In vitro assembly of a prohead-like structure of the
Rhodobacter capsulatus gene transfer agent
Anthony J. Spanoa,⁎,1, Frank S. Chena,1, Benjamin E. Goodmana,2, Agnes E. Sabata,
Martha N. Simonb, Joseph S. Wallb, John J. Correiac, Wilson McIvora, William W. Newcombd,
Jay C. Brownd, Joel M. Schnure, Nikolai Lebedeva,e
aDepartment of Biology, University of Virginia, Charlottesville, VA 22904, USA
bBiology Department, Brookhaven National Laboratory, Upton, NY 11973, USA
cDepartment of Biochemistry, University of Mississippi Medical Center, Jackson, MS 39216, USA
dDepartment of Microbiology and Cancer Center, University of Virginia Health System, Charlottesville, VA 22908, USA
eCenter for BioMolecular Sciences and Engineering, U.S. Naval Research Laboratory, Washington, DC 20375, USA
Received 24 November 2006; returned to author for revision 26 December 2006; accepted 23 February 2007
Available online 3 April 2007
The gene transfer agent (GTA) is a phage-like particle capable of exchanging double-stranded DNA fragments between cells of the
photosynthetic bacterium Rhodobacter capsulatus. Here we show that the major capsid protein of GTA, expressed in E. coli, can be assembled
into prohead-like structures in the presence of calcium ions in vitro. Transmission electron microscopy (TEM) of uranyl acetate staining material
and thin sections of glutaraldehyde-fixed material demonstrates that these associates have spherical structures with diameters in the range of 27–
35 nm. The analysis of scanning TEM images revealed particles of mass ∼4.3 MDa, representing 101±11 copies of the monomeric subunit. The
establishment of this simple and rapid method to form prohead-like particles permits the GTA system to be used for genome manipulation within
the photosynthetic bacterium, for specific targeted drug delivery, and for the construction of biologically based distributed autonomous sensors for
© 2007 Elsevier Inc. All rights reserved.
Keywords: Rhodobacter capsulatus; Gene transfer agent; Prohead-like; Assembly in vitro
The gene transfer agent (GTA) of Rhodobacter capsulatus is
a tailed phage-like particle that mediates an unusual form of
transfer of genetic information between bacterial cells (Hoch-
man, 1997; Marrs, 1974, 1978, 2002). Unlike many known
viruses, GTA does not form plaques and does not propagate via
independent packaging of its own DNA. Instead, it owes its
continued existence entirely to its presence on the non-excisable
GTA gene cluster found on the bacterial chromosome (Lang and
Beatty, 2000, 2001; Marrs, 1974). From this viewpoint, the
GTA can be considered either as a defective prophage, or a
phage precursor, or even a cell-based system for the transfer of
genetic information in bacterial populations (Lang and Beatty,
2000, 2001; Yen et al., 1979). GTA randomly packages small
amounts (≤5 kb) of R. capsulatus DNA and transfers the
packaged DNA to capsulatus strains capable of GTA uptake
(Solioz and Marrs, 1977; Solioz et al., 1975; Wall et al., 1975a;
Yen et al., 1979). This feature allowed the GTA to be used to
map R. capsulatus genes (Donohue and Kaplan, 1991; Scolnik
et al., 1980; Wall and Braddock, 1984; Wall et al., 1975b, 1984;
Yen and Marrs, 1976).
Virology 364 (2007) 95–102
Abbreviations: PBS, phosphate-buffered saline (pH 7.4); NiNTA, nickel
nitrilotriacetic acid; GTA, gene transfer agent; STEM, scanning transmission
electron microscopy; TEM, transmission electron microscopy.
⁎Corresponding author. University of Virginia, Department of Biology,
Gilmer Hall, Charlottesville, VA 22903, USA. Fax: +1 434 982 5626.
E-mail address: firstname.lastname@example.org (A.J. Spano).
1These authors contributed equal effort.
2Present address: Johns Hopkins University, Graduate Program in Cell and
Molecular Biology, Baltimore, MD, USA.
0042-6822/$ - see front matter © 2007 Elsevier Inc. All rights reserved.
The non-lytic nature of GTA makes it very attractive as a tool
for manipulation of genetic information in bacterial, and
particularly in photosynthetic bacterial populations (Brussaard,
2004; Danovaro et al., 2003; Munn, 2006). Since the GTA
packages DNA randomly, it is attractive as a universal vehicle
for targeted drug delivery (Hendrie and Russell, 2005; Normand
et al., 2005; Seth, 2005; Tomanin and Scarpa, 2004). It also can
be used as a platform for nano-electronic and bio-electronic
applications (Blum et al., 2004, 2005; Merzlyak and Lee, 2006;
Singh et al., 2006; Slocik et al., 2005; Soto et al., 2006; Souza et
al., 2006; Wang et al., 2002a, 2002b; Zhang, 2003). This
requires a detailed characterization of the structure and the
mechanism of assembly of the GTA prohead.
Pioneering work by the Marrs laboratory identified 5 to 8
polypeptides ranging in size between 13 and 40 kDa that were
attributed to the GTA particle purified from the Y262 over-
expression strain of R. capsulatus (Yen et al., 1979). Compara-
tive bioinformatics approaches were used to predict structural
components and regulatory proteins of the GTA. That includes
the identification of the members of the R. capsulatus GTA gene
cluster and the histidine kinase/sensor (ctrA/cckA) of GTA
regulatory genes (Lang and Beatty, 2000, 2001, 2002; Lang et
al., 2002). Genetic analysis has shown that GTA genes form a
cluster of about 18 genes, with an upstream cluster-specific
promoter element (Lang and Beatty, 2000), driving expression
GTA cluster (Lang and Beatty, 2000; Schaefer et al., 2002). The
promoter activity is upregulated as the cells enter the stationary
phase of the growth cycle (Lang and Beatty, 2000; Solioz et al.,
1975). Small molecules of the acyl-homoserine lactone class are
involved in the production of the GTA during the growth of R.
and the mechanism of its assembly both in vivo and in vitro have
not been identified.
