- Email: [email protected]
Light-dependent dimerisation in the N-terminal sensory module of cyanobacterial phytochrome 1
FEBS Letters 579 (2005) 3970–3974
FEBS 29732
Light-dependent dimerisation in the N-terminal sensory module of cyanobacterial phytochrome 1 Holger M. Straussa,*, Peter Schmiedera, Jon Hughesb a
Forschungsinstitut fu¨r Molekulare Pharmakologie, Robert-Ro¨ssle Strasse 10, D-13125 Berlin, Germany b Pflanzenphysiologie, JLU Giessen, Senckenbergstrasse 3, D-35390 Giessen, Germany Received 18 March 2005; revised 3 June 2005; accepted 6 June 2005 Available online 27 June 2005 Edited by Richard Cogdell
Abstract Phytochromes, photoreceptors controlling important physiological processes in plants and many prokaryotes, are photochromic biliproteins. The red-absorbing Pr ground state is converted by light into the farred-absorbing Pfr which can be photoconverted back to Pr. In plants at least Pfr is the physiologically active signalling state. Here, we show that the N-terminal photochromic module of Cph1 homodimerises reversibly and independently in Pr and Pfr, Pfr-dimers being significantly more stable. Implications for the mechanism of signal transduction are discussed. 2005 Published by Elsevier B.V. on behalf of the Federation of European Biochemical Societies. Keywords: Phytochrome; Two-component system; Analytical ultracentrifugation; Pfr; Histidine sensor kinase; Signal transduction
1. Introduction Phytochromes are an important class of sensory biliproteins in all plants, most algae, as well as certain fungi and many prokaryotes including most cyanobacteria. Typically phytochromes are photochromic in the 600–750 nm wavelength range: the red-absorbing ground state Pr (kmax ca. 660 ± 10 nm, R) is converted to the farred (kmax ca. 720 ± 15 nm, FR) absorbing Pfr. FR converts Pfr back to Pr. This photochromicity reflects a cis/trans-isomerisation around the double bond between position 15 and 16 of the bilin prosthetic group. Note that due to the spectral overlap of the two forms in the red region, a photoequilibrium mixture will be produced by irradiating with R, which in case of Cph1 contains maximally 0.7/0.3 Pfr/Pr. In plants, phytochromes regulate numerous aspects of development including germination, shade avoidance and flowering [1], while in prokaryotes, phytochromes have been implicated * Corresponding author. Fax: +49 30 94793 169. E-mail address: [email protected] (H.M. Strauss).
Abbreviations: AUC, analytical ultracentrifugation; Cph1D2, N-terminal, photochromic module of cyanobacterial phytochrome 1; FR, farred light (720 nm); HSK, histidine sensor kinase; PCB, Phycocyanobilin; Pfr, farred-absorbing form of phytochrome; Pr, red-absorbing form of phytochrome; R, red light (660 nm); r.m.s.d, root mean squared deviation; SAR, specific absorbance ratio; SEC, size-exclusion chromatography
in the control of photosystem biosynthesis [2], phototaxis [3] and input to the circadian clock [4]. In neither case, despite intensive studies worldwide, is the signalling mechanism understood. Cyanobacterial phytochrome 1 (Cph1) was discovered by sequence comparison in the genome of Synechocystis sp. PCC 6803 and subsequently shown to be a bona fide phytochrome acting as a light-regulated histidine sensor kinase (HSK) involved in a ‘‘two-component’’ phosphotransfer reaction with a response regulator, Rcp1, encoded in the same cistron [5–7]. Many, if not most, HSKÕs show a common architecture with a catalytic module attached to a regulatory module adapted to sense a particular aspect of the extra- or intracellular environment (hormones, ions, amino acids, sugars, osmolarity or, as in the case of Cph1, light). All known HSKÕs are stable dimers bound by strong interactions between double antiparallel a-helices constituting the H subdomain of the catalytic module. Kinase activation leads to mutual transphosphorylation of a conserved histidine residue in the H subdomain, followed by phosphotransfer to an aspartate residue of the cognate response regulator or an intermediate carrier [8]. While sensor modules are appropriately diverse, the kinase modules are remarkably strongly conserved. Hence it is likely that the regulatory mechanism used in all histidine sensor kinases is similar. Given the great importance of this class of sensor amongst pathogens, the regulatory mechanism is of considerable interest. Cph1 thus offers a most useful experimental model, as the regulatory agent is light rather than a chemical ligand. The similarities between Cph1 and other phytochromes make it additionally interesting in this context too. In parallel with attempts to elucidate the 3D structure of Cph1, we have been studying the physicochemistry of the N-terminal sensor module. The truncated translation product Cph1D2 (molecular mass 59.3 kDa, residues 1–514 plus a C-terminal His6-tag) lacking the catalytic domain covalently binds its native chromophore phycocyanobilin (PCB) in an autocatalytic fashion to give a red-/farred photochromic holoprotein, photochemically almost indistinguishable from the full-length wildtype. In early studies we noticed that Cph1D2 mobility in size-exclusion chromatography was Pr/Pfr state dependent, even allowing Pfr to be purified from mixtures of the two forms. In order to establish the physical basis of this effect, we subsequently carried out an intensive quaternary structural investigation using analytical ultracentrifugation. We conclude that hydrophobic interactions lead to both (Pr)2 and (Pfr)2 homodimers, the latter being much more stable. As there is no evidence for (PrPfr) heterodimers, the
0014-5793/$30.00 2005 Published by Elsevier B.V. on behalf of the Federation of European Biochemical Societies. doi:10.1016/j.febslet.2005.06.025
H.M. Strauss et al. / FEBS Letters 579 (2005) 3970–3974
interaction involved in (Pr)2 formation is probably qualitatively different from that in (Pfr)2, perhaps involving completely different residues and/or portions of the surface. The implications of this for models of intramolecular regulation in histidine sensor kinases and in phytochromes are discussed.
