Báo cáo Y học: Domain organization, folding and stability of bacteriophage T4 fibritin, a segmented coiled-coil protein docx

The analysis of DSC curves indicates that full-length fibritin has three thermal heat-absorption transitions that were reasonably assigned to the N-terminal, segmented coiled-coil, and C

Domain organization, folding and stability of bacteriophage T4

fibritin, a segmented coiled-coil protein

Sergei P Boudko1,2, Yuri Y Londer1, Andrei V Letarov1, Natalia V Sernova1, Juergen Engel2

and Vadim V Mesyanzhinov1

1

Shemyakin-Ovchinnikov Institute of Bioorganic Chemistry, Moscow, Russia;

2

Biozentrum der Universitaet Basel, Switzerland

Fibritin is a segmented coiled-coil homotrimer of the

486-residue product of phage T4 gene wac This protein

attaches to a phage particle by the N-terminal region and

forms fibrous whiskers of 530 A˚, which perform a chaperone

function during virus assembly The short C-terminal region

has a b-annulus-like structure We engineered a set of fibritin

deletion mutants sequentially truncated from the N-termini,

and the mutants were studied by differential scanning

calorimetry (DSC) and CD measurements The analysis

of DSC curves indicates that full-length fibritin exhibits

three thermal-heat-absorption peaks centred at 321 K

(DH¼ 1390 kJÆmol trimer)1), at 336 K (DH¼ 7600 kJÆmol

trimer)1), and at 345 K (DH¼ 515 kJÆmol trimer)1) These

transitions were assigned to the N-terminal, segmented

coiled-coil, and C-terminal functional domains, respectively

The coiled-coil region, containing 13 segments, melts

co-operatively as a single domain with a mean enthalpy

DHres¼ 21 kJÆmol residue)1 The ratio of DHVH/DHcalfor

the coiled-coil part of the 120-, 182-, 258- and 281-residue per

monomer mutants, truncated from the N-termini, and for full-length fibritin are 0.91, 0.88, 0.42, 0.39, and 0.13, respectively This gives an indication of the decrease of the Ôall-or-noneÕ character of the transition with increasing protein size The deletion of the 12-residue-long loop in the 120-residue fibritin increases the thermal stability of the coiled-coil region According to CD data, full-length fibritin and all the mutants truncated from the N-termini refold properly after heat denaturation In contrast, fibritin XN, which is deleted for the C-terminal domain, forms aggregates inside the cell The XN protein can be partially refolded by dilution from urea and does not refold after heat denatur-ation These results confirm that the C-terminal domain is essential for correct fibritin assembly both in vivo and in vitro and acts as a foldon

Keywords: bacteriophage; foldon; microcalorimetry; protein engineering; segmented coiled coil

Fibritin, a structural protein of bacteriophage T4 encoded

by gene wac (named for whisker’s antigen control), belongs

to a specific class of accessory proteins that act in the virus

assembly process Six fibritin molecules form the collar/

whisker complex that consists of a ring embracing the phage

neck with thin filaments (whiskers) protruding from the

collar [1] This complex is a sensing device that controls the

retraction of the long tail fibers in adverse environments and

thus prevents undesirable infection [2] The whiskers act also

as a chaperone and help the proximal and distal parts of the

long tail fibers to join correctly by increasing the effective

target sizes and thereby increasing the rates of otherwise

slow diffusion–limited bimolecular interactions [3]

The structure of fibritin was predicted from sequence and

biochemical analyses to be mainly a parallel segmented

triple-helical coiled-coil [4,5] Fibritin is a homotrimer of 486

residues per monomer and consists of three functional parts

Its predominant central region has 13 consecutive a helical coiled-coil segments linked by loops The protein is attached

to a phage particle by the N-terminal part that does not have heptad periodicity [6], and the short C-termini is essential for in vivo protein folding and trimerization [5] Functional activities of fibritin can be related to the exposure of hydrophobic patches in the coiled-coil [7] The full-length fibritin of 530 A˚ could not be crystallized, probably because of its inherent flexibility However, a set of smaller fibritin mutants was engineered and expressed in the soluble trimeric forms in an Escherichia coli system [5,8,9] The structures of the E and M fibritins, which are truncated for the last 120 and 75, respectively, C-terminal residues per monomer were solved to atomic resolution by X-ray crystallography [8,9] Three identical subunits form a trimeric parallel coiled-coil domain and a small a structural C-terminal domain The coiled-coil part of fibritin E is divided into three segments separated by short sequences called insertion loops The C-terminal domain, which consists of 30 residues from each monomer, contains a b-annulus-like structure with a hydrophobic interior Residues within the C-terminal domain make extensive hydrophobic and some polar inter–subunit interactions [8] This is consistent with the C-terminal domain being important for the correct assembly of fibritin, as shown by mutational studies ([5] and S P Boudko, unpublished results) Tight interactions between C-terminal residues of adjacent subunits counteract the latent instability that is

Correspondence to V V Mesyanzhinov, Howard Hughes Medical

Institute, Shemyakin-Ovchinnikov Institute of Bioorganic Chemistry,

Miklukho-Maklaya Street 16/10, 117997 Moscow, Russia.