As a first step in this direction in the present work, we have
developed a strategy for assembly of the GTA in vitro. To
achieve this goal we cloned and over-expressed the 42 kDa Orf5
gene product, the major GTA capsid protein, in E. coli. Then we
purified its using size exclusion and ion exchange chromato-
graphy and assembled the pure protein into prohead-like
structures in vitro in the presence of calcium ions. Our results
show that the in vitro assembled prohead-like structures form
hollow spheres of about 27–35 nm diameter.
Results and discussion
Purification and subunit structure of the Orf5
To identify the complete DNA sequence corresponding to
the major capsid protein (Orf5) of the GTA particle, we
amplified the corresponding coding region from R. capsulatus
genomic DNA using the Orf5 nucleotide sequence available
from the R. capsulatus Genome Database website (http://www.
integratedgenomics.com). The full-length capsid gene, corre-
sponding to a protein of 398 amino acids, was amplified using
oligodeoxynucleotide primers, and the product was digested
and subcloned into the pET17b bacterial protein expression
vector. The construct was propagated in E. coli strain BL21
(DE3) for protein expression.
The Orf5 capsid protein, expressed in E. coli, was purified
by size exclusion chromatography followed by ion exchange
chromatography. On SDS–PAGE the highly purified Orf5
monomer resolved as a single major polypeptide at 42 kDa
accompanied by a minor polypeptide at approximately 35 kDa
(Fig. 1A). The origin of the second (35 kDa) band seen in the
gel in addition to the expected main purified Orf5 polypeptide
(42 kDa) was not initially clear. To identify it we performed
Edman degradation chemistry on the 35 kDa protein. The N-
terminus of the 35 kDa protein begins with the amino acid
residues YAGRH corresponding to a position 59 residues from
the N-terminus of the 42 kDa full-length Orf5 protein and is
completely consistent with the mass shift observed on SDS–
PAGE. Thus, the 35 kDa polypeptide is a proteolytic fragment
of the 42 kDa form of the Orf5 polypeptide. The staining profile
of the lane for pure Orf5 protein indicates that the 42 kDa
protein in our preparations is about 90% pure as determined by
quantitative gel scanning (Fig. 1A).
When we performed size exclusion chromatography on a
calibrated Superdex 200 column, we observed that the Orf5
migrated at an apparent mass of 217 kDa (Fig. 1B), that is about
5.2 fold higher than the monomer observed by SDS–PAGE.
This discrepancy can be explained if we assume that the R.
capsulatus Orf5 protein is expressed, folded and assembled
within E. coli into a pentamer, a biologically relevant structure
that should make subsequent assembly into a prohead-like
particle possible. As it is not possible to rule out the existence of
hexameric forms of Orf5 using size exclusion chromatography,
we subjected the purified Orf5 protein to both sedimentation
velocity and sedimentation equilibrium centrifugation. The best
fit to the equilibrium data reveals a mixture of trimer (3.26
Fig. 1. (A) Coomassie-stained SDS–PAGE of purified native GTA Orf5 protein
(Orf5). Fractions containing highly purified Orf5 were denatured and subjected
to electrophoresis on a 12% polyacrylamide Laemmli SDS-denaturing gel. The
purified protein migrates at the mass of 42 kDa expected for the Orf5 monomer.
The band at 35 kDa is a proteolytic fragment of the 42 kDa monomer. The lane
shows 12 μg of recombinant protein after Coomassie staining. (B) Molecular
weight of E. coli-expressed Orf5 protein as determined by size exclusion
chromatography on Superdex 200. Purified GTA Orf5 (∼100 mg; ■) was
injected onto a calibrated size exclusion column. Standard proteins (♦) include
alcohol dehydrogenase (AD, 150 kDa), bovine serum albumin (B, 67 kDa), and
carbonic anhydrase (C, 29 kDa). The Orf5 was found to migrate as a pentamer
with an estimated mass of 217 kDa (square, above). The Superdex 200 column
was equilibrated in PBS and run at 0.4 ml/min at 4 °C.
96A.J. Spano et al. / Virology 364 (2007) 95–102
<2.63, 3.39>) and pentamer (4.66 <4.29, 5.08>) where the
units are full-length subunits in size and the error bars are two
standard deviations. We conclude that the velocity and
equilibrium data are consistent with the existence of both
trimers and pentamers in 40 mM Tris, pH 8.0 with 150 mM
NaCl at 3.6 °C that undergo a reversible reaction.
Assembly of Orf5 into particles
Experiments performed with the coliphage HK97 have
shown that its prohead assembly in vitro can be initiated by the
incubation of capsomeres composed of a single major capsid
protein (gp5) with calcium and PEG-8000 (Xie and Hendrix,
1995). To test the possibility of using a similar strategy for the
GTA, we incubated Orf5 with 10 mM CaCl2,or both 10 mM
CaCl2and 4% PEG-8000 at room temperature for 1 h. Samples
were then placed on ice and then subjected to electrophoresis in
a 1% (w/v) agarose gel containing 40 mM Tris–HCl (pH 8.0)
and 1 mM CaCl2. Nearly all of the Orf5 pentamer without added
calcium migrated as a single band at the capsomere position
(Fig. 2, “C”). When calcium was added, we observed the
appearance of a band running faster than the Orf5 pentamer
indicating the formation of higher molecular mass species (Fig.