3971
monomer. This was expected as deletion of the C-terminal catalytic module removed the H box subdomain primarily responsible for dimerisation of the full-length molecule. Unexpectedly, however, the photoequilibrium Pfr/Pr mixture achieved in R eluted as two peaks, a smaller corresponding
2. Material and methods 2.1. Preparation of chromophore and Cph1D2 Preparation of apo- and holoprotein was essentially performed as described [9]. Briefly, overnight cultures of XL1-blue E. coli (Stratagene) harbouring the p929.5 expression plasmid were grown to early exponential phase (OD600 nm 0.4) at 30 C and cooled on ice. Expression was induced with 20 lM IPTG at 18 C for 18 h, then the cells were washed and concentrated in cold TES (50 mM Tris pH 7.8, 5 mM EDTA, 300 mM NaCl, including 1 mM DTT) before lysis by two passages through a French press at 120 MPa. Before the first passage, phycocyanobilin (PCB) was added to the cell suspension. The crude extract was clarified at 25 000 · g, precipitated with ammonium sulphate, resuspended in 50 mM Tris pH 7.8, 300 mM NaCl and subjected to Ni2+-affinity chromatography. Eluted holo-Cph1D2 was further purified by preparative SEC in a Sephacryl S300 (Pharmacia/ GE) column in TES. Protein preparations used in analytical ultracentrifugation experiments had an SAR (specific absorbance ratio, A656/ A280) of >1.1, implying a holoprotein purity of >80%. All manipulations of holoprotein were carried out in darkness or under green safelight (light-emitting diodes (LEDs), kmax 520 nm). 2.2. Analytical gel filtration and preparation of pure Pfr Affinity-purified Cph1D2 in TES was concentrated by ultrafiltration (Centricon 15 K) to 5 mg/mL. Pr and 0.7 Pfr/0.3 Pr photoequilibrium mixture were produced by saturating irradiation at 730 or 660 nm (using appropriate LEDs) with short pathlengths to avoid spectral gradients. 200 ll samples were loaded onto a Superdex 200 (Pharmacia/ GE) size-exclusion column running at 250 lL/min with TES. The column was calibrated regularly using six marker proteins between 12.4 and 443 kDa (Sigma). 500 lL fractions were collected and UV–Vis spectra measured with a Uvikon 941 spectrophotometer. 2.3. Analytical ultracentrifugation AUC experiments were conducted in a Beckmann Coulter Xl-I analytical ultracentrifuge. Data were collected at appropriate wavelengths ranging from 245 to 330 nm. Three different speeds in the range 8– 20 krpm were used for the determination of equilibrium coefficients. Experiments were performed in 50 mM Tris, pH 7.8, 5 mM EDTA and 0/150/300 mM NaCl, at 10/20/30 C. Solvent densities were calculated using SEDNTERP (www.bbri.org/RASMB/rasmb.html). The partial specific volume (m) of Cph1D2 was calculated from the sequence and the contribution of the (unprotonated) chromophore was accounted for by summing tabulated density increments of the respective chemical groups. An extinction coefficient e280 of 59 mM 1 cm 1 was used to determine molar concentrations and extinction coefficients at all other wavelengths were calculated relative to this value. Attainment of apparent sedimentation and chemical equilibrium was checked with MATCH. Equilibrium data were analyzed using UltraScan6.0 (http://www.ultrascan.uthscsa.edu/), NONLIN (MATCH and NONLIN are available from http://vm.uconn.edu/~wwwbiotc/ uaf.html) and the M\-function implemented in a script for Origin6.0 (K. Schilling, Nanolytics, Dallgow, Germany; www.bbri.org/RASMB/ rasmb.html). A detailed description of the experimental procedures can be found in the supplementary material.
3. Results Size-exclusion chromatography in darkness revealed a different elution behaviour for the two forms of Cph1D2 (Fig. 1A). Pr (derived from saturating FR irradiation) eluted at the position expected for a globular protein of 60 kDa, that is as a
Fig. 1. (A) SEC-profile of Pr and Pr/Pfr-mixtures. (B) spectrum of pure Pfr after SEC overlaid with a Pfr spectrum calculated on the basis of initial rate measurements. (C) equilibrium gradients at 12 krpm of sequentially irradiated Cph1D2 (10 C, 150 mM NaCl). The arrow indicates the position at which the wavelength spectra shown in (D) were taken.
