Structure of bacteriophage T4 fibritin: a segmented coiled coil and the role of the C-terminal domain.

Abstract

Oligomeric coiled-coil motifs are found in numerous protein structures; among them is fibritin, a structural protein of bacteriophage T4, which belongs to a class of chaperones that catalyze a specific phage-assembly process. Fibritin promotes the assembly of the long tail fibers and their subsequent attachment to the tail baseplate; it is also a sensing device that controls the retraction of the long tail fibers in adverse environments and, thus, prevents infection. The structure of fibritin had been predicted from sequence and biochemical analyses to be mainly a triple-helical coiled coil. The determination of its structure at atomic resolution was expected to give insights into the assembly process and biological function of fibritin, and the properties of modified coiled-coil structures in general.
The three-dimensional structure of fibritin E, a deletion mutant of wild-type fibritin, was determined to 2.2 A resolution by X-ray crystallography. Three identical subunits of 119 amino acid residues form a trimeric parallel coiled-coil domain and a small globular C-terminal domain about a crystallographic threefold axis. The coiled-coil domain is divided into three segments that are separated by insertion loops. The C-terminal domain, which consists of 30 residues from each subunit, contains a beta-propeller-like structure with a hydrophobic interior.
The residues within the C-terminal domain make extensive hydrophobic and some polar intersubunit interactions. This is consistent with the C-terminal domain being important for the correct assembly of fibritin, as shown earlier by mutational studies. Tight interactions between the C-terminal residues of adjacent subunits counteract the latent instability that is suggested by the structural properties of the coiled-coil segments. Trimerization is likely to begin with the formation of the C-terminal domain which subsequently initiates the assembly of the coiled coil. 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 alpha-fibrous proteins.

Full text

Page 1

Structure of bacteriophage T4 fibritin: a segmented coiled coil
and the role of the C-terminal domain
Yizhi Tao1, Sergei V Strelkov1,2, Vadim V Mesyanzhinov2 and
Michael G Rossmann1*
Background: Oligomeric coiled-coil motifs are found in numerous protein
structures; among them is fibritin, a structural protein of bacteriophage T4,
which belongs to a class of chaperones that catalyze a specific phage-assembly
process. Fibritin promotes the assembly of the long tail fibers and their
subsequent attachment to the tail baseplate; it is also a sensing device that
controls the retraction of the long tail fibers in adverse environments and, thus,
prevents infection. The structure of fibritin had been predicted from sequence
and biochemical analyses to be mainly a triple-helical coiled coil. The
determination of its structure at atomic resolution was expected to give insights
into the assembly process and biological function of fibritin, and the properties
of modified coiled-coil structures in general.
Results: The three-dimensional structure of fibritin E, a deletion mutant of wild-
type fibritin, was determined to 2.2Å resolution by X-ray crystallography. Three
identical subunits of 119 amino acid residues form a trimeric parallel coiled-coil
domain and a small globular C-terminal domain about a crystallographic
threefold axis. The coiled-coil domain is divided into three segments that are
separated by insertion loops. The C-terminal domain, which consists of 30
residues from each subunit, contains a ?-propeller-like structure with a
hydrophobic interior.
Conclusions: The residues within the C-terminal domain make extensive
hydrophobic and some polar intersubunit interactions. This is consistent with
the C-terminal domain being important for the correct assembly of fibritin, as
shown earlier by mutational studies. Tight interactions between the C-terminal
residues of adjacent subunits counteract the latent instability that is suggested
by the structural properties of the coiled-coil segments. Trimerization is likely to
begin with the formation of the C-terminal domain which subsequently initiates
the assembly of the coiled coil. 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 ?-fibrous proteins.
Introduction
Phage T4 is a bacterial virus that consists of a prolate head,
containing double-stranded DNA, and a sixfold-symmet-
ric tail, through which the viral DNA is extruded during
infection [1] (Fig. 1). In addition, there is a set of six long
and six short tail fibers, which are responsible for receptor
recognition and initialization of the infectious process.
The fibers are connected to a complex substructure, called
the baseplate, located at the distal end of the tail.
Fibritin is a homotrimer of the protein gpwac, a 52kDa
product of the gene wac (whisker antigen control) that is
expressed late in the virus life cycle. After the phage head
is joined to a fiberless tail, six fibritin molecules attach to
the neck of the virion and form the collar with six fibers
(‘whiskers’). Interaction of the whiskers with the distal
and proximal parts of the long tail fibers stimulates assem-
bly of the long tail fibers and their subsequent attachment
to the tail baseplate [2,3]. Viruses without or with fewer
than six long tail fibers are either non- or less infectious
[4]. Additional to their role in virus assembly, the whiskers
are a rudimentary environment-sensing device. The
whiskers control the retraction of the long tail fibers in
response to certain adverse environments, therefore pre-
venting infection [5]. It is unclear whether long tail fiber
assembly and retraction involve the same interaction of
the whiskers with the fibers.
Recombinant gpwac protein, expressed in Escherichia coli,
assembles into filamentous molecules with a length of
530Å. Sequence analysis of the gpwac protein has pred-
icted that there are three major structural regions — the
Addresses: 1Department of Biological Sciences,
Purdue University, West Lafayette, Indiana 47907-
1392, USA and 2Howard Hughes Medical Institute,
Bach Institute of Biochemistry, 33 Leninsky
Prospect, 117071 Moscow, Russia.
*Corresponding author.
E-mail: mgr@indiana.bio.purdue.edu
Key words: bacteriophage T4, coiled coil, fibritin,
oligomerization, X-ray crystallography
Received: 20 March 1997
Revisions requested: 21 April 1997
Revisions received: 30 April 1997
Accepted: 1 May 1997
Structure 15 June 1997, 5:789–798
http://biomednet.com/elecref/0969212600500789
© Current Biology Ltd ISSN 0969-2126
Research Article789