2, “P”). At a calcium ion concentration of ∼10 mM a nearly
complete conversion to the fast-migrating prohead-like particle
was observed. Moreover, comparing the experiments with and
without PEG showed that the presence of only Ca2+ion is
enough to effect the conversion (Fig. 2, lanes 2 and 4). When
assembled into a prohead-like particle the GTA migrates faster
on a native agarose gel, and the protein band is sharper.
Although the faster migration of the assembled prohead-like
particle compared to unassembled subunits appears counter-
intuitive, the same phenomenon has been noted before (Xie and
Hendrix, 1995). The precise explanation for this migration
pattern is unclear.
To test whether the increased mobility of the Orf5 sample
after pre-incubation with calcium±PEG 8000 is consistent with
an increase in size indicative of prohead formation we subjected
the Orf5 preparation to size exclusion chromatography. A shift
from ∼225 kDa to a higher molecular weight position at or near
the void of the column was seen (Fig. 3), completely consistent
with the conversion of a slower migrating to a faster migrating
form seen on agarose gel electrophoresis (Fig. 2) (Xie and
Imaging of the assembled Orf5 prohead structure
To elucidate the structure of assembled high molecular
weight particles, we performed an electron microscopic
examination. Transmission electron microscopy of uranyl
acetate staining material (negatively stained TEM) revealed
round particles with diameter 34.7±3.7 nm (Fig. 4A). The
diameters of the particles estimated by scanning transmission
electron microscopy (STEM) are in the range of 27.6±2.0 nm
(Fig. 4B). Both diameter estimates bracket the size of the GTA
capsids (diameter ∼30 nm) isolated from R. capsulatus Y262
cells (Yen et al., 1979). TEM images of the thin sections of
glutaraldehyde-fixed material revealed that the particles were
round, hollow particles (data not shown).
To further identify the structure of the assembled particles we
took STEM measurements of the calcium-treated Orf5 material.
Fig. 2. Calcium-induced assembly of GTA Orf5 protein into high molecular
weight particlesin vitro. Orf5 (1.2 mg/ml final concentration) in 10 mM Tris–HCl
(pH 8.0) was mixed with calcium chloride (10 mM final), PEG-8000 (4%w/v
final), or both, and incubated for 1 h at room temperature. Lane 1) Orf5 no
additions, 12 μg; lane 2) Orf5 with 10 mM CaCl2, 12 μg; lane 3) Orf5 no
additions, 6 μg; lane 4) Orf5 with 10 mM CaCl2, 6 μg; lane 5) Orf5 with 10 mM
CaCl2and 4% PEG, 6 μg; lane 6) Orf5 4% PEG-8000, 6 μg. Electrophoresis was
performed in 40 mM Tris–HCl (pH 8.0) with 1.0 mM CaCl2at 40 Vat 4 °C until
the dye front reached the bottom of the gel. The agarose gel was stained with
Coomassie Blue, destained and photographed. C=position of capsomeres,
P=position of prohead-like particle.
Fig. 3. Superdex 200 elution profile of GTA Orf5 protein before and after
assembly into high molecular weight particles in vitro. Purified untagged GTA
Orf5 protein (■) was subjected to size exclusion chromatography after assembly
into particles by the addition of 10 mM CaCl2(●) or 10 mM CaCl2with 4%
PEG-8000 (▴). The elution volume of the untreated (control) sample
corresponds to the position expected of capsomere-like particles while the
assembled particles elute at a position corresponding to a very large size (at or
beyond the exclusion limit of the Superdex 200 column (≥1.3×106Da).
Column dimensions are 30 cm×1 cm. Protein samples (∼300 mg in 300 ml)
were chromatographed in 40 mM Tris–HCl (pH 8.0) containing 150 mM NaCl
and 1.0 mM CaCl2. Values of A280for the three separate size exclusion runs
were plotted so that the shape of the elution curves could be visually compared
97 A.J. Spano et al. / Virology 364 (2007) 95–102
We observed roughly spherical particles having mass 4.27±
0.468 MDa (Fig. 4B), measured using PCMass29a software.
Given the molecular weight of a monomeric subunit, this
corresponds to roughly 101±11 subunits. This is more than the
number expected for a T=1 (60 subunits) particle and less than
that anticipated for a T=3 (180 subunits) particle.
The higher-order hollow spherical particles assembled when
the Orf5 GTA protein is exposed to divalent calcium ions are
highly reminiscent of viral proheads (Fig. 4A). There is a
general uniformity in both the size and shape of the particles.
The mass of the structures assembled in vitro estimated by
STEM is not, however, consistent with that predicted from the
standard formulas for triangulation number, if we assume that
pentons and hexons are made up of five and six copies of Orf5,
The GTA prohead particles may not have strict icosahedral
symmetry at this stage. It is possible to generate models for a
prohead consisting of approximately 100 subunits that would
result in the spherically shaped objects close to the mass we
measure. The Johnson solid known as a pentagonal orthobir-
otunda is a closed convex polygon composed of only pentamers
and trimers that gives a predicted estimate reasonably close to
the highest measured estimate. This model can be made
compatible with our size exclusion data if we assume that
dimers of trimers co-migrate with the pentamers, or if the
pentamers convert to trimers. Similarly, a T=3 icosahedron
with hexons composed of trimers would give a similar mass
estimate. The above models have the support of the data derived
from equilibrium sedimentation analysis we have performed,
which shows the existence of only pentameric and trimeric
forms of the Orf5 protein.
It is also possible to build a model consisting of a 20-sided
polyhedron containing 8 hexamers (dimers of trimers) and 12
pentamers, where six hexamers occupying an equatorial
position are joined to two hexamers above and below by rings
of six pentamers. Such an unusual object, while not representing
a classic polyhedron of the Archimedian or Platonic type, might
still self-assemble if there is enough angular flexibility
conferred by the protein–protein interactions of the capsomeres
to permit the faces to form a closed structure.