3972
H.M. Strauss et al. / FEBS Letters 579 (2005) 3970–3974
to Pr at 60 kDa and a larger corresponding to 113 ± 10 kDa, a putative dimer. The UV–Vis absorbance spectrum of this peak corresponded exactly to that calculated for pure Pfr from the 0.70/0.30 Pfr/Pr molar ratio at photoequilibrium derived from initial rate measurements according to ButlerÕs equations for photochromic pigments [10,11]. This is the first Pfr absorbance spectrum to have been measured directly. The UV/Vis properties of such preparations kept in total darkness for two weeks at 20 C did not change. Thus Cph1-Pfr is thermodynamically stable, showing no tendency for dark reversion to Pr. In order to understand whether this difference in elution behaviour was due to a change in the hydrodynamic shape of the molecule or rather an increase in the apparent molar mass of the Pfr-fraction, sedimentation equilibrium experiments were performed. Cph1D2 solutions were sequentially irradiated with FR (Pr 1), R (Pfr/Pr) and FR (Pr 2) (Fig. 1C). After acquisition of the respective equilibrium gradients, a wavelength spectrum was recorded (Fig. 1D). This experiment showed that generation of Pfr increases the slope of the equilibrium gradient, which is directly proportional to the average molar mass of the system. Regeneration of Pr by simple irradiation with FR demonstrates the reversibility of this increase. Furthermore, this experiment confirms the thermal stability of Pfr under the conditions used in the AUC and for the time needed to attain sedimentation equilibrium (ca. 30 h for one gradient). A quantitative analysis of the weightaverage molar masses confirms this increase and also shows that no protein is lost due to precipitation during the photocycling (Table 1).
Table 1 Results of an M*-analysis of the experiment shown in Fig. 1C 0 mM NaCl
150 mM NaCl
300 mM NaCl
Pr 1 Mwav (kg/mol) cload (mg/mL)
57.25 0.53
54.86 0.41
61.68 0.68
Pr/Pfr Mwav (kg/mol) cload (mg/mL)
72.33 0.55
69.18 0.46
70.9 0.73
Pr 2 Mwav (kg/mol) cload (mg/mL)
57.52 0.53
59.29 0.42
64.38 0.69
We performed model-dependent fitting of preparations containing Pr or Pfr only (>95% purity), allowing the molar mass parameter to be floated. This analysis showed that both Pr and Pfr of Cph1D2 undergo a reversible monomer–dimer equilibrium, albeit with an association coefficient differing by more than one order of magnitude (Table 2). This difference is large enough to separate the two forms of the protein by non-equilibrium methods such as SEC at appropriate concentrations (vide supra). To understand the nature of the driving forces for Pr and Pfr dimerisation, we systematically varied both the experimental temperature and the ionic strength of the solvent. The Pr and the Pfr interactions increased with increasing temperature and ionic strength, indicating a hydrophobic interaction in both cases (Fig. 2B). Under all conditions tested, Pfr showed a stronger tendency to dimerise. Note that using our experimental protocol (cf. supplementary material), we cannot exclude the contribution of oligomers other than (Pfr)2 to the second dimerisation coefficient. Taken together, these data demonstrate the presence of two different, hydrophobic interfaces which are controlled by the cis-/trans-isomerisation of the bilin prosthetic group.
4. Discussion In this work, we have shown that light regulates the quaternary structure of the N-terminal signalling module of Cph1. Our data demonstrate that both (Pr)2 and (Pfr)2 dimers occur and that the interaction is hydrophobic, but that the affinity is about an order of magnitude stronger in the latter case. Although we cannot exclude their occurrence, we found no evidence for the existence of (PrPfr) heterodimers. Photodynamically, the mole fraction of Cph1-Pfr in red light is calculated to be 0.70, whereas if Pr/Pfr interactions were equally strong these oligomers would be abundant in the dimer fraction. On the contrary, no such heterodimers are detectable, implying that Pr and Pfr present different interaction surfaces (see Fig. 3A). We cannot exclude the possibility that a much weaker interaction allows Pr/Pfr heterodimerisation at the same surface, however. Fo¨rster resonance energy transfer studies with Cph1D2 and the stably dimerised full-length Cph1 are inconclusive in this regard [12]. These findings are interesting in at least three contexts.
Table 2 Predicted values of the molar mass (expected value 59 kg/mol) and association coefficients of pure Pr and Pfr KA monomer–dimer (mM 1)
r.m.s.d. (mOD)
Model
Molar mass (kg/mol)
Pr Monomer–dimer
61.3 56–66.7
19.8 9.54–41.3
4.95 4.86–5.14
Pfr Monomer–dimera
67.2 63.8–70.6
96 62.4–147.5
4.87 4.6–5.12
Pfr Monomer–dimer–tetramerb
56.1 50.1–62.2
334.9 147.9–758.7
4.83 4.56–5.07
95%-confidence intervals are given below the best-fit parameter values, in bold. Experimental conditions: 10 C, 300 mM NaCl. The corresponding fits are depicted in supplementary materials. a A monomer–dimer only model is not sufficient to describe the data as the predicted molar mass deviates significantly from the expected value. This model is thus shown for completeness only. b The formation of oligomers higher than the dimer is likely, however, their nature cannot be established unambiguously from our data. The tetramer model is shown as a plausible example only.
H.M. Strauss et al. / FEBS Letters 579 (2005) 3970–3974
Fig. 2. (A) Sedimentation data at different volume fractions of Pfr (UPfr) and fits to a model containing association coefficients describing the dimerisation of Pr and Pfr as detailed in supplementary material (20 C, 150 mM NaCl). The respective residuals are shown on top. Experimental data at all other conditions were similar to the ones depicted. (B) Variation of the association coefficient of Pr and Pfr with solvent conditions, as determined from experiments similar to those in (A). Error bars represent 95%-confidence intervals.
