synaptobrevin-2 bind BoNT/F5 differently, resulting in dif-ferential cleavage. It should be noted that BoNT/F5 was found tocleave only synaptobrevin-2 and not additional SNARE proteinssuch as SNAP-25 or a truncated version of syntaxin (data notshown).

Cleavage between 54L and 55E of synaptobrevin-2, has neverbeen reported for any botulinum neurotoxin. BoNT/B, /D, and /Gare also known to cleave snaptobrevin-2, but these toxins cleavebetween 76Q and 77F, 59K and 60L, and 81A and 82A of synaptobre-vin-2, respectively [7,10,11]. Additionally, up to this point, all sub-types within a serotype have been reported to share the samecleavage site. Knowledge of a new cleavage site of BoNT will beof particular interest to scientists involved in developing medicalcountermeasures for botulism. Many scientists are investigatinginhibitors such as peptide substrate blockers and mAb as next gen-eration botulism therapeutics. Additionally, some botulism diag-nostic methods are being developed which are based on activitytowards a protein or peptide substrate. This report of a unique sub-strate cleavage site by BoNT/F5 subtype will alert scientists to po-tential functional variability within BoNT serotypes. Therefore, it isimportant to study the enzymatic activity of newly-emerging bot-ulinum neurotoxins in order to design effective inhibitors of BoNTsand assays to detect all BoNTs.

Acknowledgements

This work was supported by NIAID IAA 120-B18. The opinions,interpretations, conclusions, and recommendations are those ofthe authors and are not necessarily endorsed by the Centers forDisease Control and Prevention, the U.S. Army, the National Insti-tute of Allergy and Infectious Diseases, or the National Institutesof Health.

References

[1] Binz, T., Blasi, J., Yamasaki, S., Baumeister, A., Link, E., Sudhof, T.C., Jahn, R. andNiemann, H. (1994) Proteolysis of SNAP-25 by types E and A botulinalneurotoxins. J Biol Chem 269, 1617–1620.

[2] Blasi, J. et al. (1993) Botulinum neurotoxin A selectively cleaves the synapticprotein SNAP-25. Nature 365, 160–163.

[3] Foran, P., Lawrence, G.W., Shone, C.C., Foster, K.A. and Dolly, J.O. (1996)Botulinum neurotoxin C1 cleaves both syntaxin and SNAP-25 in intact andpermeabilized chromaffin cells: correlation with its blockade of catecholaminerelease. Biochemistry 35, 2630–2636.

[4] Schiavo, G., Rossetto, O., Catsicas, S., Polverino de Laureto, P., DasGupta, B.R.,Benfenati, F. and Montecucco, C. (1993) Identification of the nerve terminaltargets of botulinum neurotoxin serotypes A, D, and E. J Biol Chem 268,23784–23787.

[5] Schiavo, G., Santucci, A., Dasgupta, B.R., Mehta, P.P., Jontes, J., Benfenati, F.,Wilson, M.C. and Montecucco, C. (1993) Botulinum neurotoxins serotypes Aand E cleave SNAP-25 at distinct COOH-terminal peptide bonds. FEBS Lett 335,99–103.

[6] Williamson, L.C., Halpern, J.L., Montecucco, C., Brown, J.E. and Neale, E.A.(1996) Clostridial neurotoxins and substrate proteolysis in intact neurons:botulinum neurotoxin C acts on synaptosomal-associated protein of 25 kDa. JBiol Chem 271, 7694–7699.

[7] Schiavo, G., Benfenati, F., Poulain, B., Rossetto, O., Polverino de Laureto, P.,DasGupta, B.R. and Montecucco, C. (1992) Tetanus and botulinum-Bneurotoxins block neurotransmitter release by proteolytic cleavage ofsynaptobrevin. Nature 359, 832–835.

[8] Schiavo, G., Malizio, C., Trimble, W.S., Polverino de Laureto, P., Milan, G.,Sugiyama, H., Johnson, E.A. and Montecucco, C. (1994) Botulinum Gneurotoxin cleaves VAMP/synaptobrevin at a single Ala-Ala peptide bond. JBiol Chem 269, 20213–20216.

