Time-resolved fluorescence of the bacteriophage T4 capsid protein gp23

www.elsevier.com/locate/jphotobiol

Journal of Photochemistry and Photobiology B: Biology 78 (2005) 53–60

Time-resolved fluorescence of the bacteriophageT4 capsid protein gp23

Aike Stortelder a, Joost B. Buijs a, Jaap Bulthuis b, Saskia M. van der Vies c,Cees Gooijer a, Gert van der Zwan a,*

a Department of Analytical Chemistry and Applied Spectroscopy, Faculty of Sciences, Vrije Universiteit,

Laser Centre VU, De Boelelaan 1083, 1081 HV Amsterdam, The Netherlandsb Department of Physical Chemistry, Faculty of Sciences, Vrije Universiteit, Laser Centre VU,

De Boelelaan 1083, 1081 HV Amsterdam, The Netherlandsc Department of Biochemistry and Molecular Biology, Faculty of Sciences, Vrije Universiteit, Laser Centre VU,

De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands

Received 18 March 2004; received in revised form 10 August 2004; accepted 13 September 2004

Available online 11 November 2004

Abstract

The time-resolved fluorescence properties of the bacteriophage T4 capsid protein gp23 are investigated. The structural charac-

teristics of this protein are largely unknown and can be probed by recording time-resolved and decay-associated fluorescence spectra

and intensity decay curves using a 200 ps-gated intensified CCD-camera. Spectral and decay data are recorded simultaneously,

which makes data acquisition fast compared to time-correlated single-photon counting. A red-shift of the emission maximum within

the first nanosecond of decay is observed, which can be explained by the different decay-associated spectra of fluorescence lifetimes

of the protein in combination with dipolar relaxation. In addition, iodide quenching experiments are performed, to study the degree

of exposure of the various tryptophan residues. A model for the origin of the observed lifetimes of 0.032 ± 0.003, 0.39 ± 0.06,

2.1 ± 0.1 and 6.8 ± 0.8 ns is presented: the 32 ps lifetime can be assigned to the emission of a buried tryptophan residue, the 0.4

and 2.1 ns lifetimes to two partly buried residues, and the 6.8 ns lifetime to a single tryptophan outside the bulk of the folded gp23.

� 2004 Elsevier B.V. All rights reserved.

Keywords: Time-resolved emission spectra; Fluorescence lifetimes; Decay-associated spectra; Fast-gated CCD; gp23; Gobal analysis; Fluorescence

quenching

1. Introduction

gp23 is the major capsid protein of the bacteriophage

T4. It consists of 521 amino acids, including four trypt-

ophan residues at positions 13, 247, 309 and 345. In

vivo, gp23 is folded with assistance of the molecular

chaperone complex GroEL–gp31 [1]. Chaperones are

proteins that are indispensable for folding of many pro-

1011-1344/$ - see front matter � 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.jphotobiol.2004.09.007

* Corresponding author. Tel.: +31 20 444 7635; fax: +31 20 444

7543.

E-mail address: [email protected] (G. van der Zwan).

teins. They stabilize the newly synthesized amino acid

chain and thus prevent formation of a misfolded pro-tein. There are several chaperone systems that operate

in different compartments of the cell [2]. The GroEL–

GroES combination is present in bacteria. The bacteri-

ophage T4 uses the GroEL–GroES chaperones to fold

most of its structural elements. However, the main cap-

sid protein gp23 needs a special co-chaperone with

GroEL instead of GroES. This co-chaperone, denoted

gp31, is encoded by T4 itself [1]. In the T4-head, 120hexamers of gp23 are present, thus a total 720 gp23 mol-

ecules [3]. This interesting folding pathway, which can

54 A. Stortelder et al. / Journal of Photochemistry and Photobiology B: Biology 78 (2005) 53–60

also be studied in vitro, is not yet well understood; the

same holds for most of the secondary and tertiary struc-

ture of gp23. In solution, gp23 can be present in several

forms: unfolded, as a monomer, as a hexamer or as lar-

ger ordered aggregates, so-called polyheads.