Firstly, it allows Pfr to be produced free of Pr. This might be valuable in NMR and diffraction-based 3D structural studies of phytochrome function. Secondly, the change in interaction might represent the physical basis for catalytic regulation in HPKÕs and/or phytochromes (see Fig. 3B). In the former case, changing the contact surfaces might reorient and thus activate the C-terminal kinase module, thus triggering autophosphorylation and downstream signalling via the two-component pathway, while in the case of plant phytochromes the change might rather affect intermolecular signalling interactions of the N-terminal region. Weak (PrPfr) interaction might mean that for the dimeric full-length protein, both subunits need to be phototransformed before downstream signalling can take place. This would introduce a combinatorial threshold for signalling activity. Interestingly, a particular mode of phytochrome signalling at very low fluence rates has been proposed to derive from (PrPfr) heterodimers [13,14]. As movement of the sensor module would be constrained by the hinge mechanism, the Pr/Pr interaction might be much weaker than in the case of free Cph1D2. Indeed, it is possible that the hydrophobic interface of Pr is a peculiarity of the recombinant deletion construct: removal of the C-terminus might have left a hydrophobic patch exposed to the solvent, which would be responsible for the weak dimerisation seen in Cph1D2. However, this potential artefact is inconsequent for the signalling hypothesis discussed above. Finally, due to the reversibility of the interaction and the stability of both Pr and Pfr states, photoreversible dimerisation of
3973
Fig. 3. Models for quaternary structure changes. (A) Dimerisation behaviour of Cph1D2 as Pr and Pfr. (Pr)2 and (Pfr)2 homodimers of the sensor module (Cph1D2) are photoreversible and have different stabilities, indicating different contact surfaces in each case. (B) Quaternary structural changes as a basis for regulating the signalling activity of phytochromes and two-component type histidine sensor kinases. The full-length molecule (Cph1) is a stable dimer, the binding site(s) lying in the C-terminal module. The N-terminal sensor module is connected to this via a hinge region to form a Y-shaped structure in the Pr state (as proposed by Jones and Erikson in 1989 [17]). Signalling is regulated by the conformation of the N-termini, the two arms of the molecule coming together in the Pfr state (as proposed by Furuya and Song, 1994 [18]). Our data provides the physicochemical basis for this model.
Cph1D2 is interesting biotechnologically. It might be possible to generate light-controllable systems in the context of proteins requiring dimerisation for function. As holoCphD2 can be generated in vivo [15,16] or by an in vitro autoassembly with apophytochrome and free bilins, recombinant proteins could be engineered in which the Cph1D2 module replaces the normal dimerisation site. Acknowledgements: This work was supported by the Deutsche Forschungsgemeinschaft (SFB498, TP B2 (J.H.) and B6 (P.S.)).
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version at doi:10.1016/j.febslet.2005. 06.025.
References [1] Smith, H. (2000) Phytochromes and light signal perception by plants-an emerging synthesis. Nature 407, 585–591. [2] Giraud, E., Fardoux, J., Fourrier, N., Hannibal, L., Genty, B., Bouyer, P., Dreyfus, B. and Vermeglio, A. (2002) Bacteriophytochrome controls photosystem synthesis in anoxygenic bacteria. Nature 417, 202–205.
3974 [3] Yoshihara, S. and Ikeuchi, M. (2004) Phototactic motility in the unicellular cyanobacterium Synechocystis sp. PCC 6803. Photochem. Photobiol. Sci. 3, 512–518. [4] Schmitz, O., Katayama, M., Williams, S.B., Kondo, T. and Golden, S.S. (2000) CikA, a bacteriophytochrome that resets the cyanobacterial circadian clock. Science 289, 765–768. [5] Kaneko, T., Sato, S., Kotani, H., Tanaka, A., Asamizu, E., Nakamura, Y., Miyajima, N., Hirosawa, M., Sugiura, M., Sasamoto, S., Kimura, T., Hosouchi, T., Matsuno, A., Muraki, A., Nakazaki, N., Naruo, K., Okumura, S., Shimpo, S., Takeuchi, C., Wada, T., Watanabe, A., Yamada, M., Yasuda, M. and Tabata, S. (1996) Sequence analysis of the genome of the unicellular cyanobacterium Synechocystis sp. strain PCC6803. II. Sequence determination of the entire genome and assignment of potential protein-coding regions. DNA Res. 30, 109–136. [6] Hughes, J., Lamparter, T., Mittmann, F., Hartmann, E., Ga¨rtner, W., Wilde, A. and Bo¨rner, T. (1997) A prokaryotic phytochrome. Nature 386, 663. [7] Yeh, K.C., Wu, S.H., Murphy, J.T. and Lagarias, J.C. (1997) A cyanobacterial phytochrome two-component light sensory system. Science 277, 1505–1508. [8] Stock, A., Robinson, V.L. and Goudreau, P.N. (2000) Twocomponent signal transduction. Annu. Rev. Biochem. 69, 183–215. [9] Lamparter, T., Esteban, B. and Hughes, J. (2001) Phytochrome Cph1 from the cyanobacterium Synechocystis PCC6803. Purification, assembly, and quaternary structure. Eur. J. Biochem. 268, 4720–4730. [10] Lamparter, T., Mittmann, F., Ga¨rtner, W., Bo¨rner, T., Hartmann, E. and Hughes, J. (1997) Characterization of recombinant phytochrome from the cyanobacterium Synechocystis. Proc. Natl.