[9] Schiavo, G., Shone, C.C., Rossetto, O., Alexander, F.C. and Montecucco, C. (1993)Botulinum neurotoxin serotype F is a zinc endopeptidase specific for VAMP/synaptobrevin. J Biol Chem 268, 11516–11519.

[10] Yamasaki, S. et al. (1994) Cleavage of members of the synaptobrevin/VAMPfamily by types D and F botulinal neurotoxins and tetanus toxin. J Biol Chem269, 12764–12772.

[11] Yamasaki, S., Binz, T., Hayashi, T., Szabo, E., Yamasaki, N., Eklund, M., Jahn, R.and Niemann, H. (1994) Botulinum neurotoxin type G proteolyses the Ala81-Ala82 bond of rat synaptobrevin 2. Biochem Biophys Res Commun 200, 829–835.

[12] Barr, J.R. et al. (2005) Botulinum neurotoxin detection and differentiation bymass spectrometry. Emerg Infect Dis 11, 1578–1583.

[13] Boyer, A.E. et al. (2005) From the mouse to the mass spectrometer: detectionand differentiation of the endoproteinase activities of botulinum neurotoxinsA-G by mass spectrometry. Anal Chem 77, 3916–3924.

[14] Gibson, A.M., Modi, N.K., Roberts, T.A., Hambleton, P. and Melling, J. (1988)Evaluation of a monoclonal antibody-based immunoassay for detecting type BClostridium botulinum toxin produced in pure culture and an inoculatedmodel cured meat system. J Appl Bacteriol 64, 285–291.

[15] Kozaki, S., Nakaue, S. and Kamata, Y. (1995) Immunological characterization ofthe neurotoxin produced by Clostridium botulinum type A associated withinfant botulism in Japan. Microbiol Immunol 39, 767–774.

[16] Smith, T.J. et al. (2005) Sequence variation within botulinum neurotoxinserotypes impacts antibody binding and neutralization. Infect Immun 73,5450–5457.

[17] Kalb, S.R. et al. (2008) The use of Endopep-MS to detect multiple subtypes ofbotulinum neurotoxins A, B, E, and F. International Journal of MassSpectrometry 278, 101–108.

[18] Raphael, B.H., Choudoir, M.J., Luquez, C., Fernandez, R. and Maslanka, S.E.(2010) Sequence diversity of genes encoding botulinum neurotoxin type F.Appl Environ Microbiol 76, 4805–4812.

[19] Gimenez, D.F. and Ciccarelli, A.S. (1968) Clostridium botulinum type F in thesoil of Argentina. Appl Microbiol 16, 732–734.

[20] Kalb, S.R., Baudys, J., Egan, C., Smith, T.J., Smith, L.A., Pirkle, J.L. and Barr, J.R.(2011) Different substrate recognition requirements for cleavage ofsynaptobrevin-2 by Clostridium baratii and Clostridium botulinum type Fneurotoxins. Appl Environ Microbiol 77, 1301–1308.

[21] Sikorra, S., Henke, T., Galli, T. and Binz, T. (2008) Substrate recognitionmechanism of VAMP/synaptobrevin-cleaving clostridial neurotoxins. J BiolChem 283, 21145–21152.

[22] Maslanka, S.E., Luquez, C., Raphael, B.H., Dykes, J.K. and Joseph, L.A. (2011)Utility of Botulinum Toxin ELISA A, B, E, F kits for clinical laboratoryinvestigations of human botulism. The Botulism Journal 2, 72–92.

[23] Kalb, S.R. et al. (2010) Extraction of BoNT/A, /B, /E, and /F with a single, highaffinity monoclonal antibody for detection of botulinum neurotoxin byEndopep-MS. PLoS One 5, e12237.

[24] Agarwal, R., Schmidt, J.J., Stafford, R.G. and Swaminathan, S. (2009) Mode ofVAMP substrate recognition and inhibition of Clostridium botulinumneurotoxin F. Nat Struct Mol Biol 16, 789–794.