Conformational changes in proteins can be moni-tored using an intrinsic fluorescent probe, i.e. the amino

acid tryptophan [4–6]. The emission characteristics of

this amino acid are dependent on the microenvironment

[7]. Both lifetimes and emission wavelengths may change

when the conformation of (parts of) the protein changes.

In a polar environment, the emission maximum of trypt-

ophan is at about 350 nm. It shifts to shorter wave-

lengths of up to 305 nm in apolar surroundings, forexample if the tryptophan is located inside a protein

[8,9].

Tryptophan in water shows three fluorescence life-

times, around 0.5, 3 and 7–9 ns [10,11]. Under certain

conditions, a shorter lifetime of about 13 ps has also

been observed [12]. The emission spectrum associated

with the longer lifetime is shifted to longer wavelengths

by about 30 nm, compared to that of the shorter lifetime[13]. An explanation for this wavelength dependence and

for the multi-exponentiality of the decay is provided by

the tryptophan rotamer model [14], in which the wave-

length and lifetime of the indole emission are assumed

to be dependent on the distance between the amino

group and the indole ring [15,16]. In proteins, where tryp-

tophan residues experience a much more rigid environ-

ment because of the bulky amino acid chain, there stillseems to be enough space for the tryptophan residue

to adopt various conformations [17]. This leads to mul-

tiple lifetimes even in single tryptophan proteins [4]. Of

course, when multiple tryptophan residues are present,

the lifetime distribution becomes increasingly compli-

cated, and separate lifetimes cannot always be resolved

[18]. Lifetimes might change quite dramatically due to

the different quenching characteristics of amino acidssuch as tyrosine, histide or cysteine that surround the

tryptophan residue [19]. Also, lifetimes are influenced

by solvent that penetrates or is expelled from the protein

when its conformation changes. External quenchers

such as iodide and acrylamide are often used to probe

the surface residues of proteins [20,21], and can provide

information on the degree of exposure of the different

residues. Acrylamide is a neutral quencher, which maypenetrate into the protein bulk and also quench buried

residues to some extent. Iodide is negatively charged,

and will not enter the non-polar interior of the folded

protein and quenches only the exposed residues. An

electrostatic effect due to charged amino acids next to

tryptophan residues leading to over- or under-estima-

tion of the degree of quenching may be avoided by keep-

ing the ionic strength constant.gp23 contains four tryptophan residues, so that in

principle structural changes in this protein can be stud-

ied by monitoring the intrinsic tryptophan fluorescence

emission characteristics. Since three of these residues

are located close together in the amino acid sequence,

it seems not unreasonable to assume that they experi-

ence similar surroundings. The fourth tryptophan is lo-

cated in a part of the protein that can be cleaved off veryeasily [22], and is likely to be located outside of the bulk

of the folded protein.

The time-resolved fluorescence properties of both

monomeric and hexameric gp23 were studied using a re-

cently commercialized fast-gated charge-coupled device

(CCD) camera [10]. Until recently, available CCD cam-

eras were too slow for this purpose. With this 200 ps-

gated ICCD-camera, it was possible to directly recordtime-resolved emission spectra (TRES) with very good

spectral resolution and to resolve lifetimes on a sub-

nanosecond time-scale. Quenching of tryptophan fluo-

rescence by iodide was used to determine the degree of

exposure of the different lifetimes and, hence, to help

assign the observed lifetimes to the four tryptophan

residues.