Page 2

N-terminal, the central and the C-terminal domains. The
central domain, which is 80% of the entire sequence, con-
tains the heptad repeat amino acid sequence pattern
(abcdefg)n, where a and d are preferably occupied by
hydrophobic residues [6]. Such a motif is characteristic of
an ?-helical coiled coil [7]. In gpwac this periodicity is
interrupted at several places, which suggests the existence
of twelve coiled-coil segments connected by linker
regions. Deletion of the N-terminal domain abolishes
attachment of the protein to the virion neck, whereas
certain mutations in the C-terminal domain, or its partial
deletion, result in an insoluble expression product.
However, a partial deletion of the central domain or fusion
of a foreign peptide to the C terminus affects neither fib-
ritin assembly nor attachment of the whiskers to the virion
neck [8]. An antigen-display system has been developed,
in which the extended fibritin gene, with foreign genetic
material appended to the C terminus, is re-introduced into
T4 phage [9].
Our structural study of fibritin had two main objectives.
Firstly, to study fibritin’s function as a specialized chaper-
one during T4 phage assembly. Secondly, to use fibritin as
a convenient model for studying the mechanisms that
determine protein folding, oligomerization and stability of
coiled-coil proteins. As full-length fibritin could not be
crystallized, a series of mutant proteins A, B, C, D, E, F
and M were engineered that had the N-terminal domain
and progressively more of the ?-helical (central) domain
deleted. Fibritin E consists of the final 119 residues of the
wild-type fibritin (from Glu368 to Ala486), comprising the
last three putative coiled-coil segments and the C-termi-
nal domain [10]. Here, we present the X-ray structure of
fibritin E at 2.2Å resolution. It is the first structural
protein of bacteriophage T4 for which the structure has
been solved to atomic resolution, and it provides some
insights into how fibritin can act as a specialized chaper-
one during phage assembly.
Results and discussion
The fibritin E structure was determined using the multi-
ple isomorphous replacement (MIR) method (Table 1).
The final atomic model included 113 out of the 119
residues, lacking three residues at the N terminus and
three at the C terminus. A subunit of fibritin E consists of
an ?-helical region and a C-terminal region (Fig. 2).
Three identical fibritin E subunits associate into a trimer,
forming a parallel ?-helical coiled coil along the crystallo-
graphic threefold axis (Fig. 1). In the C-terminal domain,
there is a small ? sheet formed by threefold-related ?
790Structure 1997, Vol 5 No 6
Figure 1
Bacteriophage T4, with the fibritin whiskers
shown in color. Adapted from a drawing by
Eiserling and Black [1]. On the right is an
enlarged ribbon diagram of fibritin E.

Page 3

hairpins. This ? sheet is in a plane perpendicular to the
coiled-coil axis.
Coiled-coil domain
Unlike most other coiled-coil structures [11], each ? helix
in fibritin E is interrupted by two insertion loops, resulting
in three segments each containing 2, 1.5 and 5.5 heptad
repeats, respectively. The downstream helices almost
exactly follow the continuation of the preceding ones
(Fig. 3a). Two mainchain hydrogen bonds are formed
between consecutive helical segments which bring the
helices into register after omitting approximately one turn
(Fig. 3b,c). In a coiled coil, residues at the a or d positions
are arranged in alternate layers along the coiled-coil axis.
As each helical turn contains about 3.5 residues, the
residues at a and d have to be positioned at different sides
of the supercoil axis. In this arrangement, every internal
sidechain, termed a ‘knob’, is fitted into the empty space,
a ‘hole’, provided by other peptide chains in the coiled
coil [12,13]. Helical segments ?10 and ?11 end at position
a and helical segments ?11 and ?12 start at position d.
This makes the abutting helical segments consistent with
the requirements for forming a continuous coiled coil.
Four mainchain hydrogen bonds are omitted at the site of
every insertion loop, which creates greater flexibility in
comparison with an uninterrupted ? helix (Fig. 3d). This
Research Article Bacteriophage T4 fibritin Tao et al. 791
Table 1
Phasing statistics.
Phasing
power†
DerivativeRcullis*SiteLigand
UO2(CH3COO)2
0.64 1.87a
b
a
c
D449,E475
E455
D449,E475
Q467
Pb(CH3COO)2
K2PtCl4
0.72
0.90
1.56
0.77
*Rcullis= Σ||FPH–FP|–|FH(calc)||/?|FPH–FP|, where the FP, FPHand FHare
the structure amplitudes of the native, heavy atom derivative and heavy
atom compounds, respectively. †Phasing power= rms (FH/E), where E
is the lack of closure error.
Figure 2
Fibritin E secondary and tertiary structure.
(a) Assignment of secondary structure
elements to the fibritin E sequence. Rods
represent ? helices and arrows represent
? strands. The periodic pattern (abcdefg)n
denotes heptad repeats. Loops are shown by
thin lines. The numbering of the ?-helical
segments and insertion loops is derived from
the predicted structure of the wild-type fibritin
[8]. Fibritin E contains ?-helical segments ?10
(partially), ?11 and ?12, as well as loops L10
and L11. (b) Stereo diagram of the fibritin E
C? backbone, drawn with the program
MolView [42]. Each monomer is shown in a
different color.
abcdefgabcdefga------defgabcdefga--------------defgabcdefgabcde
VSGLNNAVQNLQVEIGNNSAGIKGQVVALNTLVNGTNPNGSTVEERGLTNSIKANETNIASVT
fgabcdefgabcdefgabcdefg-<<<< C-terminal domain >>>>
QEVNTAKGNISSLQGDVQALQEAGYIPEAPRDGQAYVRKDGEWVLLSTFL
α10
α11
L10L11
α12
β1
β2
434371
483435
371
383
398
411
428
438
453
466
483
371
383
398
411
428
438
453
466
483
(a)
(b)