Fax: + 7 095 336 6022, Tel.: + 7 095 335 5588,

E-mail: vvm@ibch.ru

Abbreviations: DSC, differential scanning calorimetry;

IPTG, isopropyl thio-b- D -galactoside.

(Received 20 July 2001, revised 6 December 2001, accepted

11 December 2001)

suggested by the structural properties of the coiled-coil

segments [8] Trimerization is likely to begin with the

formation of the C-terminal domain that acts as a folding

nucleus domain (foldon) and subsequently initiates the

assembly of the coiled coil [8,10] The interplay between the

stabilizing effect of the C-terminal domain and the labile

coiled-coil domain may be essential for the fibritin function

and for the correct functioning of many other a helix fibrous

proteins as well

In the present work, we obtained a set of fibritin mutants

sequentially truncated from the N-termini We engineered

also mutant S1 that have deleted for one loop of 12 residues

in fibritin E To characterize the thermodynamic properties,

stability, and domain organizations, we analysed these

fibritin mutants by differential scanning calorimetry (DSC)

and CD measurements The analysis of DSC curves

indicates that full-length fibritin has three thermal

heat-absorption transitions that were reasonably assigned to the

N-terminal, segmented coiled-coil, and C-terminal

func-tional domains, respectively

Full-length fibritin and all the mutants truncated from the

N-termini refold properly after heat denaturation We

designed also the XN mutant, a full-length fibritin that has

no C-terminal domain (Fig 1) that forms aggregates inside

the cell The XN protein can be partially refolded by fast

dilution from urea and does not refold after heat

denatur-ation The XN protein can be refolded by fast dilution from

urea and does not refold after heat denaturation

M A T E R I A L S A N D M E T H O D S

E coli strains and plasmids

The Top10 E coli strain (Invitrogen, USA) was used for the

selection of recombinant clones and plasmid DNA

purifi-cation Protein expression was performed in the BL21

(DE3) strain (Promega, USA) containing the T7 RNA

polymerase gene under lac UV5 control in the E coli

chromosome DNA fragments encoding truncated fibritin

mutants were cloned in the pET19b (+) and pET23d (+)

expression vectors containing the ribosome-binding site for

effective translation (Novagen, USA), that allow

transcrip-tion from the T7 RNA polymerase promoter

Design of fibritin mutants

We used previously designed expression vectors for a

full-length fibritin [8], fibritin XN [10,11], E, M [8,9], F (V V

Mesyanzhinov, unpublished results), and the S1 fibritin

[12,13] To create the B1, SM1, SM4 mutants, we

amplified the DNA fragments of interest by PCR and

introduce the NcoI and BamHI restriction sites for

subsequent cloning into plasmid vectors Cloning was

performed using the common techniques described in [14]

The S1 mutant that lacks 12 residues of the L11 loop

(residues

Asn-Gly-Thr-Asn-Pro-Asn-Gly-Ser-Thr-Val-Glu-Glu, Asn404-Glu415) was constructed on the basis of

fibritin E We have used an overlapping PCR method to

delete the DNA piece encoding this loop [13] Sequencing

was carried out by the dideoxy chain termination method

using a DNA sequencing kit/BigDye terminator cycle

sequencing ready reaction (Applied Biosystems) and an

automated DNA sequencer

Expression and purification of fibritin mutants The cell culture of the E coli BL21 (DE3) strain carrying the respective vector was grown at 37°C in 500 mL of

2· tryptone/yeast medium [14] until the density reached a

D600value of 0.6 Protein expression was induced by 1 mM IPTG with subsequent incubation for 3 h at 37°C with vigorous aeration We used a modification of the previously

Fig 1 Schematic presentations and amino-acid sequence of fibritin (a) Schematic presentation of the fibritin mutants used in this work: full-length fibritin (wac), XN, B1, SM1, SM4, E, S1, M, and F For each mutant, the range of amino-acid sequence that it comprises of the full-length fibritin sequence is given The N-terminal domain is a broad box; coil regions are narrow boxes; the loops, separating coiled-coil segments, are hexamers; the C-terminal domain (foldon) is a sphere (b) Amino-acid sequence of full-length fibritin and heptad scheme of the fibritin coiled coil part The hydrophobic residues in the

a and d positions are shown in bold The coiled-coil segments are indicated by roman (I–XIII), and the loops are marked [L1–L11] The bacteriophage T4 gene wac nucleotide sequence is deposited in the EMBL Gene Data Bank: accession number X12888 Atomic coordi-nates of fibritin E and fibritin M, deposited in PDB, are 1AA0 and 1AVY, respectively.

described method for purification of fibritin mutants [5].