2. Experimental

2.1. Setup

The setup consisted of a laser system combined with

an intensified CCD (ICCD)-camera detection and time-

resolved photon counting detection. The main part ofthe laser system is a Mira 900-P (Coherent, Santa Clara

CA, USA) laser, which emits 3 ps pulses at a repetition

rate of 76 MHz, pumped by an Innova-300 argon ion

laser (Coherent). The Mira 900-P is a mode-locked tita-

nium–sapphire laser, tunable from approximately 700

to 1000 nm. The laser emission is led through a pulse

picker (APE PulseSelect, Berlin, Germany), which re-

duced the repetition rate to 3.8 MHz to avoid doubleexcitation of molecules. A frequency tripler (Oplaz

Technologies fs-tripler, Chatsworth CA, USA) was

used to provide the used output wavelength of 291

nm. As the laser power was about 100 lW no signifi-

cant heating or photodegradation of the sample was ex-

pected. This light was used to excite the protein sample

in a cell with 1 cm pathlength (type 104F, Hellma

GmbH & Co KG, Mullheim, Germany). Fluorescenceemission was collected at 90� and dispersed by a spect-

rograph (TVC JarrellAsh Monospec 18, Grand Junc-

tion CO, USA). A WG-320 filter (Schott, Mainz,

Germany) was used to filter out laser scatter. Spectra

were recorded from 320 to 450 nm with an ICCD-

camera with ultrafast gate times (LaVision, Gottingen,

Germany). This system contains an image intensifier

that provides very good sensitivity and acts as an extre-mely fast shutter with a shortest gate time of 200 ps

FWHM. Validation of this camera using tryptophan

A. Stortelder et al. / Journal of Photochemistry and Photobiology B: Biology 78 (2005) 53–60 55

fluorescence as a model system has been described in

[10]. Detection was triggered by a reflection of the laser

beam on a photodiode. The intensity decay was re-

corded with 25 ps steps over a range of 10 ns, so per

decay curve 400 data points were collected. Image col-

lection and image processing was performed under con-trol of DaVis software v6.2 (LaVision).

For time-correlated single-photon counting (TCSPC)

experiments an SPC-630 (Becker & Hickl GmbH, Ber-

lin, Germany) system with a time-resolution of about

15 ps was used. A laser pulse focused on a photodiode

provided the synchronization signal. Fluorescence emis-

sion was detected by a photomultiplier tube after being

dispersed by a monochromator (TVC JarrellAsh Mono-spec 18, Grand Junction CO, USA). Decays were re-

corded at wavelengths between 330 and 440 nm in 10

nm steps. Quenching experiments were performed using

TCSPC detection at a single wavelength (342 nm).

2.2. Sample preparation

Samples of 1 lM purified gp23 were prepared (from96.6 lM stock solution) by dilution in a buffer contain-

ing 50 mM Tris pH 7.4, 10 mM KCl, 10 mM MgCl2 and

0.01% Tween 20. Experiments were performed at room

temperature and 4 �C. From size-exclusion chromatog-

raphy experiments, it was concluded that at room tem-

perature gp23 is mainly present in hexameric form,

whereas at 4 �C it is in monomeric form [23]. This was

confirmed by steady-state fluorescence anisotropy exper-iments: a higher anisotropy was found for gp23 at room

temperature than at 4 �C, indicating a slower rotation

and thus a larger complex.

For the quenching experiments, KI was added to ob-

tain final concentrations from 0.0 to 0.5 M. Appropriate

amounts of KCl were added to keep the total ionic

strength at 0.5 M. To avoid formation of I2 10 mM so-

dium isothiocyanate was added to the solution.

2.3. Data analysis

Fluorescence decay curves were analyzed using a glo-

bal fit procedure based on the Levenberg–Marquardt

algorithm. This procedure uses the gate function (ob-

tained by recording scattered laser-light) for deconvolu-

tion of the recorded decays and determination oflifetimes and wavelength dependent amplitudes. Life-

times are assumed to be constant for the decays in-

cluded, but with variable amplitudes. This assumption

is correct for proteins in many cases [24]. A maximum

number of four lifetimes is fitted simultaneously to the

array of experimental decays at different wavelengths,

to get a good fit. The goodness of fit was determined

on the basis of v2, residuals and correlation coefficients.A maximum of 12 decays were included in the global

lifetime analysis. The time-resolved fluorescence spectra

were plotted on a wavenumber scale and fitted to a com-

bination of Gaussian functions using Origin 7.0 (Origin-

Lab, Northampton MA, USA). This is justified, since

the shape of a fluorescence spectrum is dominated by

inhomogeneous band broadening [25,26]. The purpose

of this fit was simply to determine the emission maxima,and no physical meaning is ascribed to the individual

components.