Page 4

is consistent with the differences of curvature of fibritin
molecules viewed by electron microscopy [8] and with the
increased temperature factors of atoms in the insertion
loops of fibritin E. This pliability may permit the whiskers
to bind the long tail fibers.
There is a glycine residue near the beginning and end of
each insertion loop. The mutation of these glycines to
non-glycine residues, especially to residues with large
sidechains, would result in steric hindrance because the
mainchain dihedral angles of these four glycines are in the
regions of the Ramachandran plot that are unfavored for
other residues. These glycines might be important for
helix initiation and termination. The hydrogen bonds
between sidechains within the loop structures might
provide some stability to the loop conformation. Hence,
the glycines at the loop termini may provide some flexibil-
ity between the rigid ?-helical segments and the some-
what rigid loop structure itself.
No aromatic residues have been found at the a or d posi-
tions in parallel coiled-coil structures with strict heptad
repeats and which have been solved to atomic resolution
(for examples see Table 2). Although there are two pheny-
lalanines at a positions in the coiled-coil region of hemag-
glutinin [14], the irregularity of the heptad repeats shifts
these two a positions further out of the interior than in a
regular coiled-coil structure. For a coiled-coil structure
with exact heptad repeats, like fibritin, aromatic residues
at either the a or d position are unlikely because of their
bulky sidechains, which would result in steric hindrance.
The above considerations lead to some guidelines for the
sequence requirements in the formation of coiled-coil struc-
tures with insertion loops, such as those that occur in fib-
ritin. Firstly, a helix prior to an insertion loop ends at
position a, and the next helix starts at position c or d in order
to facilitate the helix continuation. Secondly, glycines or
small polar residues are generally found at the ends of inser-
tion loops. Thirdly, there are no aromatic residues at the a
792Structure 1997, Vol 5 No 6
Figure 3
N
L
Q
V
EIIKGQ
VV
A
NNNNNNNNNNNNN
O
OOOOOOOOOOOO
gnnsag
Loop L10
A
L
N
T
L
V
LTNSIK
A
p
n
g
s
t
NNNNNNNNNNNNN
O
OOOOOOOOOOOO
n
g
t
n
v
e
e
r
g
Loop L11
(a)(d) (b)(c)
Insertion loops in the coiled-coil region of fibritin E. (a) Superposition
of the fibritin E C? backbone (blue) onto an ideal structure (red), which
is calculated from parameters defining helix ?12. (b,c) The
arrangement of hydrophobic sidechains and hydrogen bonds at the
insertion loops L10 and L11 (parts (b) and (c) are rotated by 50° and
100° about the coiled-coil axis relative to (a), respectively). ? helices
are in green and the insertion loops are in white. (d) The pattern of
hydrogen bonds across each insertion loop. Hydrogen bonds are
represented by black solid lines, whereas missing hydrogen bonds are
in grey. Part (a) was drawn with the program MolView [42]; parts (b)
and (c) were drawn with MOLSCRIPT [43] and RASTER3D [44].
Table 2
Comparison of various coiled-coil structures.
Supercoil
Oligomerization Protein Radius(Å)Pitch(Å)
Dimer
Trimer
Average*
Average†
Hemagglutinin‡
Fibritin E§
GCN4 pLI#
COMP**
5.0
6.6
6.7
6.4
7.6
8.6
128.5
175.5
335.1
137.8
200.5
204.0
Tetramer
Pentamer
*Average of GCN4 leucine zipper p1 [15], catabolite gene activator
protein [45] and GAL4 yeast transcription factor [46]. †Average of
GCN4 leucine zipper pII [13], mannose-binding protein (rat) [47] and
mannose-binding protein (human) [20]. ‡Hemagglutinin [14] has
occasional additional inserted residues that interrupt the regular repeat
pattern. §Average of fibritin E segments ?10, ?11 and ?12. #GCN4
leucine zipper pLI [16]. **Cartilage oligomeric matrix protein (rat) [48].