The pellet from 500 mL of the E coli culture was

resuspended in 10 mL of Tris/EDTA buffer (50 mM

Tris/HCl, pH 8.0, 1 mM EDTA) and sonicated with

cooling The cell debris was removed by centrifugation at

25 000 g for 20 min To precipitate nucleic acids, 1 mL of

30% (w/v) streptomycin sulfate (Sigma, USA) solution in

Tris/EDTA buffer was added; the concentrated protein

solution was kept on ice for 15 min After centrifugation,

ammonium sulfate was added to the supernatant to a final

concentration of 20–50% saturation, depending on the

particular mutant, and the mixture was incubated overnight

at 4°C Protein precipitate was collected by low-speed

centrifugation, and resuspended in 3–10 mL of Tris/EDTA

buffer Nucleic acid and protein precipitation procedures

were skipped for protein S1 After ammonium sulfate

precipitation, the protein solution was applied to a 10-mL

hydroxyapatite column (Bio-Rad; DNA grade) equilibrated

with 10 mM Na phosphate (pH 8.0) and washed with

10 mMNa phosphate The flow-through fractions,

contain-ing recombinant proteins, were dialysed against Tris/EDTA

buffer and stored at 4°C The E, S1 and F proteins were

additionally applied to a 15-mL DEAE–Sephacryl column

and eluted with a linear gradient of NaCl Fractions

containing proteins were dialysed against Tris/EDTA buffer

and stored at 4°C

The protein purity was judged by denaturing SDS/PAGE

using two systems: for proteins with Mrlarger than 12 kDa

we used the Laemmli system [15]; for smaller ones we

applied the Schaegger and Jagow system [16] Protein

concentration was determined by measuring the absorbency

at 280 nm in 6MGdnHCl, and the extinction coefficient

was calculated as described in [17] For the DSC procedure

the proteins were dialysed against NaCl/Pi [10 mM Na

phosphate (pH 8.0), 150 mMNaCl or 10 mMNa phosphate

(pH 8.0)], centrifuged at 10 000 g for 30 min, and degassed

for 5 min

Purification and refolding of the XN fibritin

The pellet from 500 mL of the E coli cells expressing fibritin

XN was suspended in 10 mL of Tris/EDTA buffer (50 mM

Tris/HCl (pH 8.0), 1 mM EDTA) and sonicated under

cooling The cell extract was centrifuged at 3500 g for

30 min and supernatant was removed The pellet was

resuspended in 0.5 mL of 8M urea for 10 min and the

suspension was centrifuged at 10 000 g for 30 min to

remove insoluble particles The supernatant was mixed with

50 mL of the refolding buffer (50 mM Tris/HCl, pH 8.0,

2 mM EDTA, 2 mM phenylmethanesulfonyl fluoride),

incubated at 4°C for 3–4 days and then concentrated to

2 mL The protein solution was further purified on the

hydroxyapatite column as described above The yield of the

soluble protein was 15% of initial concentration

indicat-ing weak refoldindicat-ing

DSC

Calorimetric measurements were performed using a

VP-DSC Microcalorimeter (Microcal Inc.) equipped with

a cell (covered with Tantaloy 61TM) of 0.5 mL volume at

a heating rate of 1 KÆmin)1 Baseline subtraction,

calcu-lation of DH for different peaks and determination of

absolute heat capacity were performed using the MicroCal ORIGIN5.0 program To determine absolute heat capacity

of proteins, we used the following parameters in the equation:

DCp¼ g0qðtÞV0ð1 þ 0:00002tÞ CAbs

p ðtÞ ÿ vð1 þ atÞCW

pðtÞ

where DCp is the sample-buffer baseline minus the buffer-buffer baseline, g0is the concentration of protein (gÆmL)1), q(t) is the relative density of water (stored in the ORIGIN program [18]), V0is the nominal volume (0.5194 mL) of the sample cell, t is temperature in°C, CAbs

p (t) is the absolute heat capacity (calÆdeg)1Æg)1) of the protein in solution, v is the partial specific volume of the protein (0.717 mLÆmg)1),

a is the coefficient of thermal expansion of the protein (0.0007 1/a°C), and CW

p(t) is the unit-volume heat capacity

of water (calÆdeg)1ÆmL)1) (stored in Origin) The thermal coefficient of cubic expansion of tantalum is 0.00002 The values of the van’t Hoff enthalpy of the process for the peaks representing the melting of coiled coil region were calculated as for a first order reaction [19]:

D10Hvh ¼ 4RT

2 maxðhDCpimaxÿ D10Cp= Þ

D10Hcal where D10Hvhis the van’t Hoff enthalpy for transition from state 0 to state 1, D10Hcalis the calorimetric enthalpy, Tmaxis the temperature of the maximum heat capacity,ÆDCpæmaxis the excess heat capacity of proteins in the maximum of the peak, and D10Cpis the difference between heat capacities for state 1 and 0 (after and before the transition)

CD measurements

CD spectra of mutant proteins were recorded with an Aviv 62DS circular dichroism spectrometer (Aviv Inc., USA), equipped with a thermostatic quartz cell having a 1-mm path length CD data were analysed using the CONTIN program [20]

R E S U L T S

Engineering and properties of fibritin deletion mutants

To investigate the stability and thermodynamic properties

of T4 fibritin, a set of recombinant truncated mutants was designed and analysed All these molecules contained an intact C-terminal part and had different numbers of coiled-coil segments and separating segments loops (Fig 1 and Table 1) Fibritin S1, based on fibritin E with 120 resides per chain, had a deleted loop L11 of 12 residues, and fibritin XN had no C-terminal region of 30 residues