3. Results

The fluorescence emission of a 1 lM solution of both

the monomeric and hexameric form of bacteriophage T4capsid protein gp23 was recorded using the fast-gated

ICCD-camera. Emission spectra and intensity decay

curves were measured simultaneously using a 300 ps

gate width. The three-dimensional plot in Fig. 1 illus-

trates the relation between time and spectral data: fluo-

rescence emission spectra along the x-axis, decay curves

along the y-axis.

Changes in the emission spectra of gp23 were moni-tored during the decay of fluorescence, and lifetimes

were determined. In addition, iodide quenching experi-

ments were performed to determine the degree of solvent

exposure of the tryptophan residues. An excitation

wavelength of 291 nm was chosen to avoid tyrosine

interference.

3.1. Time-resolved emission spectra

Fig. 1 shows TRES of hexameric gp23 along the x-

axis. The emission spectrum shifts to the red on a nano-

second time-scale. This is more clearly depicted in Fig. 2,

where the position of the emission maximum in time is

plotted. A large shift in the emission maximum is seen,

which starts already in the first nanosecond of decay.

A red-shift in the emission maximum of about 2 nmcan be seen even within the first 250 ps of decay (Fig.

2, left). This shift continues for another 2 nm in the de-

cay period up to 4 ns (Fig. 2, right). Dipolar relaxation

times of 1.0 ± 0.1 ns and 1.4 ± 0.2 ns were found for the

hexamer and the monomer, respectively. The somewhat

slower relaxation for the monomer is probably due to

the lower temperature at which the experiments were

done.The emission spectra were fitted to Gaussian func-

tions to determine the exact emission maximum. An

adequate fit of the spectra (320–450 nm) was obtained

with two Gaussian functions. Fitting with other func-

tions, such as stretched Gaussians, did not significantly

alter the results.

The TRES of gp23 were recorded at a concentration

of 1 lM, but a 10-fold lower concentration could also beused. However, under such conditions acquisition times

were very long, which gave rise to problems with laser

Fig. 2. Shift of the hexamer emission maximum with time (left: 0–250 ps, right: 250–4500 ps).

Fig. 1. Time-resolved emission of 1 lM of hexameric gp23 recorded with the fast-gated camera using a gate-width of 300 ps. On the time axis,

intensity decays can be seen; on the wavelength axis, emission spectra are shown.

56 A. Stortelder et al. / Journal of Photochemistry and Photobiology B: Biology 78 (2005) 53–60

and sample stability. The same problems occur when

using the minimum gate width of 200 ps (instead of

300 ps).

3.2. Fluorescence lifetimes

Along with TRES, intensity decay curves at different

wavelengths were recorded (seen in Fig. 1 along the y-

axis). Twelve decay curves were considered over the

range from 320 to 430 nm (each corresponding to 50

pixels, i.e. a spectral bandwidth of 7 nm). These 12 decay

curves were globally fitted to a sum of exponentials,using the procedures described in Section 2.3:

IðtÞ ¼X

iAiðkÞe

�tsi : ð1Þ

This approach yielded a good fit with four lifetimes

for data recorded with the ICCD-camera: for hexameric

gp23, these lifetimes are 0.032 ± 0.003, 0.39 ± 0.06,

2.1 ± 0.1 and 6.8 ± 0.8 ns, whereas for monomericgp23 we find 0.028 ± 0.005, 0.37 ± 0.07, 2.3 ± 0.1 and