The virtue of the above models is that they predict the
measured mass of the prohead-like particles we have produced
much more accurately than do standard T=1 or T=3 icosahedra
consisting of pentamers alone, or pentamers and hexamers.
Lastly we note that it is possible that the prohead-like particle
we form in the presence of calcium does not represent a native
GTA prohead. It has been shown that aberrant prohead
structures form in cells and in vitro when components needed
for normal assembly are either defective or are missing
(Earnshaw and King, 1978; Howatson and Kemp, 1975; Li et
al., 2005) In the present case, if the pentameric form of
capsomeres produced in E. coli is made, whereas the hexamers
are not, or if the hexamers decay into trimers, then it is possible
that an alternate structure such as the one we have isolated could
be formed. Additional experimental evidence is needed to
discriminate among these possibilities, and this is the subject of
our current work.
Siphoviridae. A BLAST search using the GTA Orf5 protein
sequence against the non-redundant database demonstrates that
Fig. 4. Transmission electron micrographs of the in vitro assembled GTA
particles. (A) TEM micrograph of negative stained material. Freshly prepared
GTA prohead particles were mounted and stained with 2% uranyl acetate and
imaged. Small spherically shaped particles are the GTA prohead particles.
Herpes Simplex Virus capsids were added to the grid and used for scale. Scale
bar=125 nm. A single Herpes capsid is seen in the center of the image. (B)
STEM micrograph of the GTA particles. The particles were assembled from
purified, unmodified GTA Orf5 protein in the presence of 10 mM CaCl2. The
cylindrical object at the bottom right of the image is tobacco mosaic virus
(TMV) used as a standard for determining GTA particle mass. The diameter of
TMV is 18 nm. Scale bar=50 nm. Measured mass of the GTA prohead=4.27±
0.468 MDa (mean±std deviation; 80 particles independently measured). The
calculated mass of the GTA Orf5 monomer is 42,170.6 Da.
98A.J. Spano et al. / Virology 364 (2007) 95–102
this protein bears the highest degree of sequence similarity to
HK97-family capsid proteins from the Roseobacter clade that
includes the marine bacteria Jannaschia, Oceanicola, and Ro-
general physical resemblance,most HK97“family members”do
not bear high homology in terms of their primary amino acid
sequence (Helgstrand et al., 2003).
The lengths of GTA Orf5 protein and the functionally equi-
valent HK97 gp5 protein are quite similar. When we compared
primary sequence of the mature capsid protein of the GTA Orf5
against the same set of sequences chosen by Helgstrand et al.
(2003), only 9 positions were found to be conservatively
substituted, and only 2 positions were identical among all nine
viruses chosen from the HK97-like family. The situation
improved significantly if we aligned mature Orf5 protein
against the HK97 capsid protein alone. In this case, out of a total
of 282 residues in the mature HK97 capsid protein, 63 positions
are conservatively substituted with respect to Orf5, and 50 are
identical. The primary sequence similarity between GTA Orf5
and HK97 capsid protein (gp5) is not the only common property
found between these phages. In R. capsulatus, the product of
GTA Orf5 is a 42 kDa protein that gets modified by the removal
of about 11 kDa at the N-terminus to produce a mature capsid
protein in these cells (Lang and Beatty, 2000), much like the
maturation step seen in HK97 (Hendrix, 2005). Still, important
differences remain. GTA capsid proteins are not covalently
cross-linked, as evidenced by a prominent 31 kDa protein found
in SDS–PAGE of the isolated mature virus-like particle (Lang
and Beatty, 2000; Yen et al., 1979). In the HK97 system, Lys169
is one of the pair of residues that is covalently cross-linked to its
partner, Asn356(Duda et al., 1995; Ross et al., 2005; Wikoff et
al., 2000). Neither residue is present at equivalent positions in
the GTA Orf5, nor is the GTA Orf5 equivalent of the
mechanistically important Glu363position conserved (Wikoff
et al., 2000). Nonetheless, given the similarity in size of
precursor and mature product, the proteolytic cleavage of the N-
terminal domain and the ability to form prohead-like structures
in the presence of calcium, our observations and data suggest
the possibility of a similar protein folding and association
mechanism in the GTA particle, common to other members of
the larger HK97 family.
What is the role of calcium in the assembly of the prohead-
like particle? It should be noted that there are 52 charged
carboxylates present in each 42 kDa GTA Orf5 molecule
(number of Asp+number of Glu+one C-terminal carboxy-
late). If we assume that (1) some of the negatively charged
residues are unavailable because they are sequestered in salt–
bridge interactions within or between Orf5 proteins, and (2)
that one Ca2+ion effectively neutralizes one carboxylate, the
ratio of calcium ions per 42 kDa subunit needed to initiate
prohead formation is nearly stoichiometric. This might indicate
that a major part of the GTA assembly pathway in vitro
involves metal-dependent charge neutralization to permit
capsomeres to approach each other and associate in a
productive fashion. How this compares to the situation within
the bacterial cell is uncertain, as that would be expected to be a
chaperone-mediated process (Xie and Hendrix, 1995), and the
metal-dependent folding seen in vitro is not even necessarily
reflected in incorporation of metal ions into the capsid shells
of the assembled particles (Helgstrand et al., 2003). Resolution
of this matter awaits a direct structural analysis of the GTA
The design and construction of new molecular tools to
implement specific, efficient drug delivery are required to
realize the promise of genetic therapy. Similar tools would also
open unprecedented possibilities for genetic manipulation in
ecosystems. In this study, we have demonstrated that we can
express high levels of the major capsid protein (Orf5) of the
GTA, a phage-like particle capable of exchanging genetic
information between cells of R. capsulatus. Under the control
of the T7-based pET expression system, tens of milligrams of
the Orf5 can be expressed per liter of E. coli culture, and then
assembled into prohead-like particles in the presence of calcium
ions. The analysis performed using transmission electron
microscopy of uranyl acetate stained material and using
scanning transmission electron microscopy demonstrates that
these nanoparticles are spherical and hollow, having diameters
in the range of 27–35 nm and mass of ∼4.3 MDa, representing
101±11 copies of the monomeric subunit.