To enhance the protein stability, five mutations were introduced into the 74 residues of fibritin M that forms the last coiled-coil segment (5,5 heptad repeats) and the complete C-terminal domain [9] Particularly, the Ser421 residue was substituted for Lys to test the possible formation of interchain salt bridge with Glu426 The substitutions Asn428 to Asp and Thr433 to Arg were designed to create a similar interchain salt bridge between these two residues Residue 425, an Asp in a d position, was replaced by an Ile, which is generally a favourable residue in this position for a trimeric coiled coil [21]

The crystal structure of two fibritin truncated mutants,

E and M, that have 120 and 74 residues per monomer,

respectively, have been determined to atomic resolution [8]

X-ray crystallography confirmed that both mutants are

trimeric, parallel, coiled coils with a small C-terminal

domain that has a b-annulus structure In addition, we

were able to obtain crystals of fibritin B1, that has 281

residues per monomer Crystals belong to space group P21,

and existence of threefold noncrystallographic symmetry

pattern in observed X-ray diffraction data indicates that the

B1 protein is a trimer too (N V Sernova, unpublished

results) These data and the repetitive segmented structure

of fibritin suggest that other fibritin mutants studied that

have b-annulus C-terminal domain mentioned above also

should have a parallel trimeric coiled-coil structure

Indeed, all these recombinant mutants, except fibritin

XN, expressed from the plasmids in E coli cells were

soluble and proteins were purified by ammonium sulfate

precipitation followed by chromatography on

hydroxy-apatite Fibritin XN was refolded from inclusion bodies as

described in Materials and methods It is known that

full-length fibritin, as well as some N-terminally truncated

mutants, are resistant to 1% SDS [5,10,13] These proteins

do not dissociate to the monomer chains in the presence of

SDS at room temperature, and they migrate on SDS/PAGE

as trimers All the mutants used in this research have such a

resistance to SDS again except fibritin XN (data not

shown)

Figure 2 shows the CD spectra of the purified fibritin

mutants These spectra indicate that all mutants, except the

shortest fibritin F, exhibited properties characteristic of a

high content of a helicity The a helical contents slightly

decreased with decreasing size of the mutants The mean

residue ellipticity at 220 nm was )32 800 degÆcm2Ædmol)1

for full-length fibritin and)25 800 and )21 900 degÆcm2Æ

dmol)1 for fibritin B1 and fibritin SM4, respectively

Interestingly, fibritin M exhibited more a helicity than

fibritin E, probably due to the absence of insertion loops

The CD spectrum of fibritin F represented mostly the

secondary structure of the C-terminal domain, which is in a

good agreement with published data [22]

Assignment of the fibritin thermal transitions

to functional domains

The full-length fibritin, and the N-terminally truncated B1,

SM1, SM4, E, M, and F mutants were analyzed by DSC

The DSC data were also collected for fibritin XN that had

no C-terminal domain Our goal was to answer a question about how many thermodynamically independent domains fibritin has, and to assign the thermal transitions to individual functional regions Measurements were per-formed in 10 mM Na phosphate buffer, pH 8.0 with 0.15M NaCl In these conditions, the endotherm for a full-length fibritin exhibited three well-resolved heat-absorption peaks centred at 321 K (DH¼ 1390 kJÆmol trimer)1), 336 K (DH¼ 7600 kJÆmol trimer)1), and 345 K (DH¼ 515 kJÆmol trimer)1), respectively (Fig 3A) The transition at 321 K can be assigned to the N-terminal region (residues 1–50), which has no heptad periodicity, and most probably to the first adjacent downstream putative coiled-coil segment (residues 51–83) and the large loop L1 (residues 84–96) (Fig 1B) All the fibritin mutants, of different length, truncated from the N-termini had no corresponding peak Additionally, fibritin XN, that con-tained the N-terminus, had a heat absorption peak at 321 K

of the same enthalpy as wild-type fibritin (see below) The transition at 345 K was clearly related to the C-terminal domain The DSC endotherm showed that all truncated fibritin molecules, containing the C-terminal domain, had the heat absorption peak (Fig 3A,B) Its enthalpy was approximately equal for all studied fibritin mutants (Fig 3A, internal) as well as for the isolated C-termini [22] The highest transition temperature of the different oligomeric protein domains was usually

Table 1 Thermodynamic properties of fibritin truncated mutants.

No of amino-acid

residues

DH cal of all transitions (total) (JÆmol)1)

DH cal coiled-coil transition (JÆmol)1)

DH cal folding nucleus (JÆmol)1)

DH vh /DH cal coiled-coil transition

Fig 2 Far CD spectra of wac, B1, SM1, E, M and F fibritins.

concentration dependent [22] Indeed, the 345 K transition

of fibritin was concentration dependent (data not shown) as

was found for the isolated C-termini [22]

In addition, the CD spectrum of fibritin SM4 indicated

that the secondary structure of the C-terminal domain melts

between 335 and 358 K (Fig 4A) The DSC endotherms for

B1, SM1, and SM4 mutants (all containing the C-terminal

domain) revealed that the 330 K heat adsorption transition

was almost accomplished at 335 K, while the 345 K

transition was just beginning According to the CD data,

the SM4 protein was completely unfolded at 358 K The

CD spectrum of fibritin’s C-terminal domain was calculated

as the difference of spectra at 335 K and 358 K It had a

characteristic positive peak centered at 229 nm with molar

ellipticity hmolar¼ 12 000 degÆcm2Ædmol)1 (Fig 4B) that

was in agreement with the CD spectrum of the purified

C-terminal domain [22]