5.6 ± 0.6 ns. The finding of the shortest (�30 ps) lifetime

was somewhat surprising: because a time-step of 25 ps

was used, this lifetime is based on a small number of

data points. However, the fit of the data improved sig-

nificantly when it was included. In addition, this 30 ps

lifetime, as well as the 0.4 and 2.1 ns lifetimes were, quite

satisfying, the same lifetimes as we obtained fromTCSPC experiments, occasionally performed to check

results found with the camera. The longest lifetimes

are for both hexameric and monomeric gp23 somewhat

A. Stortelder et al. / Journal of Photochemistry and Photobiology B: Biology 78 (2005) 53–60 57

different than that observed with TCSPC, in which we

found the longest lifetimes to be 5.3 and 6.1 ns, respec-

tively. This deviation may be attributed to the fact that

the decay period recorded with the camera is not longer

than 12 ns. For longer lifetimes this implies that the

intensity has not yet completely decayed to zero, whichmay cause the tail to be fitted less accurately. In the

TCSPC experiments such problems will not play a role,

since 50 ns periods of decay are recorded.

3.3. Decay-associated spectra

A decay-associated spectrum represents the emission

spectrum originating from one lifetime component.Although it is tempting to ascribe such a spectrum to

a single fluorophore, it can also come from a combina-

tion of fluorophores having quite similar lifetimes.

Fig. 3. Decay-associated data of the four lifetimes of gp23 in monomeric

amplitudes obtained from global fits, relative to the normalized maximum in

Fig. 4. Stern–Volmer graphs of the 5–7 ns (e), 2 ns (h), and 0.4 ns (s) lifetime

hexamer (right). s = lifetime without quencher,KSV = quenching constant, kq =

Using Eq. (2), decay-associated spectra (DAS) can be

constructed by distributing the total intensity at a cer-

tain wavelength among the four lifetimes of gp23

according to their relative amplitudes at that wavelength

as found by global analysis

IsjðkÞ ¼AjsjPisiAi

: ð2Þ

Interestingly, quite different emission spectra were

obtained for the four lifetimes (Fig. 3), with maxima

at 344 nm (0.03 ns), 339 nm (0.4 ns), 344 nm (2.1 ns),

and 348 nm (6.8 ns) for the hexamer, and 344 nm

(0.03 ns), 334 nm (0.4 ns), 339 nm (2.3 ns), and 346

nm (5.6 ns) for monomeric gp23.

Apparently, the spectra of the three shorter lifetimesare more blue than that of the longest lifetime. This

may partly be due to incomplete dipolar relaxation of

(left) and hexameric form (right). The DAS were constructed from

tensity of the total spectrum.

s of gp23 fluorescence quenched by iodide, for gp23 monomer (left) and

quenching rate, fa = fraction of total emission accessible to quenching.

58 A. Stortelder et al. / Journal of Photochemistry and Photobiology B: Biology 78 (2005) 53–60

the surroundings of the tryptophan residues [27]. This

would also explain the bluer DAS for the lifetimes in

monomeric gp23. Since the residues are likely to be more

exposed in the monomer, the spectra are expected to be

more red than in the hexamer. However, dipolar relaxa-

tion is slower at 4 �C, were the monomer is observed, sothe dipoles will not be completely relaxed and the spec-

tra will be less red-shifted. Noteworthy is that the DAS

of the 32 ps lifetime has a maximum at a longer wave-

length than the DAS corresponding to the 0.4 ns life-

time, indicating a more polar environment for this

residue.

3.4. Quenching with iodide

Changes in the fluorescence lifetimes of 1 lM solu-

tions of monomeric and hexameric gp23 due to addition

of an external quencher were studied using TCSPC

detection. Fig. 4 shows the Stern–Volmer plots of the

three longest lifetimes.

Large differences in quenching between the two con-

formations are seen for the 2 ns lifetime. The longestlifetime is also somewhat more quenched in the mono-

mer than in the hexamer, the change is not as much as

for the 2 ns lifetime. Quenching for the 0.4 ns lifetime

is similar in both forms. The quenching curves are slop-

ing downwards for hexameric gp23, showing that a sig-

nificant fraction of tryptophan emission will not be

quenched and thus comes from residues inaccessible to

iodide.