Page 5

and d positions in regular parallel coiled-coil structures.
These guidelines have now been applied to the complete
wild-type fibritin sequence, leading to small revisions in the
secondary structure predictions of Efimov et al. [8] (Fig. 4).
The differences to the earlier prediction are in the position-
ing of the N termini of ?-helical segments and the introduc-
tion of new loops where there are aromatic residues.
Simple rules have been proposed that predict whether
the spontaneous oligomerization of coiled coils forms
dimers, trimers or tetramers. These are based on the
mutant studies of the GCN4 leucine zipper and are the
consequence of efficient ‘knobs-into-holes’ packing of
sidechains at the a and d positions (Table 3) [13,15,16].
However, analysis of a large variety of coiled-coil struc-
tures shows less clear preference of polypeptide sequence
for different oligomerization states [17] (Table 3),
although it is not clear whether the state of oligomeriza-
tion of these structures has been affected by adjacent
globular domains. Furthermore, a particular isolated
heptad-repeat sequence can form a mixture of dimers and
trimers if the residues at the a and d positions do not fit
easily into available holes [18,19]. Out of 46 residues at a
positions in the wild-type fibritin, there are 16 isoleucines,
14 valines and 10 leucines, and, out of 43 residues at d
positions, there are 15 leucines, 13 isoleucines and 14
Research Article Bacteriophage T4 fibritin Tao et al. 793
Figure 4
Predicted secondary structure of the coiled-
coil domain of complete wild-type fibritin.
Residues listed in each column headed
‘abcdefg’ are predicted to form ?-helices
(?1–?12) and residues at the right side of
each column form insertion loops (L1–L11)
between the two neighboring ?-helical
segments. Residues in the a or d positions are
highlighted in boxes. Residues corresponding
to the known fibritin E structure are shown in
bold italics.
(α7)
(α6)
(α11)
(α10)
(α9)
(α8)
(α12)
(L6)
(L11)
(L10)
(L9)
(L8)
(L7)
(α1)
(α5)
(α4b)
(α4a)
(α3)
(α2)
(L1)
(L5)
(L4b)
(L4a)
(L3)
(L2)
b c d e f g a
51 . . V L R N V
56 E V L D K N I
63 G I L K T S L
70 E T A N S D I
77 K T I Q G I L
DVSGDIE
ALAQIG
97 I N K K D I
103 S D L K T L T
110 S E H T E I L
117 N G T N N T V
124 D S I L A D I
GPFNAEA
NSVYRT
144 I R N D L
149 L W I K R E L
GQYTGQD
INGLPVV
GNPSSG
176 M K H R I
181 I N N T D V I
188 T S Q G I R L
SELETKF
IESDVG
208 S L T I E V
214 G N L R E E L
GPKPPSF
SQN
231 V Y S R L
236 N E I D T K Q
243 T T V E S D I
250 S A I K T S I
GYPGNN
b c d e f g a
263 S I I T S V
269 N T N T D N I
276 A S I N L E L
NQSGG
288 I K Q R L
293 T V I E T S I
GSDDIPS
S
308 I K G Q I
313 K D N T T S I
320 E S L N G I V
GENTSSG
334 L R A N V
339 S W L N Q I V
GTDSSGG
QPSPPG
359 S L L N R V
365 S T I E T S V
372 S G L N N A V
379 Q N L Q V E I
GNNSAG
392 I K G Q V
397 V A L N T L V
NGTNPNG
STVEERG
418 L T N S I
423 K A N E T N I
430 A S V T Q E V
437 N T A K G N I
444 S S L Q G D V
451 Q A L Q E A
GYIP....
Table 3
Preferred residues in parallel coiled-coil oligomers.
Rules based on GCN4 leucine zipper mutants*Analysis of various coiled-coil structures†
OligomerPosition aPosition dPosition aPosition d
Dimer
Trimer
Tetramer
?-branched residues LeucinesFew valines Few isoleucines, few valines
No preference
Not studied
Predominantly ?-branched
Leucines
?-branched residues
*See references [13,15,16]. †See reference [17].

Page 6

valines. This distribution is roughly consistent with an
expectation for a typical trimer, rather than a dimer or a
tetramer. On the other hand, in fibritin E, of the 10
residues at the a position, 6 are valines and 4 are
isoleucines (all ? branched) and of the 10 residues at the d
positions, there are 6 leucines, 1 isoleucine, 1 alanine, 1
valine and 1 asparagine (i.e. leucines are predominant).
This suggests that fibritin E would be more stable as a
dimeric, rather than a trimeric, coiled-coil structure. A
similar situation exists in the trimeric coiled-coil domain
of human mannose-binding protein [20]. The lack of suit-
able residues at the a and d positions in fibritin E is also
apparent in the large distances between closest contacts of
symmetry-related sidechain atoms. In fibritin E, the
average distance is 4.42Å, whereas in an isoleucine zipper
trimer it is 3.95Å [13]. The larger distance in fibritin E is
because of the dominance of valines (‘knobs’), which do
not extend far enough to completely fill the holes in the
presence of some isoleucines that keep the ? helices sepa-
rated. This accounts for the larger distance between
valines than there would be between isoleucines. Thus,
the coiled coil in fibritin appears to be less energetically
favorable than a perfect trimer.
A least-squares procedure was used to parameterize
coiled-coil segments, as described in Materials and
methods (Table 4). These parameters were compared to
those for other coiled-coil structures (Table 2). The rela-
tively short pitch of fibritin E compared to other trimeric
coiled-coil structures can be explained by the efficient
packing of valines at the a position [18].
There was unassigned electron density on the coiled-coil
axis, at the center between the symmetry-related Asn425
residues at the d position. This density was modeled as a
chloride ion, which was present in the protein buffer. If
this site had been a Zn2+ion, it would probably have
been substituted by Pb2+in the heavy-atom derivative.
In wild-type fibritin, the predicted structure has five
asparagines at d positions. It has been suggested that
asparagines at the d positions, which are stabilized by a
chloride ion, increase the specificity of alignment
[21,22]. In addition, there are salt bridges between
residues from adjacent subunits—Glu384–Lys393,
Glu414–Lys423 and Arg416–Glu426, which help to stabi-
lize the trimeric coiled coil. However, the predicted salt
bridge between residues Glu435 and Lys440 [8], at posi-
tions g and e, respectively, is not formed in the crystal
structure.
C-terminal domain and its role in folding and trimerization
Each subunit in the fibritin E trimer contains a C-termi-
nal region that has a ?-hairpin motif and a loop connect-
ing the ? hairpin to the end of the last ?-helical segment
(Figs 1,2). Within a subunit, the connecting loop is stabi-
lized by hydrophobic interactions and hydrogen bonds.
Residues Tyr458, Ile459, Trp476, Val470, Leu479,
Phe482 and Leu483 form a hydrophobic interior within
the C-terminal domain of the trimer (Fig. 5a). In particu-
lar, Trp476 interacts with the hydrocarbon chain of
Arg471, which in turn interacts with a symmetry-related
Trp476 in another subunit. The three symmetry-related
? hairpins, each stabilized by five mainchain hydrogen
bonds, form a propeller-like structure (Fig. 5b). There are
two mainchain hydrogen bonds between each pair of
neighboring ? hairpins (Fig. 5c). In addition, there is a
salt bridge between Glu461 and Arg471 from adjacent
subunits.
In order to investigate the folding and oligomerization
pathway of fibritin, a number of truncated fibritin mol-
ecules were expressed in E. coli [23]. These mutant mol-
ecules lacked the N-terminal domain and a progressively
larger amount of the amino end of the ?-helical domain.
The shortest mutant was composed of only the last 58
residues. All of the mutants formed soluble expression
products, except a mutant without the last 18 residues of
the C-terminal domain which produced an insoluble
product. Similarly, substitution of Trp476 by leucine or
serine in the C-terminal domain also produced an insolu-
ble product (MM Shneider & VVM, unpublished results).
On the other hand, substitution of Trp476 by tyrosine or
phenylalanine had no effect on protein solubility. These
observations are consistent with intersubunit interactions
seen in the fibritin E structure. Substitution of Trp476 by
leucine or serine disrupts this interaction, as the relatively
small sidechains do not reach to the hydrocarbon chain of
Arg471. However, mutations of Trp476 to aromatic
residues preserve this interaction. Therefore, the residues
which form the hydrophobic core of the C-terminal
domain are essential for trimerization, so that mutations
which disrupt the hydrophobic core apparently result in
improper subunit association, leading to the formation of
inclusion bodies.
794Structure 1997, Vol 5 No 6
Table 4
Parameters of the three fibritin E ?-helical segments.
Parameters*
?10
?11
?12
Supercoil radius (Å)
Number of residues per supercoil turn
Supercoil pitch (Å)
Supercoil cross angle (°)
Radius of curvature (Å)
Radius of ?-helix (Å)
Number of residues per ?-helix turn
Rise per residue for ?-helix (Å)
?-helix pairwise cross angle (°)
Pairwise inter-helical distance (Å)
Number of heptad repeats
Rms deviation (Å)†
6.4
112.4
165.2
27.2
114.9
2.27
3.64
1.51
23.4
11.0
2
0.32
6.6
76.8
109.7
41.3
53.0
2.24
3.66
1.53
34.8
11.4
1.5
0.49
6.2
95.1
138.6
31.5
84.4
2.30
3.63
1.51
26.9
10.8
5.5
0.33
*See references [13] and [41].†Rms deviation is between calculated
and observed positions of C? atoms.