H.M. Strauss et al. / FEBS Letters 579 (2005) 3970–3974
[11]
[12]
[13] [14] [15] [16] [17] [18]
Acad. Sci. USA 94, 11792–11797, Erratum in: Proc. Natl. Acad. Sci. USA (1998), 95, 2714. van Thor, J.J., Borucki, B., Crielaard, W., Otto, H., Lamparter, T., Hughes, J., Hellingwerf, K.J. and Heyn, M.P. (2001) Lightinduced proton release and proton uptake reactions in the cyanobacterial phytochrome Cph1. Biochemistry 40, 11460– 11471. Otto, H., Lamparter, T., Borucki, B., Hughes, J. and Heyn, M.P. (2003) Dimerization and inter-chromophore distance of Cph1 phytochrome from Synechocystis, as monitored by fluorescence homo and hetero energy transfer. Biochemistry 42, 5885–5895. van der Woude, W.J. (1985) A dimeric mechanism for the action of phytochrome: evidence from photothermal interactions in lettuce seed germination. Photochem. Photobiol. 42, 655–661. Brockmann, J., Rieble, S., Kazarinova-Fukshansky, N., Seyfried, M. and Scha¨fer, E. (1987) Phytochrome behaves as a dimer in vivo. Plant Cell Environ. 10, 105–111. Gambetta, G.A. and Lagarias, J.C. (2001) Genetic engineering of phytochrome biosynthesis in bacteria. Proc. Natl. Acad. Sci. USA 98, 10566–10571. Landgraf, F.T., Forreiter, C., Hurtado Pico, A., Lamparter, T. and Hughes, J. (2001) Recombinant holophytochrome in Escherichia coli. FEBS Lett. 508, 459–462. Jones, A.M. and Erickson, H.P. (1989) Domain structure of phytochrome from Avena sativa visualized by electron microscopy. Photochem. Photobiol. 49, 479–483. Furuya, M. and Song, P.S. (1994) Assembly and properties of holophytochrome, in: Photomorphogenesis in plants (Kendrick, R.E. and Kronenberg, G.H.M., Eds.), second ed, pp. 105–140, Kluwer Academic Publishers, Dordrecht, The Netherlands.
FEBS 29732
Light-dependent dimerisation in the N-terminal sensory module of cyanobacterial phytochrome 1 Holger M. Straussa,*, Peter Schmiedera, Jon Hughesb a
Forschungsinstitut fu¨r Molekulare Pharmakologie, Robert-Ro¨ssle Strasse 10, D-13125 Berlin, Germany b Pflanzenphysiologie, JLU Giessen, Senckenbergstrasse 3, D-35390 Giessen, Germany Received 18 March 2005; revised 3 June 2005; accepted 6 June 2005 Available online 27 June 2005 Edited by Richard Cogdell
Abstract Phytochromes, photoreceptors controlling important physiological processes in plants and many prokaryotes, are photochromic biliproteins. The red-absorbing Pr ground state is converted by light into the farred-absorbing Pfr which can be photoconverted back to Pr. In plants at least Pfr is the physiologically active signalling state. Here, we show that the N-terminal photochromic module of Cph1 homodimerises reversibly and independently in Pr and Pfr, Pfr-dimers being significantly more stable. Implications for the mechanism of signal transduction are discussed. 2005 Published by Elsevier B.V. on behalf of the Federation of European Biochemical Societies. Keywords: Phytochrome; Two-component system; Analytical ultracentrifugation; Pfr; Histidine sensor kinase; Signal transduction
1. Introduction Phytochromes are an important class of sensory biliproteins in all plants, most algae, as well as certain fungi and many prokaryotes including most cyanobacteria. Typically phytochromes are photochromic in the 600–750 nm wavelength range: the red-absorbing ground state Pr (kmax ca. 660 ± 10 nm, R) is converted to the farred (kmax ca. 720 ± 15 nm, FR) absorbing Pfr. FR converts Pfr back to Pr. This photochromicity reflects a cis/trans-isomerisation around the double bond between position 15 and 16 of the bilin prosthetic group. Note that due to the spectral overlap of the two forms in the red region, a photoequilibrium mixture will be produced by irradiating with R, which in case of Cph1 contains maximally 0.7/0.3 Pfr/Pr. In plants, phytochromes regulate numerous aspects of development including germination, shade avoidance and flowering [1], while in prokaryotes, phytochromes have been implicated * Corresponding author. Fax: +49 30 94793 169. E-mail address: [email protected] (H.M. Strauss).