4. Discussion

The TRES showed a shift of about 4 nm in the first 4

ns of decay. This shift may be due to relaxation of the

protein environment around the tryptophan residues,

despite the fact that solvent relaxation around a fluoro-phore is usually a femto- to pico-second process. Only if

the solvent is very viscous will reorientation of the sol-

vent dipoles around the fluorophore be much slower

and therefore compatible with nanosecond experiments,

a situation that presumably applies for proteins [14]. In

proteins, the amino acid chain around the fluorescing

tryptophan residue behaves as a viscous solvent, and

will be slow in reorienting.For the observed lifetimes and DAS two possible

explanations are considered. The most obvious option

is to assume that each lifetime originates from one indi-

vidual tryptophan residue, an explanation in line with

the fact that four lifetimes were obtained

(0.032 ± 0.003, 0.39 ± 0.06, 2.1 ± 0.1 and 6.8 ± 0.8 ns

for hexameric gp23, 0.028 ± 0.005, 0.37 ± 0.07,

2.3 ± 0.1 and 5.6 ± 0.6 ns for monomeric gp23), whilegp23 contains four tryptophan residues. However, the

primary amino acid sequence of gp23 reveals three try-

ptophans located close together (at positions 247, 309

and 345) and might experience similar environments.

Thus, a 10- to 100-fold difference in fluorescence life-

times from these residues is not to be expected.

A second, more probable explanation can be pro-

posed. The longest (5–7 ns) lifetime may be attributedto the fourth tryptophan in gp23 at position 13, which

experiences a very different environment. It is located

in a part of the protein that is cleaved off after assembly

of the head-structure [19]. Therefore, it is probably lo-

cated in a flexible part of the chain outside the bulk of

the folded protein, which is not expected to change

much if monomers form a hexamer. This assumption

is confirmed by the quenching experiments, which showsimilar quenching for this lifetime for gp23 in mono-

meric and hexameric form. It is not quenched as much

as could be expected for an exposed tryptophan residue,

which probably means that there still is some degree of

folding in this tailing part of the protein, and since the

quenching rate is somewhat diferent in both cases, it

may also be partly shielded from the quencher in the

hexamer. This tryptophan residue will be more exposedto the solvent than the other tryptophan residues, and

hence will experience a more polar environment. This

is reflected in the DAS of the 5–7 ns lifetime, which

has a relatively red-shifted emission maximum, a shift

that will at least partly be due to spectral relaxation,

which is especially strong in polar surroundings [14].

If this interpretation is correct, the remaining life-

times of 30 ps, 0.4 ns and 2 ns originate from the com-bined emission of the three tryptophan residues on

positions 247, 309 and 345. The fact that the lifetimes

have similar emission maxima suggests that they arise

from fluorophores in the same environment. The relative

blue emission found for these lifetimes suggests a non-

polar environment, which is generally the case inside a

protein. Amino acids found in the vicinity of these try-

ptophans may quench the fluorescence [19], hence lead-ing to shorter fluorescence lifetimes than expected for

free tryptophan (or N-acetyl-tryptophan amide, which

is the actual residue [14]). This is also the reason why

the added intensity of the 30 ps, 0.4 ns and 2 ns DAS

is lower than three times that of the 5–7 ns DAS. The

assignment is confirmed, and can even be refined some-

what by taking into account the results of the quenching

experiments. The 2 ns lifetime is quenched to a large ex-tent in the monomer, whereas in the hexamer the

quenching is much less, indicating that this residue can

be found on the interaction surface of the different

monomers in a hexamer. The 0.4 ns lifetime is quenched

to similar extent in both forms of gp23, so its degree of

exposure is the same. This lifetime will originate from

residue(s) that are partially exposed, but are not on

the interaction surface of the monomers if they form ahexamer. The shortest (30 ps) lifetime is probably due

to a deeply buried tryptophan residue, quenched to a

A. Stortelder et al. / Journal of Photochemistry and Photobiology B: Biology 78 (2005) 53–60 59

large extent by close lying amino acids. This is sup-

ported by the DAS which show the same emission max-

imum for this lifetime in both forms of gp23, indicating

that its environment is the same in both cases. Its emis-

sion maximum is not as blue as could be expected, indi-

cating some kind of polar interaction with surroundingamino acids. Another indication that it is deeply buried

is that it is not quenched by iodide: the 30 ps lifetime had

to be included in fits of all decay curves, also those of

samples including 0.5 M iodide.