Page 7

A monomeric ?-helical structure, with its amphipathic
character, would be unstable. There is also some evidence
that, in a coiled coil, the ? helices are not formed until the
individual chains come together [24]. The formation of
inclusion bodies implies random association of the
polypeptide chains due to non-specific hydrophobic inter-
actions between the helices. Therefore, the likely func-
tion of the C-terminal domain of fibritin is perhaps to
provide correct alignment of the subunits to each other.
Thus, after expression of the polypeptide, the correctly
aligned trimer would only be able to form once the com-
plete polypeptide has been synthesized. By that time,
however, the rest of the molecule may already have been
associated randomly. Presumably, the instability of the
random associations would permit their re-association into
correctly aligned trimers upon the formation of ? helices
and coiled coils. Apparently, the interactions within the
stable trimeric C-terminal domain appear to override the
tendency of the last three segments of the coiled-coil
domain to form dimers.
Examples of proteins that contain coiled-coils associated
with non-coiled-coil C-terminal domains are the type I
macrophage receptor [25] and the large class of ‘collectin’
proteins [26]. The assembly pathway of fibritin proposed
here might apply especially to ?-fibrous proteins, such as
tropomyosin [27], intermediate filament protein [28] and
lamin [29]. A C-terminal domain is also required for the
assembly of procollagen, in which folding of the collagen
helix proceeds from the carboxy to the amino end [30].
The structure of fibritin E should, therefore, be valuable
for the study of mechanisms that determine folding and
oligomerization and to provide insights into the function
of coiled-coil proteins with stabilizing non-coiled-coil
domains.
Biological implications
F ibritin is a fibrous structural protein of bacteriophage
T 4. During phage assembly, six fibritin molecules attach
to each virion neck through their N -terminal domains, to
form a collar with six fibers (‘whiskers’). T he whiskers
Research Article Bacteriophage T4 fibritin Tao et al. 795
Figure 5
A
V
Y
W
VLL
R
E
G
D
K
G
N
N
N
N
N
N
N
N
N
N
N
O
O
O
O
O
O
OO
O
O
N
N
N
N
N
N
N
N
N
N
O
O
O
O
O
O
O
O
O
O
N
N
N
N
N
N
N
N
N
N
O
O
O
O
O
O
O
O
O
O
K
D
G
E
R
W
V
V
L
Y
L
A
E
W
V
L
L
D
K
R
V
Y
A
479
473
473
473
479
479
(a)(b)
(c)
The C-terminal domain of fibritin E. (a) Stereo diagram of the
C-terminal domain of a fibritin E subunit. The sidechains shown are
those located in the hydrophobic interior formed at the interface
between three symmetry-related subunits. The vertical line shows the
trimer axis. Atoms are shown in standard colors. (b) Ribbon diagram of
the C-terminal domain looking along the trimer axis, each subunit is
shown in a different color. (c) Mainchain hydrogen bonds formed
within the C-terminal domain of a fibritin E trimer. Parts (a) and (b)
were drawn with MOLSCRIPT [43] and RASTER3D [44].