Abbreviations: AUC, analytical ultracentrifugation; Cph1D2, N-terminal, photochromic module of cyanobacterial phytochrome 1; FR, farred light (720 nm); HSK, histidine sensor kinase; PCB, Phycocyanobilin; Pfr, farred-absorbing form of phytochrome; Pr, red-absorbing form of phytochrome; R, red light (660 nm); r.m.s.d, root mean squared deviation; SAR, specific absorbance ratio; SEC, size-exclusion chromatography
in the control of photosystem biosynthesis [2], phototaxis [3] and input to the circadian clock [4]. In neither case, despite intensive studies worldwide, is the signalling mechanism understood. Cyanobacterial phytochrome 1 (Cph1) was discovered by sequence comparison in the genome of Synechocystis sp. PCC 6803 and subsequently shown to be a bona fide phytochrome acting as a light-regulated histidine sensor kinase (HSK) involved in a ‘‘two-component’’ phosphotransfer reaction with a response regulator, Rcp1, encoded in the same cistron [5–7]. Many, if not most, HSKÕs show a common architecture with a catalytic module attached to a regulatory module adapted to sense a particular aspect of the extra- or intracellular environment (hormones, ions, amino acids, sugars, osmolarity or, as in the case of Cph1, light). All known HSKÕs are stable dimers bound by strong interactions between double antiparallel a-helices constituting the H subdomain of the catalytic module. Kinase activation leads to mutual transphosphorylation of a conserved histidine residue in the H subdomain, followed by phosphotransfer to an aspartate residue of the cognate response regulator or an intermediate carrier [8]. While sensor modules are appropriately diverse, the kinase modules are remarkably strongly conserved. Hence it is likely that the regulatory mechanism used in all histidine sensor kinases is similar. Given the great importance of this class of sensor amongst pathogens, the regulatory mechanism is of considerable interest. Cph1 thus offers a most useful experimental model, as the regulatory agent is light rather than a chemical ligand. The similarities between Cph1 and other phytochromes make it additionally interesting in this context too. In parallel with attempts to elucidate the 3D structure of Cph1, we have been studying the physicochemistry of the N-terminal sensor module. The truncated translation product Cph1D2 (molecular mass 59.3 kDa, residues 1–514 plus a C-terminal His6-tag) lacking the catalytic domain covalently binds its native chromophore phycocyanobilin (PCB) in an autocatalytic fashion to give a red-/farred photochromic holoprotein, photochemically almost indistinguishable from the full-length wildtype. In early studies we noticed that Cph1D2 mobility in size-exclusion chromatography was Pr/Pfr state dependent, even allowing Pfr to be purified from mixtures of the two forms. In order to establish the physical basis of this effect, we subsequently carried out an intensive quaternary structural investigation using analytical ultracentrifugation. We conclude that hydrophobic interactions lead to both (Pr)2 and (Pfr)2 homodimers, the latter being much more stable. As there is no evidence for (PrPfr) heterodimers, the
0014-5793/$30.00 2005 Published by Elsevier B.V. on behalf of the Federation of European Biochemical Societies. doi:10.1016/j.febslet.2005.06.025
H.M. Strauss et al. / FEBS Letters 579 (2005) 3970–3974
interaction involved in (Pr)2 formation is probably qualitatively different from that in (Pfr)2, perhaps involving completely different residues and/or portions of the surface. The implications of this for models of intramolecular regulation in histidine sensor kinases and in phytochromes are discussed.
3971
monomer. This was expected as deletion of the C-terminal catalytic module removed the H box subdomain primarily responsible for dimerisation of the full-length molecule. Unexpectedly, however, the photoequilibrium Pfr/Pr mixture achieved in R eluted as two peaks, a smaller corresponding
2. Material and methods 2.1. Preparation of chromophore and Cph1D2 Preparation of apo- and holoprotein was essentially performed as described [9]. Briefly, overnight cultures of XL1-blue E. coli (Stratagene) harbouring the p929.5 expression plasmid were grown to early exponential phase (OD600 nm 0.4) at 30 C and cooled on ice. Expression was induced with 20 lM IPTG at 18 C for 18 h, then the cells were washed and concentrated in cold TES (50 mM Tris pH 7.8, 5 mM EDTA, 300 mM NaCl, including 1 mM DTT) before lysis by two passages through a French press at 120 MPa. Before the first passage, phycocyanobilin (PCB) was added to the cell suspension. The crude extract was clarified at 25 000 · g, precipitated with ammonium sulphate, resuspended in 50 mM Tris pH 7.8, 300 mM NaCl and subjected to Ni2+-affinity chromatography. Eluted holo-Cph1D2 was further purified by preparative SEC in a Sephacryl S300 (Pharmacia/ GE) column in TES. Protein preparations used in analytical ultracentrifugation experiments had an SAR (specific absorbance ratio, A656/ A280) of >1.1, implying a holoprotein purity of >80%. All manipulations of holoprotein were carried out in darkness or under green safelight (light-emitting diodes (LEDs), kmax 520 nm). 2.2. Analytical gel filtration and preparation of pure Pfr Affinity-purified Cph1D2 in TES was concentrated by ultrafiltration (Centricon 15 K) to 5 mg/mL. Pr and 0.7 Pfr/0.3 Pr photoequilibrium mixture were produced by saturating irradiation at 730 or 660 nm (using appropriate LEDs) with short pathlengths to avoid spectral gradients. 200 ll samples were loaded onto a Superdex 200 (Pharmacia/ GE) size-exclusion column running at 250 lL/min with TES. The column was calibrated regularly using six marker proteins between 12.4 and 443 kDa (Sigma). 500 lL fractions were collected and UV–Vis spectra measured with a Uvikon 941 spectrophotometer. 2.3. Analytical ultracentrifugation AUC experiments were conducted in a Beckmann Coulter Xl-I analytical ultracentrifuge. Data were collected at appropriate wavelengths ranging from 245 to 330 nm. Three different speeds in the range 8– 20 krpm were used for the determination of equilibrium coefficients. Experiments were performed in 50 mM Tris, pH 7.8, 5 mM EDTA and 0/150/300 mM NaCl, at 10/20/30 C. Solvent densities were calculated using SEDNTERP (www.bbri.org/RASMB/rasmb.html). The partial specific volume (m) of Cph1D2 was calculated from the sequence and the contribution of the (unprotonated) chromophore was accounted for by summing tabulated density increments of the respective chemical groups. An extinction coefficient e280 of 59 mM 1 cm 1 was used to determine molar concentrations and extinction coefficients at all other wavelengths were calculated relative to this value. Attainment of apparent sedimentation and chemical equilibrium was checked with MATCH. Equilibrium data were analyzed using UltraScan6.0 (http://www.ultrascan.uthscsa.edu/), NONLIN (MATCH and NONLIN are available from http://vm.uconn.edu/~wwwbiotc/ uaf.html) and the M\-function implemented in a script for Origin6.0 (K. Schilling, Nanolytics, Dallgow, Germany; www.bbri.org/RASMB/ rasmb.html). A detailed description of the experimental procedures can be found in the supplementary material.