This is a first step towards an understanding of the GTA
assembly within R. capsulatus. Since GTA cannot propagate in
mammals and is non-infectious in humans, it is in some respects
more attractive for specific targeted drug delivery than human
viruses. Lastly, as the GTA can be co-expressed with
photosynthetic proteins in the same bacterial cell, it is very
suitable for the construction of biologically based distributed
autonomous sensors for environmental monitoring.
Materials and methods
Construction of Orf5
The Orf5, which codes for the major GTA coat protein, was
amplified from genomic DNA of strain Y262 using the primer
pair 5′AAACATATGAAGACCGAGACCAAGG 3′ and 5′
the following conditions for amplification: initial 98 °C dena-
turation, 10 s; 59 °C annealing, 30 s; 72 °C extension, 50 s; 30
cycles. A final extension was carried out at 72 °C for 8 min.
Amplifications were carried out using Phusion DNA polymer-
ase according to the manufacturer's recommendations. After
amplification, the DNA fragments were purified from the
amplification cocktail using a Qiagen PCR Purification Kit and
then digested with NdeI and HindIII, fractionated on low
melting temperature agarose and ligated into pET17b to
produce pFCOrf5 Nat.
The nucleotide sequence of all constructs was confirmed
Purification of the GTA Orf5
E. coli strain BL21(DE3) bearing pFCOrf5-Nat were grown
in 1 l volume of LB medium on a shaker at 150 rpm at 37 °C to
the optical density at 600 nm=0.5, induced with 0.5 mM IPTG,
99A.J. Spano et al. / Virology 364 (2007) 95–102
and immediately shifted to 25 °C for additional 16 h. The cells
were harvested and washed by centrifugation in phosphate-
buffered saline (PBS) and frozen in this condition at −70 °C.
Upon thawing, the cells were sonicated on a Heat Systems
Sonifier equipped with a macrotip for 2–3 cycles on wet ice for
30 s per cycle.
To prepare Orf5, the cell sonicate was centrifuged at
85,000×g for 1 h. Supernatant fluid (clear) was recovered and
placed on ice. Pellets were discarded. The supernatant contain-
ing the Orf5 was then loaded onto a 44 cm×3 cm Sephadex
G200 column at 4 °C equilibrated in PBS. Fractions containing
the bulk of the 42 kDa protein Orf5 protein eluted at or near the
void volume of the column. They were collected, pooled and
analyzed by SDS–PAGE.
The pooled Sephadex fractions were dialyzed overnight
against 10 mM Tris–HCl (pH 8.0) and applied to a preparative
MonoQ ion exchange resin in 10 mM Tris–HCl (pH 8.0) and
eluted in a 0.0–1.0 M NaCl gradient in the same buffer. The
Orf5 protein was eluted at approximately 250 mM NaCl. After
elution it was dialyzed against 10 mM Tris–HCl (pH 8.0)
Estimation of protein concentration of purified 42 kDa
protein was made using absorption at 280 nm (1.0 O.D. at
280 nm ∼1 mg/ml).
Quantitative gel scanning
Coomassie Blue-stained SDS polyacrylamide gels of
purified Orf5 protein were scanned in a Molecular Dynamics
Scanner using ImageQuant 5.0 software. Pixel densities of the
individual 42 and 35 kDa bands were obtained from the linear
portion of data set, and the pixel density of each band was
compared to the total pixel density of the two proteins
combined, to obtain the relative amount of each present in the
purified protein fraction.
Size exclusion chromatography of the bacterially expressed
42 kDa Orf5 gene product
The Orf5 gene product was subjected to size exclusion
chromatography on a calibrated Pharmacia Superdex 200 co-
lumn, in 50 mM sodium phosphate buffer (pH 7.4) containing
150 mM NaCl, by injecting approximately 100 μg protein in
100 μl into the column at a flow rate of 0.4 ml/min at 4 °C.
Calibration standards included thyroglobulin (660 kDa), apo-
ferretin (440 kDa), β-amylase (200 kDa), alcohol dehydrogen-
ase (150 kDa), bovine serum albumin (67 kDa) and carbonic
anhydrase (29 kDa).
Assembly of particles and agarose gel electrophoresis
GTA particles were assembled from purified 42 kDa Orf5
protein (1.2 mg/ml) in 10 mM Tris–HCl (pH 8.0) at 22 °C for 1
h in the presence or absence of 10 mM CaCl2, with or without
4% PEG 8000. After incubation samples were mixed with 1/10
volume of 60% glycerol with 0.001% bromophenol blue.
Electrophoretic separation of the unassembled or assembled
Orf5 protein was performed in a 1 mm thick vertical 1% agarose
gel made up in 40 mM Tris–HCl (pH 8.0) with 1 mM CaCl2at
40 V for 2–3 h at 4 °C.
Transmission electron microscopy (TEM) of negatively stained
For routine analysis, portions of the incubation were diluted
10- to 50-fold with 20 mM Tris–HCl (pH 8.0) and were
deposited onto carbon-Formvar-coated copper electron micro-
scope grids, washed with 20 mM Tris–HCl (pH 8.0), stained
with 1% uranyl acetate, air dried and photographed in a Philips
400 T electron microscope operated at 80 keV.
Scanning transmission electron microscopy (STEM)
Grids for the STEM were prepared by the wet film
technique as described previously (Wall et al., 1998; Wall and
Simon, 2001). Briefly, 2.3 mm titanium grids, coated with a
thick holey film, are placed on a floating thin (2–3 nm)
carbon film prepared by ultra-high vacuum evaporation onto
freshly cleaved rock salt. The grids are picked up one at a
time such that a thin layer of liquid is retained. They are
washed extensively (by washing and wicking). TMV (tobacco
mosaic virus, both a qualitative and a quantitative control) is
allowed to adsorb to the carbon film for 1 min. After further
washings and wickings, the sample is allowed to adsorb for
1 min. After additional washes, ending with a volatile buffer,
the grid is blotted to a very thin layer of liquid and plunged
into liquid nitrogen slush. Six grids are transferred to an ion-
pumped freeze drier, freeze-dried overnight, and transferred
under vacuum to the STEM.