This assignment seems to indicate four different fluo-

rescence lifetimes for four different tryptophan residues.

However, as mentioned above, it is unlikely that exactly

one lifetime originates from one residue. Therefore, it islikely that each lifetime found is a mixture of lifetimes

from partly exposed residues that show multiple life-

times each, which cannot be mutually distinguished with

an exponential fit method.

Of course, it is obvious that additional fluorescence

experiments should be done to check the above assign-

ment. Such experiments could, for example, make use

of mutation of one (or several) tryptophan residues intophenylalanines, or be based on removing the tail con-

taining Trp-13.

5. Conclusions

Time-resolved fluorescence studies using a fast-gated

ICCD-camera provide adequate detailed information onthe bacteriophage T4 capsid protein gp23, using the

intrinsic fluorescence of its four tryptophan residues.

TRES show a shift of 4 nm within the first 4 ns of de-

cay. The observed shifts in emission maxima and the

red-shifted DAS were attributed to spectral relaxation

of the protein environment around the excited state try-

ptophan dipoles, with relaxation times of 1.0 ± 0.1 ns

and 1.4 ± 0.2 ns for the hexamer and the monomer,respectively.

The lifetimes obtained from global fitting procedures

are 0.032 ± 0.003, 0.4 ± 0.06, 2.1 ± 0.1 and 6.8 ± 0.8 ns

for the gp23 hexamer, and 0.028 ± 0.005, 0.37 ± 0.07,

2.3 ± 0.1 and 5.6 ± 0.6 ns for monomeric gp23. The 30

ps lifetime can be ascribed to a deeply buried tryptophan

residue; the 0.4 ns lifetime to a residue partly buried, and

not on the interaction surface of monomers in a hex-amer; the 2 ns lifetime to the combined emission of three

tryptophans that lie relatively close together in the ami-

no acid sequence of gp23; the 6.8 ns lifetime may be

attributed to the fourth tryptophan residue located far

away from the other three at position 13. This assign-

ment is only tentative. Further experiments, for instance

on single tryptophan mutants of gp23, are necessary to

achieve solid conclusions. However, the results obtainedso far are already useful for the study of the hexameriza-

tion equilibrium.

Acknowledgements

Purified gp23 was kindly provided by Patrick J. Bak-

kes and Els Kroezinga (Vrije Universiteit Amsterdam,

Department of Biochemistry and Molecular Biology).

References

[1] F.A. Eiserling, L.W. Black, Pathways in T4 morphogenesis, in:

J.D. Karam (Ed.), Molecular Biology of Bacteriophage T4,

ASM Press, 1994, pp. 209–212.

[2] F.U. Hartl, Molecular chaperones in cellular protein folding,

Nature 381 (1996) 571–580.

[3] N.H. Olson, M. Gingery, F.A. Eiserling, T.S. Baker, The structure

of isometric capsids of bacteriophage T4, Virology 279 (2001)

385–391.

[4] J.M. Beechem, L. Brand, Time-resolved fluorescence of proteins,

Annu. Rev. Biochem. 54 (1985) 43–71.

[5] M.R. Eftink, The use of fluorescence methods to monitor

unfolding transitions in proteins, Biochemistry Moscow 63

(1998) 276–284.

[6] Y. Engelborghs, Correlating protein structure and protein fluo-

rescence, J. Fluoresc. 13 (2003) 9–16.

[7] P. Callis, Molecular orbital theory of the 1Lb and the 1La states of

indole, J. Chem. Phys. 95 (1991) 4230–4240.

[8] J.T. Vivian, P. Callis, Mechanisms of tryptophan fluorescence

shifts in proteins, Biophys. J. 80 (2001) 2093–2109.