The major heat absorption peak at 336 K, observed for a

full-length fibritin, had an enthalpy that was four times

larger than the other two transitions at 321 K and 345 K,

and it definitely can be assigned to the coiled-coil part The

occurrence of only a single transition strongly supports

co-operative heat-induced unfolding of all coiled coil segments Unfolding of the coiled coil of fibritin XN gave two heat absorption peaks centred at 330 K and at 336 K (see below) The appearance of the 330 K transition can be explained by the structure destabilization at the C-terminus due to the elimination of 30 last residues

Besides the 345 K peak, fibritin B1, which consisted about half of a full-length molecule (Fig 1), as well as shorter SM1 and SM4 mutants all had another heat absorption peak with a midpoint at 330 K (Fig 3A) However, for fibritin E this peak was centred at 320 K, and the smallest fibritin M and F showed no separation of melting between the C-terminal domain and the coiled-coil region (Fig 3A) Significant stabilization of fibritin M, in comparison with a wild-type fibritin, can be explained mainly by two residues substitutions As confirmed by X-ray crystallography [9], the mutation Ser421 to Lys created a new salt bridge between residues Lys421 and Glu426 These residues occupy the g and e heptad’s positions in different chains within fibritin M trimer It is known that interchain salt bridges have a stabilizing effect

on the coiled coil [23] Another mutation, Asn425 to Ile,

Fig 4 The calorimetric enthalpy plots for the full-length fibritin (wac), B1, SM1, SM4, and F proteins in 0.01 M Na phosphate buffer (pH 8.0) and 0.15 NaCl The enthalpy assigned to the coiled-coil part represent a linear dependence with the slope of 21 kJÆmol res)1.

Fig 3 Temperature dependence of the partial heat capacity of fibritin mutants in 0.01 M Na phosphate buffer (pH 8.0) and 0.15 M NaCl Protein concentration was 16 m M chain)1for the full-length fibritin, and 50 m M chain)1for the others (a) Thermal transition profiles of the wac, B1, SM1, SM4, M, and F mutants (b) Thermal transition curves for the E, S1, and F fibritins.

eliminates an unusual interaction between the Asp in a d

position that is mediated in fibritin E by a chloride ion

located on the threefold axis [8] This interaction, also found

in other coiled-coil proteins, is considered to be important

for the correct alignment of polypeptide chains upon a

coiled-coil formation [23,24] However, in fibritin, its

C-terminal domain governs such an assembly alignment

Furthermore, Ile425 is well accommodated at its d position

in the trimeric coiled-coil structure [9], and this mutation

also seems to increase the stability of fibritin M

The DHcalvalues of the 336 K peak of full-length fibritin,

and of the 330 K peaks of the B1, SM1, SM4, and E

truncated molecules were proportional to their size (Fig 5)

The mean enthalpy, calculated from the slope of the graph,

was DHres¼ 21 kJÆmol residue)1 The singularity and

pro-portionality of that transition are consistent with the

thermal unfolding of a uniform domain By varying the

ionic strength of the sample buffer, no discrete melting of

subdomains was found for the short coiled-coil segments

(data not shown)

The melting temperature of the coiled-coil region of the

B1, SM1, SM4 (Tm¼ 330 K), and E (Tm¼ 320 K)

mutants was lower than that for the respective part of a

wild-type fibritin (Tm¼ 336 K) This was an indication that

the deletion of the N-terminal sequence of fibritin had a

destabilizing influence The ratio of DHVH andDHcalfor the

E, SM4, SM1, B1 mutants, and for a full-length fibritin were

0.91, 0.88, 0.42, 0.39 and 0.13, respectively (Table 1),

indicating a decrease of the all-or-none transition character

with increasing domain size A plot of total DHcalagainst

the number of residues for all mutants, truncated from the

N-termini, yielded a homogeneous curve with an initial slope of 6.5 ± 0.5 and a final slope of 27.5 ± 2 kJÆ(mol residue))1(Fig 5)

Preliminary results indicate that at low ionic strength (10 mMsodium phosphate buffer, pH 8.0) full-length fibr-itin exhibited two heat absorption peaks (T1m ¼ 326 K, and T2m ¼ 334 K) that are probably related to the transition of the coiled-coil region The position of the

326 K peak approximately matched the position of a single transition peak of the B1, SM1, and SM4 mutants (Tm ¼ 327–328 K) (data not shown) At the present, by varying pH and ionic strength conditions, we are trying to detect subdomain transitions of the coiled-coil region Stability of the S1 fibritin

Three coiled coil segments of fibritin E are separated by two loops: residues Gly386–Gly391 form the first one (L10) and the second one (L11) contains the residues Asn404–Gly417 [9] (Fig 1) To clarify the role of the loop regions in protein stability, we designed fibritin S1 lacking the Asn-Gly-Thr-Asn-Pro-Asn-Gly-Ser-Thr-Val-Glu-Glu sequence of loop L11 [13] The two last L11 loop residues, Arg and Gly, were preserved in S1 to made the coiled coil continuous (Fig 1B)