Page 8

interact with the long tail fibers and accelerate the assem-
bly and attachment of the long tail fibers to the baseplate.
T he structure of fibritin E , a deletion mutant consisting
of the last 119 residues of the 486-residue wild-type
protein, has been determined. T he structure of fibritin E
consists of a long coiled coil and a small globular
C -terminal domain. T he coiled coil is composed of three
segments linked together by exposed loop structures.
T hese insertion loops introduce flexibility into the other-
wise rigid coiled coil, increasing the rotational freedom
of the whiskers, which may be necessary to accelerate
the attachment of the long tail fibers to the tail baseplate.
T he sequence of the coiled-coil domain of fibritin E is
not optimal for the formation of either a stable dimer or
trimer. T he small, globular, C -terminal domain contains
a ?-hairpin structure and its residues are involved in
extensive hydrophobic and some polar interactions
between the subunits of the trimer. T here is mutational
and structural evidence that the C -terminal domain pro-
vides correct alignment of the subunits and defines the
oligomerization state. T he same strategy for attaining
the correct alignment between monomers might be used
by the large class of ?-fibrous proteins, many of which
have a globular domain associated with a coiled-coil
structure. T he structure of fibritin, therefore, provides
valuable details of the mechanisms of protein folding and
oligomerization. It also provides insights into the func-
tions of coiled-coil proteins.
U sing fibritin as a template, a surface-display T 4 vector
was recently developed. T his is the first example where
fused chimeric molecules can be designed for a phage
display system with knowledge of the atomic structure
of the template.
Materials and methods
Crystallization
Recombinant fibritin E was expressed in E. coli strain BL21(DE3) [31]
and purified as described by Efimov et al. [8]. Crystals of fibritin E were
obtained using the hanging drop vapor diffusion method [10]. The best
crystals grew at 22°C when 1?l droplets of the recombinant protein
(29mgml–1in 10mM Tris-HCl, pH 7.5) were mixed with 1µl well solu-
tion containing 34% PEG400, 0.1M Zn(CH3COO)2, 0.1M MES,
pH 6.0. Rhombohedral crystals appeared after one day. Their space
group was R32 with hexagonal cell dimensions a = 41.2, c = 358.7Å.
Data collection
Native and derivative data sets were recorded at room temperature on
an R-axis IIC image plate detector, which was mounted on a Rigaku
rotating anode X-ray source operated at 50kV and 100mA. Each data
set was from a single crystal, and all the data were processed using the
programs DENZO and SCALEPACK [32] (Table 5).
Structure determination
Difference Patterson maps were used to identify the heavy-atom sites
of the uranyl acetate derivative. Thereafter, the heavy-atom sites of the
Pb and Pt derivatives were located by cross Fourier maps combined
with difference Patterson maps. The position of the Pb site was the
same as the major site of the UO2derivative. Heavy-atom parameters
were refined using MLPHARE [33] (Table 1). The anomalous signals
from the UO2and Pb derivatives were included in the refinement and
the phasing. The overall figure of merit for the calculated phases was
0.54 for data between 15 and 3Å resolution. The initial MIR map calcu-
lated to 3.0Å resolution showed good density for most of the ? helices.
The program DM [34] (employing solvent flattening and histogram
matching) was used for further phase improvement. The CCP4 suite of
programs [35] was used for all crystallographic calculations.
Although the MIR electron density map allowed most of the amino
acids to be fitted unambiguously using the program O [36] and the
known sequence, the bigger insertion loop (L11) and the N-terminal
helical segment had been deleted by the solvent flattening procedure.
Another MIR map, modified using DM without solvent flattening, gave
good density for these regions. The initial model included 111 amino
acid residues out of a total of 119. Crystallographic refinement was
performed using the program X-PLOR [37] with idealized amino acid
parameters as defined by Engh and Huber [38]. Four percent of the
reflections were omitted during the refinement and used as a test set to
monitor the refinement process by calculating Rfree[39]. The Fobs ampli-
tudes were re-scaled to Fcalcusing an anisotropic scale factor given
by the expression kexp(B11h2a*2+ B22k2b*2+ B33k2c*2+ 2B12hka*b*+
2B23klb*c*+ 2B13hla*c*), where B11= B22= –7.36, B33= 14.71, B12=
–9.83 and B23= B13= 0.0Å2.
Individual atomic temperature factors were refined after several rounds
of positional refinement and manual correction. A total of 80 water mol-
ecules and a bulk solvent model (B = 160Å2, k = 0.43) were added.
The final Rfreewas 0.267 and Rworkwas 0.217 for data between 30.0
and 2.2Å resolution. The stereochemistry of the model had a root
mean square (rms) deviation of 0.010Å for bond lengths from the ide-
alized values and of 1.3° for bond angles. The overwhelming majority of
amino acids could be associated with well-defined electron density.
The exceptions are the sidechains of Pro408, Asn409 and Ser411,
796Structure 1997, Vol 5 No 6
Table 5
Data collection statistics.
Soaking
time
(days)
Heavy atom
concentration
(mM)
Highest
resolution
(Å)
No. of
measured
reflections
No. of unique
reflections
(% of total)Data setRmerge*Rdiff†
Native
UO2(CH3COO)2
Pb(CH3COO)2
K2PtCl4
–
1
3
3
–2.2
3.0
2.5
2.5
57901
20545
54054
30181
6490 (92.6%)
2666 (99.6%)
4527 (99.5%)
4521 (95.2%)
0.053
0.057
0.046
0.031
–
20
1
1
0.142
0.109
0.090
*Rmerge= Σ
where the FPand FPHare the structure amplitudes of the native and heavy atom derivative compounds, respectively.
hΣ
i|Ihi– 〈Ih〉|/?
h?
iIhi. †Rdiff= ?
h|FPh–FPHh|/?
hFPh,