3. Results Size-exclusion chromatography in darkness revealed a different elution behaviour for the two forms of Cph1D2 (Fig. 1A). Pr (derived from saturating FR irradiation) eluted at the position expected for a globular protein of 60 kDa, that is as a
Fig. 1. (A) SEC-profile of Pr and Pr/Pfr-mixtures. (B) spectrum of pure Pfr after SEC overlaid with a Pfr spectrum calculated on the basis of initial rate measurements. (C) equilibrium gradients at 12 krpm of sequentially irradiated Cph1D2 (10 C, 150 mM NaCl). The arrow indicates the position at which the wavelength spectra shown in (D) were taken.
3972
H.M. Strauss et al. / FEBS Letters 579 (2005) 3970–3974
to Pr at 60 kDa and a larger corresponding to 113 ± 10 kDa, a putative dimer. The UV–Vis absorbance spectrum of this peak corresponded exactly to that calculated for pure Pfr from the 0.70/0.30 Pfr/Pr molar ratio at photoequilibrium derived from initial rate measurements according to ButlerÕs equations for photochromic pigments [10,11]. This is the first Pfr absorbance spectrum to have been measured directly. The UV/Vis properties of such preparations kept in total darkness for two weeks at 20 C did not change. Thus Cph1-Pfr is thermodynamically stable, showing no tendency for dark reversion to Pr. In order to understand whether this difference in elution behaviour was due to a change in the hydrodynamic shape of the molecule or rather an increase in the apparent molar mass of the Pfr-fraction, sedimentation equilibrium experiments were performed. Cph1D2 solutions were sequentially irradiated with FR (Pr 1), R (Pfr/Pr) and FR (Pr 2) (Fig. 1C). After acquisition of the respective equilibrium gradients, a wavelength spectrum was recorded (Fig. 1D). This experiment showed that generation of Pfr increases the slope of the equilibrium gradient, which is directly proportional to the average molar mass of the system. Regeneration of Pr by simple irradiation with FR demonstrates the reversibility of this increase. Furthermore, this experiment confirms the thermal stability of Pfr under the conditions used in the AUC and for the time needed to attain sedimentation equilibrium (ca. 30 h for one gradient). A quantitative analysis of the weightaverage molar masses confirms this increase and also shows that no protein is lost due to precipitation during the photocycling (Table 1).
Table 1 Results of an M*-analysis of the experiment shown in Fig. 1C 0 mM NaCl
150 mM NaCl
300 mM NaCl
Pr 1 Mwav (kg/mol) cload (mg/mL)
57.25 0.53
54.86 0.41
61.68 0.68
Pr/Pfr Mwav (kg/mol) cload (mg/mL)
72.33 0.55
69.18 0.46
70.9 0.73
Pr 2 Mwav (kg/mol) cload (mg/mL)
57.52 0.53
59.29 0.42
64.38 0.69
We performed model-dependent fitting of preparations containing Pr or Pfr only (>95% purity), allowing the molar mass parameter to be floated. This analysis showed that both Pr and Pfr of Cph1D2 undergo a reversible monomer–dimer equilibrium, albeit with an association coefficient differing by more than one order of magnitude (Table 2). This difference is large enough to separate the two forms of the protein by non-equilibrium methods such as SEC at appropriate concentrations (vide supra). To understand the nature of the driving forces for Pr and Pfr dimerisation, we systematically varied both the experimental temperature and the ionic strength of the solvent. The Pr and the Pfr interactions increased with increasing temperature and ionic strength, indicating a hydrophobic interaction in both cases (Fig. 2B). Under all conditions tested, Pfr showed a stronger tendency to dimerise. Note that using our experimental protocol (cf. supplementary material), we cannot exclude the contribution of oligomers other than (Pfr)2 to the second dimerisation coefficient. Taken together, these data demonstrate the presence of two different, hydrophobic interfaces which are controlled by the cis-/trans-isomerisation of the bilin prosthetic group.
4. Discussion In this work, we have shown that light regulates the quaternary structure of the N-terminal signalling module of Cph1. Our data demonstrate that both (Pr)2 and (Pfr)2 dimers occur and that the interaction is hydrophobic, but that the affinity is about an order of magnitude stronger in the latter case. Although we cannot exclude their occurrence, we found no evidence for the existence of (PrPfr) heterodimers. Photodynamically, the mole fraction of Cph1-Pfr in red light is calculated to be 0.70, whereas if Pr/Pfr interactions were equally strong these oligomers would be abundant in the dimer fraction. On the contrary, no such heterodimers are detectable, implying that Pr and Pfr present different interaction surfaces (see Fig. 3A). We cannot exclude the possibility that a much weaker interaction allows Pr/Pfr heterodimerisation at the same surface, however. Fo¨rster resonance energy transfer studies with Cph1D2 and the stably dimerised full-length Cph1 are inconclusive in this regard [12]. These findings are interesting in at least three contexts.