The calorimetric transitions for the coiled-coil regions of the E and S1 mutants differed by 10 K (Fig 3B) The coiled-coil part, which lacked the loop sequence, melted at

330 K while fibritin E had a transition at 320 K The enthalpy of this transition was DHcal ¼ 656 kJÆ(mol trimer))1 for fibritin S1 and 687 kJÆ(mol trimer))1 for fibritin E Most probably, the stability of S1 increased due

to the formation of uniform coiled coil containing two segments, XI and XII Also, elimination of loop 11 might have helped to form of additional salt bridge between residues Glu435 and Lys440, at the g and e positions, respectively That bridge was initially proposed [5], but it was not found in fibritin E crystal structure [8] Crystallo-graphic investigations of fibritin S1 structure are in progress Refolding of the XN fibritin

Due to aberrant folding, fibritin XN, lacking the C-terminal domain, was not soluble during in vivo expression and it formed aggregates [10] We were able to purify and dissolve these aggregates in 8Murea Then the protein was partially refolded by the fast 100-fold dilution from 8Mto 0.08M urea in 50 mMTris/HCl buffer, pH 8.0 and purified on a hydroxyapatite column The CD spectrum of an in vitro refolded fibritin XN was similar to the spectrum of a full-length fibritin (data not shown) However, the DSC endotherm of the refolded XN fibritin did not reveal a heat-adsorption 345 K-peak characteristic for the C-termi-nal domain, and the protein had three thermal transition peaks centred at 321 K, 329 K, and 336 K (Fig 6A) The main difference between fibritin XN and other truncated fibritin molecules, which contained the C-terminal domain, was lack of ability of the XN molecule to refold after temperature-induced denaturation After one round of heating to 340 K and subsequent slow cooling to 293 K for

60 min, the protein revealed a complete lack of refolding (Fig 6A) In contrast, all fibritin mutants containing the C-terminal domain exhibited reversible refolding under the

Fig 5 Far CD spectra for the SM4, and F proteins and folding nucleus

alone in a solution of 0.01 M Na phosphate buffer (pH 8.0) and 0.15 M

NaCl (a) Spectra of the SM4 fibritin (182 residues per monomer)

were registered at 298, 335, and 358 K The protein has the native

conformation at 298 K, and is completely unfolded at 358 K The

335 K spectrum is the spectrum of the partially unfolded state in

which the coiled-coil part is disordered and the folding nucleus

domain still has its secondary structure This may be seen at 229 nm:

the 335 K spectrum has a more positive h-value than the 358 K

spectrum The difference of the signals for these two spectra assigned

only for the folding nucleus (30 residues) is presented in (b) in

comparison with the isolated the C-terminal part spectra [22] The

C-termini peak, centred at 229 nm, can easily be detected also for

fragment F that has only 58 residues per monomer (a).

same conditions As an example, Fig 6B shows the results

of heat denaturation of fibritin B1 After heating to 336 K,

the transition curves for second and third rounds differed

from the first one by only a few percent The differences

were even smaller for shorter fibritin fragments Significant

flattening of the peaks corresponding to the coiled-coil

region was observed only after heating to 369 K (see

Fig 6B, for fibritin B1) Prolonged heating led to a further

decrease of the extent of refolding Independent of

temper-ature and time of heat exposure, refolding of the C-terminal

domain was completely reversible as indicated by identical

DH°-values, sharpness and height of the 345 K peak

D I S C U S S I O N

Previous work has demonstrated that a full-length fibritin

has a complex pattern of heat-induced transitions [5] that

were difficult to assign to individual domains Also it was

not possible to determine calorimetric parameters for the

individual steps in transition curve and to investigate the

interactions between individual segments in the

three-stranded coiled-coil domain A more detailed analysis was

performed now with the help of truncated fibritin molecules

The C-terminal domain has the highest melting

temper-ature and it melts independently from all the other regions

Due to its trimeric nature, the midpoint temperature of the

C-terminal domain transition is slightly concentration

dependent, an observation which is in agreement with the

results for purified domain [22] It acts as a cross-linker

between the three chains and, as it was proposed earlier

[5,8,10], it helps to align three chains and serves as a foldon

by increasing local chain concentration at the C-terminus

In addition, the C-terminal domain of fibritin, like other

oligomerization domains [25,26], stabilizes adjacent

upstream coiled-coil segments

For the coiled-coil region of fibritin B1, which contains

about half of a fibritin sequence, only a single transition was

observed The assignment of the 330 K transition is evident

from the loss of a helicity at this temperature and changes in

the magnitude of the accompanying enthalpy The ratio of the van’t Hoff enthalpy to calorimetric enthalpy of 0.39 indicates that the nine putative segments of the coiled-coil domain of fibritin B1 do not unfold in an all-or-none manner ÔNon all-or-none transitionÕ means that we do have intermediates, but in the case of fibritin and other fibrous proteins these intermediates do not have fixed structures because these proteins have a zipper-like mechanism of folding-unfolding [27] Nevertheless, the sharpness of the transition and the failure to detect a splitting of the transition profile into individual subpeaks suggests that loop regions, connecting B1 coiled-coil segments, serve as co-operative linkers between the segments According to equilibrium criteria, the unfolding and reversible refolding

of the nine segments therefore occurs in a single step The singularity of the coiled-coil transition, midpoint temperature and peak sharpness are maintained also for the SM1 and SM4 fibritins in which the number of coiled-coil segments is reduced to eight and five, respectively The all-or-none approximation is better fulfilled for these proteins, which is expected for their smaller size and more limited contacts Interestingly, the enthalpy of the transition for the