[9] J.R. Lakowicz, Protein fluorescence, Principles of Fluorescence

Spectroscopy, first ed., Plenum Press, New York, 1983, pp. 341–

379.

[10] A. Stortelder, J.B. Buijs, J. Bulthuis, C. Gooijer, G. van der

Zwan, A fast-gated intensified CCD-camera to record time-

resolved fluorescence spectra of tryptophan, Appl. Spectrosc. 58

(2004) 705–710.

[11] W.B. De Lauder, P. Wahl, pH dependence of the fluorescence

decay of tryptophan, Biochemistry 9 (1970) 2750–2754.

[12] O.F.A. Larsen, I.H.M. Van Stokkum, A. Pandit, R. Van

Grondelle, H. Van Amerongen, Ultrafast polarized fluorescence

measurements on tryptophan and a tryptophan-containing pep-

tide, J. Phys. Chem. B 107 (2003) 3080–3085.

[13] A.G. Szabo, D.M. Rayner, Fluorescence decay of tryptophan

conformers in aqueous solution, J. Am. Chem. Soc. 102 (1980)

554–563.

[14] J.R. Lakowicz, On spectral relaxation in proteins, Photochem.

Photobiol. 72 (2000) 421–437.

[15] Y. Engelborghs, The analysis of time resolved protein fluorescence

in multi-tryptophan proteins, Spectrochim. Acta 57 (2001) 2255–

2270.

[16] J.W. Petrich, M.C. Chang, D.B. McDonald, G.R. Fleming, On

the origin of nonexponential fluorescence decay in tryptophan and

its derivatives, J. Am. Chem. Soc. 105 (1983) 3824–3832.

[17] J.R. Lakowicz, B.P. Maliwal, H. Cherek, A. Balter, Rotational

freedom of tryptophan residues in proteins and peptides, Bio-

chemistry 22 (1983) 1741–1752.

[18] J.R. Alcala, E. Gratton, F.G. Prendergast, Fluorescence lifetime

distributions in proteins, Biophys. J. 51 (1987) 597–604.

[19] Y. Chen, M.D. Barkley, Toward understanding tryptophan

fluorescence in proteins, Biophys. J. 81 (1998) 1765–1775.

[20] J.R. Lakowicz, Quenching of fluorescence, Principles of Fluores-

cence Spectroscopy, first ed., Plenum Press, New York, 1983, pp.

257–295.

[21] M.R. Eftink, C.A. Ghiron, Exposure of tryptophanyl residues in

proteins. Quantitative determination by fluorescence quenching

studies, Biochemistry 15 (1976) 672–680.

60 A. Stortelder et al. / Journal of Photochemistry and Photobiology B: Biology 78 (2005) 53–60

[22] L.G. Aijrich, L.P. Kurochkina, V.V. Mesyanzhinov, Chaperonin-

mediated folding of bacteriophage T4 major capsid protein. II.

Production of gene product 23 deletion mutants, Biochemistry

Moscow 67 (2002) 815–821.

[23] P.J. Bakkes, personal communication.

[24] J.M. Beechem, Global analysis of biochemical and biophysical

data, Method Enzymol. 210 (1992) 37–55.

[25] R. Jankowiak, Fundamental aspects of fluorescence line-narrow-

ing spectroscopy. In: Shpol�skii Spectroscopy and Other Site-

Selection Methods, Chemical Analysis, first ed. vol. 156, Wiley,

New York, 2000, pp. 235–271.

[26] S.K. Pal, J. Peon, A.H. Zewail, Ultrafast surface hydration

dynamics and expression of protein functionality: a-chymotryp-

sin, Proc. Natl. Acad. Sci. 99 (2002) 15297–15302.

[27] G. Mei, A. Di Venere, A. Finazzi Agro, F. De Matteis, N.

Rosato, Dipolar relaxation times of tryptophan and tyrosine in

glycerol and proteins: a direct evaluation from their fluorescence

decays, J. Fluoresc. 13 (2003) 467–477.