Page 9

which are located in an insertion loop (L11) and have high temperature
factors. The final protein model was analyzed with the program
PROCHECK [40]. The Ramachandran plot has 91.7% of the residues
in the most favored regions. The only residues in the generously
allowed region is Asp473, which makes a salt bridge with Lys472
within the same subunit.
Coiled-coil parameterization
A least-squares refinement procedure was used to refine 13 parame-
ters which defined the C? positions of the coiled coil in the reference
strand. Six parameters described the orientation and position of the
coiled coil. Another seven parameters [41] described the coiled coil
itself. All the C? atom positions in the coiled coils were used for refine-
ment. The positions of the other strands in the coiled coils were gener-
ated from the reference strand by applying the rotational symmetries
(two, three, four or fivefold). The computer program is available from
the authors.
Accession numbers
Coordinates and structure factors have been deposited with the
Brookhaven Protein Data Bank (entry codes 1aa0 and r1aa0sf, respec-
tively).
Acknowledgements
We would like to thank Guoguang Lu, Jordi Bella, Alan Friedman,
Prasanna Kolatkar, Sukyeong Lee, Lidiya Kurochkina, Yuri Londer and
Mikhail Shneider for their helpful advice and discussion. We also thank
Thomas J Smith for the use of the program MolView in creating Figures
2b and 3a. The work was supported by a grant from the National
Science Foundation (MCB-9102855) to MGR, and also by grants from
the Protein Engineering Council of Russia and from the Russian Foun-
dation of Basic Research (96-04-48035) to VVM. VVM is an Interna-
tional Howard Hughes Medical Institute Scholar.
References
1.Eiserling, F.A. & Black, L.W. (1994). Pathways in T4 morphogenesis.
In Molecular Biology of Bacteriophage T4. (Karam, E., ed), pp.
209–212, American Society for Microbiology, Washington, DC, USA.
2.Wood, W.B. & Conley, M.P. (1979). Attachment of tail fibers in
bacteriophage T4 assembly. J. Mol. Biol. 127, 15–29.
3.Bloomfield, V.A. (1983). Physical studies of morphogenetic reactions.
In Bacteriophage T4. (Mathews, C.K., Kutter, E.M., Mosig, G. &
Gerget, P.B., ed), pp. 270–276, American Society for Microbiology,
Washington, DC, USA.
4.Wood, W.B. & Henninger, M. (1969). Attachment of tail fibers in
bacteriophage T4 assembly: some properties of the reaction in vitro
and its genetic control. J. Mol. Biol. 39, 608–618.
5.Conley, M.P. & Wood, W.B. (1975). Bacteriophage T4 whiskers: a
rudimentary environment-sensing device. Proc. Natl. Acad. Sci. USA
72, 3701–3705.
6.Sobolev, B.N. & Mesyanzhinov, V.V. (1991). The wac gene product of
bacteriophage T4 contains coiled-coil structure patterns. J. Biomol.
Struct. Dynam. 8, 953–965.
7.Cohen, C. & Parry, D.A.D. (1990). ?-helical coiled coils and bundles:
how to design an ?-helical protein. Proteins 7, 1–15.
8.Efimov, V.P., et al., & Mesyanzhinov, V.V. (1994). Fibritin encoded by
bacteriophage T4 gene wac has a parallel triple-stranded ?-helical
coiled-coil structure. J. Mol. Biol. 242, 470–486.
9.Efimov, V.P., Nepluev, I.V. & Mesyanzhinov, V.V. (1995).
Bacteriophage T4 as a surface expression vector. Virus Genes 10,
173–177.
10. Strelkov, S.V., Tao, Y., Rossmann, M.G., Kurochkina, L.P., Shneider,
M.M. & Mesyanzhinov, V.V. (1996). Preliminary crystallographic
studies of bacteriophage T4 fibritin confirm a trimeric coiled-coil
structure. Virology 219, 190–194.
11. Lupas, A. (1996). Coiled coils: new structures and new functions.
Trends Biochem. Sci. 21, 375–382.
12. Crick, F.H.C. (1953). The packing of ?-helices: simple coiled-coils.
Acta Cryst. 6, 689–697.
13. Harbury, P.B., Kim, P.S. & Alber, T. (1994). Crystal structure of an
isoleucine-zipper trimer. Nature 371, 80–83.
14. Bullough, P.A., Hughson, F.M., Skehel, J.J. & Wiley, D.C. (1994).
Structure of influenza haemagglutinin at the pH of membrane fusion.
Nature 371, 37–43.
15. O’Shea, E.K., Klemm, J.D., Kim, P.S. & Alber, T. (1991). X-ray structure
of the GCN4 leucine zipper, a two-stranded, parallel coiled coil.
Science 254, 539–544.
16. Harbury, P.B., Zhang, T., Kim, P.S. & Alber, T. (1993). A switch
between two-, three-, and four-stranded coiled coils in GCN4 leucine
zipper mutants. Science 262, 1401–1407.
17. Woolfson, D.N. & Alber, T. (1995). Predicting oligomerization states of
coiled coils. Protein Sci. 4, 1596–1607.
18. Gonzalez Jr., L., Brown, R.A., Richardson, D. & Alber, T. (1996).
Crystal structures of a single coiled-coil peptide in two oligomeric
states reveal the basis for structural polymorphism. Nat. Struct. Biol. 3,
1002–1010.
19. Ogihara, N.L., Weiss, M.S., DeGrado, W.F. & Eisenberg, D. (1997).
The crystal structure of the designed trimeric coiled coil coil-VaLd:
implications for engineering crystals and supramolecular assemblies.
Protein Sci. 6, 80–88.
20. Sheriff, S., Chang, C.Y. & Ezekowitz, R.A. (1994). Human mannose-
binding protein carbohydrate recognition domain trimerizes through a
triple ?-helical coiled coil. Struct. Biol. 1, 789–794.
21. Fass, D., Harrison, S.C. & Kim, P.S. (1996). Retrovirus envelope
domain at 1.7Å resolution. Nat. Struct. Biol. 3, 465–469.
22. Lumb, K.J. & Kim, P.S. (1995). A buried polar interaction imparts
structural uniqueness in a designated heterodimeric coiled coil.
Biochemistry 34, 8642–8648.
23. Sobolev, B.N., Shneider, M.M., Marusich, E.I. & Mesyanzhinov, V.V.
(1995). Fibritin and adhesin of bacteriophage T4 tell us how fibrous
protein folds and assembles. In Evolutionary Biochemistry and
Related Areas of Physicochemical Biology. (Poglazov, B., Kurganov,
B.I., Kritsky, M.S. &Gladilin, K.L., eds), pp. 351–374, Bach Institute of
Biochemistry and ANKO, Moscow, Russia.
24. Sosnick, T.R., Jackson, S., Wilk, R.R., Englander, S.W. & DeGrado,
W.F. (1996). The role of helix formation in the folding of a fully
?-helical coiled coil. Proteins 24, 427–432.
25. Kodama, T., Freeman, M., Rohrer, L., Zabrecky, J., Matsudaira, P. &
Krieger, M. (1990). Type I macrophage scavenger receptor contains
?-helical and collagen-like coiled coils. Nature 343, 531–535.
26. Hoppe, H. & Reid, K.B.M. (1994). Collectins — soluble proteins
containing collagenous regions and lectin domains — and their roles
in innate immunity. Protein Sci. 3, 1143–1158.
27. Phillips Jr., G.N., Fillers, J.P. & Cohen, C. (1986). Tropomyosin crystal
structure and muscle regulation. J. Mol. Biol. 192, 111–131.
28. Parry, D.A.D. & Fraser, R.D.B. (1985). Intermediate filament structure.
1. Analysis of IF protein sequence data. Internatl. J. Biol. Macromol. 7,
203–213.
29. McKeon, F.D., Kirschner, M.W. & Caput, D. (1986). Homologies in
both primary and secondary structure between nuclear envelope and
intermediate filament proteins. Nature 319, 463–468.
30. Prockop, D.J. & Kivirikko, K.I. (1995). Collagens: molecular biology,
diseases, and potentials for therapy. Ann. Rev. Biochem. 64,
403–434.
31. Studier, F.W., Rosenberg, J.J., Dunn, J.J. & Dubendorff, J.W. (1990).
Use of T7 RNA polymerase to direct expression of cloned genes.
Methods Enzymol. 185, 60–89.
32. Otwinowski, Z. & Minor, W. (1997). Processing of X-ray diffraction
data collected in oscillation mode. Methods Enzymol. 276A,
307–326.
33. Otwinowski, Z. (1991). Maximum likelihood refinement of heavy atom
parameters. In Isomorphous Replacement and Anomalous Scattering.
Proceedings of the CCP4 Study Weekend. (Wolf, W., Evans, P.R. &
Leslie, A.G.W., eds), pp. 80–86, SERC Daresbury Laboratory,
Warrington, UK.
34. Cowtan, K. (1994). ‘dm’: an automated procedure for phase
improvement by density modification. Joint CCP4 and ESF-EACBM
Newsletter on Protein Crystallography 31, 34–38.
35. Collaborative computational project number 4 (1994). The CCP4
Suite: programs for protein crystallography. Acta Cryst. D 50,
760–763.
36. Jones, T.A., Zou, J.-Y., Cowan, S.W. & Kjeldgaard, M. (1991).
Improved methods for building protein models in electron density
maps and the location of errors in these models. Acta Cryst. A 47,
110–119.
37. Brünger, A.T. (1992). X-PLOR. Version 3.1. A system for X-ray
Crystallography and NMR. Yale university press, New Haven &
London.
Research Article Bacteriophage T4 fibritin Tao et al. 797