The STEM is operated in a dark field mode. The scattered
electrons are collected in two annular detectors from each pixel
and the number of scattered electrons (in each pixel) is directly
proportional to the mass thickness in that pixel. By summing the
scattered electrons over a particle and subtracting the back-
ground from the thin carbon film, the mass of the particle can be
GTA particles were imaged by the STEM at the Brookhaven
National Laboratory as described in more detail at http://www.
Ultrathin sectioning of GTA proheads
GTA particles were assembled in the presence of 10 mM
CaCl2 as above and were pelleted from solution by
centrifugation at 100,000×g for 2 h at 4 °C. Supernatant
fluid was carefully removed and a freshly prepared solution of
50 mM MOPS buffer (pH 8.0) containing 2.5% glutaraldehyde
was added at room temperature to cover the clear pellets.
These were incubated at 4 °C for at least 24 h. The pellet was
delivered to the UVA Advanced Microscopy Facility for
further processing at room temperature. The fixed pellet was
washed in 0.1 M phosphate buffer (pH 7.4), post-fixed for
100 A.J. Spano et al. / Virology 364 (2007) 95–102
30 min in 1.0% (w/v) osmium tetroxide, dehydrated in a
graded series of ethanol and infiltrated with epoxy resin which
was subsequently polymerized at 60 °C for 48 h. Ultrathin
sections (60–65 nm) were cut on a Leica Ultracut UCT
ultramicrotome, contrast-stained with uranyl acetate and lead
citrate according to routine procedures and examined in a
JEOL 1230 transmission electron microscope. Digital images
were acquired using a SIA-L3C digital camera system
(Scientific Instruments and Applications, Inc.).
Sedimentation velocity experiments were performed in
40 mM Tris–HCl (pH 8.0) and 150 mM NaCl, with 0.25, 0.50
and 1.0 O.D. (at 280 nm) protein at 3.6 °C, 42 K in Beckman
style centerpieces, with data collected at the absorbance peak of
278 nm at an interval of 0.002 cm with one flash of the lamp per
point. Sednterpwas usedtocalculatetheappropriate proteinand
hydrodynamic parameters used to interpret and analyze the data
(MW of monomer=42,170.74, vbar=0.7366, Ext at
278 nm = 1.005, buffer density = 1.00734 g/ml,
viscosity=1.5870 cp (Laue et al., 1992)). Data were analyzed
with DCDT+2 to produce g(s) distributions and demonstrated a
concentration-dependent shift in boundary position (Philo,
2006). The g(s) data revealed evidence of a raised fast region
in the boundary consistent with<2% aggregation in the lowest
concentrations. The data were also normalized, divided by Abs,
to verify the concentration-dependent shift in position. These
data were further analyzed with SEDANAL by direct boundary
fitting methods (Stafford and Sherwood, 2004). The data can be
described by a non-interacting two species model (s1=3.47 s,
s2=4.85 s) where the fraction of the fast species increases with
concentration from 60 to 80%, or a reversible reaction between
an intermediate and pentamer with a midpoint of ∼6 mM total
To help understand these results and to verify the size of the
hypothesized species, sedimentation equilibrium experiment
was performed in six channel centerpieces. Three loading
concentrations were run (3.6 °C) at two speeds, 8 K and 16 K,
and data were collected at both 247 nm and 278 nm during the
run to verify attainment of equilibrium, ∼47 h and ∼40 h,
respectively. The 12 data sets were analyzed globally with
SEDANAL to a two species model to determine what species
were present in the solutions. The best fit to the equilibrium data
reveals a mixture of trimer (137,678<110,819, 143,131>) and
pentamer (196,538<180,772, 214,158>) where the error bars
are 95% confidence intervals and correspond to two standard
deviations. Monomer–trimer–pentamer models also fit the data
(not shown) but with extremely large K values consistent with
very tight complexes and no free monomer.
Edman degradation sequencing
Edman degradation sequence analysis was performed on
∼35 kDa Orf5 protein transferred to an Immobilon Psq
membrane at the Tufts Core facility at the Tufts Medical School,
Sequence alignments and structure analysis
version 3.7 (Schwede et al., 2003). Sequence alignments were
performed using ClustalW (http://www.ebi.ac.uk/clustalw/)
(Thompson et al., 1994).
The support of this work by Air Force Office of Scientific
Research, Office of Naval Research through a Naval Research
Laboratory base program, and Defense Advance Research
Project Agency is gratefully acknowledged. The Brookhaven
National Laboratory STEM is an NIH Supported Resource
Center, National Institutes of Health 5 P41 EB2181, with
additional support provided by Department of Energy, and the
Office of Biological and Environmental Research. We would
like to thank Drs. Matthew Harvey (University of Virginia)
and Dr. Allan Moose (University of Wisconsin, Eau Claire)
for helpful discussions concerning the geometry of polyhedra.
We thank Jan Redick for her expert technical assistance in the
preparation of the thin sections. The AUC work was done in
the UMMC Analytical Ultracentrifuge Facility by J. J.