E, SM4, SM1 and B1 fibritins increases linearly with an increasing number of amino-acid residues in the coiled-coil region In contrast to the independent melting of the coiled-coil segments of different stability, this is additional evidence for the co-operative transition of the entire coiled-coil region The ratio of the van’t Hoff enthalpy to calorimetric enthalpy for fibritin E is 0.91, is very close to 1 for the all-or-none approximation This finding, which is in accordance with the crystallographic observation [8] that two coiled-coil segments of fibritin E is a repetitive structured domain with loop regions as a part of the structure The enthalpy change per residue in the coiled-coil domain of all the fibritin mutants (DHres¼)21 JÆmol)1) has the same magnitude as for a three-stranded coiled-coil domain of laminin [28], and for a two-stranded coiled coil of leucine zippers [29,30] According to CD data, we were able to refold fibritin

XN, which was solved in urea, by rapid dilution During the

Fig 6 Consequent DSC scans performed for the XN and B1 fibritin mutants in 0.01 M Na phosphate buffer (pH 8.0) with 0.15 M NaCl with a scan rate

of 1 KÆmin)1 The absolute heat capacity vs the temperature is shown (a) The XN fibritin scans: the first is of the folded fragment, the second is after treating the fragment at 340 K for 5 min and cooling down to room temperature for more than 1 h (b) Consequent scans of the B1 fragment (without refilling the cells): the first two scans were performed until 336 K followed by cooling down to 298 K for 1 h; the others scans were performed until 369 K.

first round of DSC, the refolded XN protein exhibits several

heat absorption peaks, one of which was assigned to the

N-terminal domain Following the first round of heat

denaturation, it was impossible to refold of the molecule by

slow cooling to low temperature In contrast, full

revers-ibility has been observed for all fragments containing the

C-terminal domain These results strongly confirm our

previous conclusion [8,10] that the C-termini is essential for

fibritin assembly in vivo and in vitro and act as a foldon

Foldon is a protein unit that forms on the initial steps of

folding [31,32] which frequently perform a specific, distinct

function that remains intact even after isolated or

trans-ferred into other proteins [22,33–35] The stabilizing and

assembly of the trimeric T4 fibritin foldon has been

demonstrated recently by protein engineering for several

chimera proteins [22,36,37]

A C K N O W L E D G E M E N T S

We thank Dr Kyle Tanner for critical reading of the manuscript,

and Dr Sergei Yu Venyaminov for providing the CONTIN

program This work was supported in part by HHMI (grants

75195–52080, and 55000324), Russian Foundation for Basic

Research (grant 99-04-48430), and by the ƠUniversities of RussiaÕ

grant to V V M, and by Swiss National Science Foundation (grant

31-49281.96) to J E.

R E F E R E N C E S

1 Coombs, D.H & Eiserling, F.A (1977) Studies on the structure,

protein composition and assembly of the neck of bacteriophage

T4 J Mol Biol 116, 375–405.

2 Conley, M.P & Wood, W.B (1975) Bacteriophage T4 whiskers: a

rudimentary environment-sensing device Proc Natl Acad Sci.

USA 72, 3701–3705.

3 Terzaghi, B.E., Terzaghi, E & Coombs, D (1979) The role of the

collar/whisker complex in bacteriophage T4D tail fiber

attach-ment J Mol Biol 127, 1–14.

4 Sobolev, B.N & Mesyanzhinov, V.V (1991) The wac gene

product of bacteriophage T4 contains coiled-coil structural

patterns J Biomol Struct Dyn 8, 953–965.

5 Efimov, V.P., Nepluev, I.V., Sobolev, B.N., Zurabishvili, T.G.,

Schulthess, T., Lustig, A., Engel, J., Haener, M., Aebi, U.,

Venyaminov, S., Yu, Potekhin, S.A & Mesyanzhinov, V.V (1994)

Fibritin encoded by bacteriophage T4 gene wac has a parallel

triple-stranded alpha-helical coiled-coil structure J Mol Biol.

242, 470–486.

6 Cohen, C & Parry, D.A (1994) Alpha-helical coiled coils: more

facts and better predictions Science 263, 488–489.

7 Siegert, R., Leroux, M.R., Scheufler, C., Hartl, F.U & Moarefi, I.

(2000) Structure of the molecular chaperone prefoldin Unique

interaction of multiple coiled coil tentacles with unfolded proteins.

Cell 103, 621–632.

8 Tao, Y., Strelkov, S.V., Mesyanzhinov, V.V & Rossmann, M.G.

(1997) Structure of bacteriophage T4 fibritin: a segmented coiled

coil and the role of the C-terminal domain Structure 5, 789–798.