Page 10

38. Engh, R.A. & Huber, R. (1991). Accurate bond and angle parameters
for X-ray protein structure refinement. Acta Cryst. A 47, 392–400.
39. Brünger, A.T. (1992). Free R value: a novel statistical quantity for
assessing the accuracy of crystal structures. Nature 355, 472–475.
40. Laskowski, R.A., MacArthur, M.W., Moss, D.S. & Thornton, J.M.
(1993). PROCHECK: a program to check the stereochemistry quality
of protein structures. J. Appl. Cryst. 26, 283–291.
41. Crick, F.H.C. (1953). The Fourier transform of a coiled-coil. Acta
Cryst. 6, 685–689.
42. Smith, T.J. (1995). MolView: a program for analyzing and displaying
atomic structures on the Macintosh personal computer. J. Mol.
Graphics 13, 122–125.
43. Kraulis, P. (1991). MOLSCRIPT: a program to produce both detailed
and schematic plots of protein structures. J. Appl. Cryst. 24,
946–950.
44. Merritt, E.A. & Murphy, M.E.P. (1994). Raster3D version 2.0: a
program for photorealistic molecular graphics. Acta Cryst. D 50,
869–873.
45. Weber, I.T. & Steitz, T.A. (1987). Structure of a complex of catabolite
gene activator protein and cyclic AMP refined at 2.5Å resolution.
J. Mol. Biol. 198, 311–326.
46. Marmorstein, R., Carey, M., Ptashne, M. & Harrison, S.C. (1992). DNA
recognition by GAL4: structure of a protein–DNA complex. Nature
356, 408–414.
47. Weis, W.I. & Drickamer, K. (1994). Trimeric structure of a C-type
mannose-binding protein. Structure 2, 1227–1240.
48. Malashkevich, V.N., Kammerer, R.A., Efimov, V.P., Schulthess, T. &
Engel, J. (1996). The crystal structure of a five-stranded coiled coil in
COMP: a prototype ion channel? Science 274, 761–765.
798Structure 1997, Vol 5 No 6