Blum,A.S.,Soto,C.M., Wilson,C.D.,Cole,J.D., Kim,M., Gnade,B.,Chatterji,
A., Ochoa, W.F., Lin, T.W., Johnson, J.E., Ratna, B.R., 2004. Cowpea
mosaic virus as a scaffold for 3-D patterning of gold nanoparticles. Nano
Lett. 4 (5), 867–870.
Blum, A.S., Soto, C.M., Wilson, C.D., Brower, T.L., Pollack, S.K., Schull, T.L.,
Chatterji,A., Lin, T.W., Johnson, J.E., Amsinck,C., Franzon, P., Shashidhar,
R., Ratna, B.R., 2005. An engineered virus as a scaffold for three-
dimensional self-assembly on the nanoscale. Small 1 (7), 702–706.
Brussaard, C.P., 2004. Viral control of phytoplankton populations–a review.
J. Eukaryot. Microbiol. 51 (2), 125–138.
Danovaro, R., Armeni, M., Corinaldesi, C., Mei, M.L., 2003. Viruses and
marine pollution. Mar. Pollut. Bull. 46 (3), 301–304.
Donohue, T.J., Kaplan, S., 1991. Genetic techniques in rhodospirillaceae.
Methods Enzymol. 204, 459–485.
Duda, R.L., Hempel, J., Michel, H., Shabanowitz, J., Hunt, D., Hendrix, R.W.,
1995. Structural transitions during bacteriophage HK97 head assembly. J.
Mol. Biol. 247 (4), 618–635.
Earnshaw, W., King, J., 1978. Structure of phage P22 coat protein aggregates
formed in the absence of the scaffolding protein. J. Mol. Biol. 126 (4),
Helgstrand, C., Wikoff, W.R., Duda, R.L., Hendrix, R.W., Johnson, J.E., Liljas,
L., 2003. The refined structure of a protein catenane: the HK97
bacteriophage capsid at 3.44 A resolution. J. Mol. Biol. 334 (5), 885–899.
Hendrie, P.C., Russell, D.W., 2005. Gene targeting with viral vectors.Mol. Ther.
12 (1), 9–17.
Hendrix, R.W., 2005. Bacteriophage HK97: assembly of the capsid and
evolutionary connections. Adv. Virus Res. 64, 1–14.
Hochman, A., 1997. Programmed cell death in prokaryotes. Crit. Rev.
Microbiol. 23 (3), 207–214.
Howatson, A.F., Kemp, C.L., 1975. The structure of tubular head forms of
bacteriophage lambda; relation to the capsid structure of petit lambda and
normal lambda heads. Virology 67 (1), 80–84.
Lang, A.S., Beatty, J.T., 2000. Genetic analysis of a bacterial genetic exchange
element: the gene transfer agent of Rhodobacter capsulatus. Proc. Natl.
Acad. Sci. U.S.A. 97 (2), 859–864.
Lang, A.S., Beatty, J.T., 2001. The gene transfer agent of Rhodobacter
101A.J. Spano et al. / Virology 364 (2007) 95–102
capsulatus and “constitutive transduction” in prokaryotes. Arch. Microbiol.
175 (4), 241–249.
Lang, A.S., Beatty, J.T., 2002. A bacterial signal transduction system controls
genetic exchange and motility. J. Bacteriol. 184 (4), 913–918.
Lang, A.S., Taylor, T.A., Beatty, J.T., 2002. Evolutionary implications of
phylogenetic analyses of the gene transfer agent (GTA) of Rhodobacter
capsulatus. J. Mol. Evol. 55 (5), 534–543.
Laue, T., Shah, B., Ridgeway, T., Pelletier, S., 1992. Computer-aided
Interpretation of analytical sedimentation data for proteins. In: Harding, S.,
Rowe, A., Horton, J. (Eds.), Analytical Ultracentrifugation in Biochemistry
and Polymer Science. Royal Society of Chemistry, Cambridge.
Li, Y., Conway, J.F., Cheng, N., Steven, A.C., Hendrix, R.W., Duda, R.L., 2005.
Control of virus assembly: HK97 “Whiffleball” mutant capsids without
pentons. J. Mol. Biol. 348 (1), 167–182.
Marrs, B., 1974. Genetic recombination in Rhodopseudomonas capsulata. Proc.
Natl. Acad. Sci. U.S.A. 71 (3), 971–973.
Marrs, B., 1978. Mutations and genetic manipulations as probes of bacterial
photosynthesis. In: Sanadi, D., Vernon, L. (Eds.), Current Topics in
Bioenergetics, vol. 8. Academic Press, New York.
Marrs, B.L., 2002. The early history of the genetics of photosynthetic bacteria: a
personal account. Photosynth. Res. 73 (1–3), 55–58.
Merzlyak, A., Lee, S.W., 2006. Phage as templates for hybrid materials and
mediators for nanomaterial synthesis. Curr. Opin. Chem. Biol. 10 (3),
Munn, C.B., 2006.Viruses as pathogensof marine organisms—Frombacteria to
whales. J. Mar. Biol. Assoc. U. K. 86 (3), 453–467.
Normand, N., Valamanesh, F., Savoldelli, M., Mascarelli, F., BenEzra, D.,
Courtois, Y., Behar-Cohen, F., 2005. VP22 light controlled delivery of
oligonucleotides to ocular cells in vitro and in vivo. Mol. Vis. 11, 184–191.
Philo, J.S., 2006. Improved methods for fitting sedimentation coefficient
distributions derived by time-derivative techniques. Anal. Biochem. 354 (2),
Ross, P.D., Cheng, N., Conway, J.F., Firek, B.A., Hendrix, R.W., Duda, R.L.,
Steven, A.C., 2005. Crosslinking renders bacteriophage HK97 capsid
maturation irreversible and effects an essential stabilization. EMBO J. 24
Schaefer, A.L., Taylor, T.A., Beatty, J.T., Greenberg, E.P., 2002. Long-chain
acyl-homoserine lactone quorum-sensing regulation of Rhodobacter capsu-
latus gene transfer agent production. J. Bacteriol. 184 (23), 6515–6521.