9 Strelkov, S.V., Tao, Y., Shneider, M.M., Mesyanzhinov, V.V &

Rossmann, M.G (1998) Structure of bacteriophage T4 fibritin M:

a troublesome packing arrangement Acta Crystallogr D54,

805–816.

10 Letarov, A.V., Londer, Y.Y., Boudko, S.P & Mesyanzhinov,

V.V (1999) The carboxy-terminal domain initiates trimerization

of bacteriophage T4 fibritin Biochemistry (Moscow) 64, 817–823.

11 Letarov, A.V (1999) Examination of the Folding Pathways of the

Bacteriophage T4 Fibritin PhD Thesis Ivanovsky Institute of

Virology, Academy of Medical Sciences, Moscow, Russia.

12 Londer, Y.Y (1999) Folding, Assembly, and Stability of Bacte-riophage T4 Fibritin PhD Thesis Bakh Institute of Biochemistry, Russian Academy of Sciences, Moscow, Russia.

13 Londer, Y & Mesyanzhinov, V.V (1999) Thermostability of bacteriophage T4 fibritin and its deletion mutants Bioorg Khim.

25, 257–263.

14 Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., Smith, J.A & Struhl, K (1992) Short Protocols in Molecular Biology John Wiley and Sons, Toronto, Canada.

15 Laemmli, U.K (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4 Nature 227, 680–685.

16 Schagger, H & Jagow, G (1987) Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation

of proteins in the range from 1 to 100 kDa Anal Biochem 166, 368–379.

17 Edelhoch, H (1967) Spectroscopic determination of tryptophan and tyrosine in proteins Biochemistry 6, 1948–1954.

18 MicroCal Inc (1998) DSC Data Analysis in OriginỊ Tutorial guide MicroCal Inc., Northampton, MA, USA.

19 Privalov, P.L & Potekhin, S.A (1986) Scanning microcalorimetry

in studying temperature-induced changes in proteins Methods Enzymol 131, 4–51.

20 Provencher, S.W & Glockner, J (1981) Estimation of globular protein secondary structure from circular dichroism Biochemistry

20, 33–37.

21 Harbury, P.B., Kim, P.S & Alber, T (1994) Crystal structure of

an isoleucine-zipper trimer Nature 371, 80–83.

22 Frank, S., Kammerer, R.A., Mechling, D., Schulthess, T., Land-wehr, R., Bann, J., Guo, Y., Lustig, A., Baechinger, H & Engel, J (2001) Stabilisation of short collagen-like triple helices by protein engineering J Mol Biol 308, 1081–1089.

23 Lumb, K.J & Kim, P.S (1995) Measurement of interhelical electrostatic interactions in the GCN4 leucine zipper Science 268, 436–439.

24 Lumb, K.J & Kim, P.S (1995) A buried polar interaction imparts structural uniqueness in a designed heterodimeric coiled coil Biochemistry 34, 8642–8648.

25 Engel, J & Kammerer, R.A (2000) What are oligomerization domains good for? Matrix Biol 19, 283–288.

26 Cohen, C & Parry, D.A (1998) A conserved C-terminal assembly region in paramyosin and myosin rods J Struct Biol 122, 180–187.

27 Engel, J & Schwarz, G (1970) Co-operative conformational transitions of linear biopolymers Angew Chem Int 9, 389–400.

28 Kammerer, R.A., Antonsson, P., Schulthess, T., Fauser, C & Engel, J (1995) Selective chain recognition in the C-terminal alpha-helical coiled-coil region of laminin J Mol Biol 250, 64–73.

29 Bosshard, H.R., Durr, E., Hitz, T & Jelesarov, I (2001) Energetics of coiled coil folding: the nature of the transition states Biochemistry 40, 3544–3552.

30 Yu, M.-H & King, J (1984) Single amino acid substitutions influencing the folding pathway of the phage P22 tailspike endorhamnosidase Proc Natl Acad Sci USA 81, 6584–6588.

31 Panchenko, A.R., Luthey-Schulten, Z & Wolynes, P.G (1996) Foldons, protein structural modules and exons Proc Natl Acad Sci USA 93, 2008–2013.

32 Inaba, K., Kobayashi, N & Fersht, A.R (2000) Conversion of two-state to multi-state folding on fusion of two protein foldons.

J Mol Biol 302, 219–233.

33 Yanagawa, H., Yoshida, K., Torigoe, C., Prak, J.S., Sato, K., Shirai, T & Go, M (1993) Protein anatomy; functional roles of barnase module J Biol Chem 268, 5861–5865.

34 Wakasugi, K., Isimori, K., Imai, K., Wada, Y & Morishima, I (1994) ƠModuleÕ substitution in hemoglobin subunits J Biol Chem 269, 18750–18756.

35 Miroshnikov, K.A., Sernova, N.V., Shneider, M.M &

Mesyanzhinov, V.V (2000) Transformation of a fragment of

beta-structural bacteriophage T4 adhesin to stable alpha-helical

trimer Biochemistry (Moscow) 65, 1346–1351.

36 Krasnykh, V., Belousova, N., Korokhov, N., Mikheeva, G & Curiel, D.T (2001) Genetic targeting of an adenovirus vector via replacement of the fiber protein with the phage T4 fibritin J Virol.

75, 4176–4183.