Schwede, T., Kopp, J., Guex, N., Peitsch, M.C., 2003. SWISS-MODEL: an
automated protein homology-modeling server. Nucleic Acids Res. 31 (13),
Scolnik, P.A., Walker, M.A., Marrs, B.L., 1980. Biosynthesis of carotenoids
derived from neurosporene in Rhodopseudomonas capsulata. J. Biol. Chem.
255 (6), 2427–2432.
Seth, P., 2005. Vector-mediated cancer gene therapy: an overview. Cancer Biol.
Ther. 4 (5), 512–517.
Singh, P., Gonzalez, M.J., Manchester, M., 2006. Viruses and their uses in
nanotechnology. Drug Dev. Res. 67, 23–41.
Slocik, J.M., Naik, R.R., Stone, M.O., Wright, D.W., 2005. Viral templates for
gold nanoparticle synthesis. J. Mater. Chem. 15 (7), 749–753.
Solioz, M., Marrs, B., 1977. The gene transfer agent of Rhodopseudomonas
capsulata. Purification and characterization of its nucleic acid. Arch.
Biochem. Biophys. 181 (1), 300–307.
Solioz, M., Yen, H.C., Marris, B., 1975. Release and uptake of gene transfer
agent by Rhodopseudomonas capsulata. J. Bacteriol. 123 (2), 651–657.
Soto, C.M., Blum, A.S., Vora, G.J., Lebedev, N., Meador, C.E., Won, A.P.,
Chatterji, A., Johnson, J.E., Ratna, B.R., 2006. Fluorescent signal
amplification of carbocyanine dyes using engineered viral nanoparticles.
J. Am. Chem. Soc. 128 (15), 5184–5189.
Souza, G.R., Christianson, D.R., Staquicini, F.I., Ozawa, M.G., Snyder, E.Y.,
Sidman, R.L., Miller, J.H., Arap, W., Pasqualini, R., 2006. Networks of gold
nanoparticles and bacteriophage as biological sensors and cell-targeting
agents. Proc. Natl. Acad. Sci. U.S.A. 103 (5), 1215–1220.
Stafford, W.F., Sherwood, P.J., 2004. Analysis of heterologous interacting
systems by sedimentation velocity: curve fitting algorithms for estimation of
sedimentation coefficients, equilibrium and kinetic constants. Biophys.
Chem. 108 (1–3), 231–243.
Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W: improving
the sensitivity of progressive multiple sequence alignment through sequence
weighting, position-specific gap penalties and weight matrix choice. Nucleic
Acids Res. 22 (22), 4673–4680.
Tomanin, R., Scarpa, M., 2004. Why do we need new gene therapy viral
vectors? Characteristics, limitations and future perspectives of viral vector
transduction. Curr. Gene Ther. 4 (4), 357–372.
Wall, J.D., Braddock, K., 1984. Mapping of Rhodopseudomonas capsulata nif
genes. J. Bacteriol. 158 (2), 404–410.
Wall, J.S., Simon, M.N., 2001. Scanning transmission electron microscopy of
DNA–protein complexes. Methods Mol. Biol. 148, 589–601.
Wall, J.D., Weaver, P.F., Gest, H., 1975a. Gene transfer agents, bacteriophages,
and bacteriocins of Rhodopseudomonas capsulata. Arch. Microbiol. 105
Wall, J.D., Weaver, P.F., Gest, H., 1975b. Genetic transfer of nitrogenase-
hydrogenase activity in Rhodopseudomonas capsulata. Nature 258 (5536),
Wall, J.D., Love, J., Quinn, S.P., 1984. Spontaneous Nif-mutants of Rhodop-
seudomonas capsulata. J. Bacteriol. 159 (2), 652–657.
Wall, J.S., Hainfeld, J.F., Simon, M.N., 1998. Scanning transmission electron
microscopy of nuclear structures. Methods Cell Biol. 53, 139–164.
Wang, Q., Lin, T., Johnson, J.E., Finn, M.G., 2002a. Natural supramolecular
building blocks. Cysteine-added mutants of cowpea mosaic virus. Chem.
Biol. 9 (7), 813–819.
Wang, Q., Lin, T., Tang, L., Johnson, J.E., Finn, M.G., 2002b. Icosahedral virus
particles as addressable nanoscale building blocks. Angew. Chem., Int. Ed.
Engl. 41 (3), 459–462.
Wikoff, W.R., Liljas, L., Duda, R.L., Tsuruta, H., Hendrix, R.W., Johnson, J.E.,
2000. Topologically linked protein rings in the bacteriophage HK97 capsid.
Science 289 (5487), 2129–2133.
Xie, Z., Hendrix, R.W., 1995. Assembly in vitro of bacteriophage HK97
proheads. J. Mol. Biol. 253 (1), 74–85.
Yen, H.C., Marrs, B., 1976. Map of genes for carotenoid and bacteriochlo-
rophyll biosynthesis in Rhodopseudomonas capsulata. J. Bacteriol. 126 (2),
Yen, H.C., Hu, N.T., Marrs, B.L., 1979. Characterization of the gene transfer
agent made by an overproducer mutant of Rhodopseudomonas capsulata.
J. Mol. Biol. 131 (2), 157–168.
Zhang, S., 2003. Fabrication of novel biomaterials through molecular self-
assembly. Nat. Biotechnol. 21 (10), 1171–1178.
102 A.J. Spano et al. / Virology 364 (2007) 95–102