Table 2 Predicted values of the molar mass (expected value 59 kg/mol) and association coefficients of pure Pr and Pfr KA monomer–dimer (mM 1)
r.m.s.d. (mOD)
Model
Molar mass (kg/mol)
Pr Monomer–dimer
61.3 56–66.7
19.8 9.54–41.3
4.95 4.86–5.14
Pfr Monomer–dimera
67.2 63.8–70.6
96 62.4–147.5
4.87 4.6–5.12
Pfr Monomer–dimer–tetramerb
56.1 50.1–62.2
334.9 147.9–758.7
4.83 4.56–5.07
95%-confidence intervals are given below the best-fit parameter values, in bold. Experimental conditions: 10 C, 300 mM NaCl. The corresponding fits are depicted in supplementary materials. a A monomer–dimer only model is not sufficient to describe the data as the predicted molar mass deviates significantly from the expected value. This model is thus shown for completeness only. b The formation of oligomers higher than the dimer is likely, however, their nature cannot be established unambiguously from our data. The tetramer model is shown as a plausible example only.
H.M. Strauss et al. / FEBS Letters 579 (2005) 3970–3974
Fig. 2. (A) Sedimentation data at different volume fractions of Pfr (UPfr) and fits to a model containing association coefficients describing the dimerisation of Pr and Pfr as detailed in supplementary material (20 C, 150 mM NaCl). The respective residuals are shown on top. Experimental data at all other conditions were similar to the ones depicted. (B) Variation of the association coefficient of Pr and Pfr with solvent conditions, as determined from experiments similar to those in (A). Error bars represent 95%-confidence intervals.
Firstly, it allows Pfr to be produced free of Pr. This might be valuable in NMR and diffraction-based 3D structural studies of phytochrome function. Secondly, the change in interaction might represent the physical basis for catalytic regulation in HPKÕs and/or phytochromes (see Fig. 3B). In the former case, changing the contact surfaces might reorient and thus activate the C-terminal kinase module, thus triggering autophosphorylation and downstream signalling via the two-component pathway, while in the case of plant phytochromes the change might rather affect intermolecular signalling interactions of the N-terminal region. Weak (PrPfr) interaction might mean that for the dimeric full-length protein, both subunits need to be phototransformed before downstream signalling can take place. This would introduce a combinatorial threshold for signalling activity. Interestingly, a particular mode of phytochrome signalling at very low fluence rates has been proposed to derive from (PrPfr) heterodimers [13,14]. As movement of the sensor module would be constrained by the hinge mechanism, the Pr/Pr interaction might be much weaker than in the case of free Cph1D2. Indeed, it is possible that the hydrophobic interface of Pr is a peculiarity of the recombinant deletion construct: removal of the C-terminus might have left a hydrophobic patch exposed to the solvent, which would be responsible for the weak dimerisation seen in Cph1D2. However, this potential artefact is inconsequent for the signalling hypothesis discussed above. Finally, due to the reversibility of the interaction and the stability of both Pr and Pfr states, photoreversible dimerisation of
3973
Fig. 3. Models for quaternary structure changes. (A) Dimerisation behaviour of Cph1D2 as Pr and Pfr. (Pr)2 and (Pfr)2 homodimers of the sensor module (Cph1D2) are photoreversible and have different stabilities, indicating different contact surfaces in each case. (B) Quaternary structural changes as a basis for regulating the signalling activity of phytochromes and two-component type histidine sensor kinases. The full-length molecule (Cph1) is a stable dimer, the binding site(s) lying in the C-terminal module. The N-terminal sensor module is connected to this via a hinge region to form a Y-shaped structure in the Pr state (as proposed by Jones and Erikson in 1989 [17]). Signalling is regulated by the conformation of the N-termini, the two arms of the molecule coming together in the Pfr state (as proposed by Furuya and Song, 1994 [18]). Our data provides the physicochemical basis for this model.
Cph1D2 is interesting biotechnologically. It might be possible to generate light-controllable systems in the context of proteins requiring dimerisation for function. As holoCphD2 can be generated in vivo [15,16] or by an in vitro autoassembly with apophytochrome and free bilins, recombinant proteins could be engineered in which the Cph1D2 module replaces the normal dimerisation site. Acknowledgements: This work was supported by the Deutsche Forschungsgemeinschaft (SFB498, TP B2 (J.H.) and B6 (P.S.)).
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version at doi:10.1016/j.febslet.2005. 06.025.
References [1] Smith, H. (2000) Phytochromes and light signal perception by plants-an emerging synthesis. Nature 407, 585–591. [2] Giraud, E., Fardoux, J., Fourrier, N., Hannibal, L., Genty, B., Bouyer, P., Dreyfus, B. and Vermeglio, A. (2002) Bacteriophytochrome controls photosystem synthesis in anoxygenic bacteria. Nature 417, 202–205.
3974 [3] Yoshihara, S. and Ikeuchi, M. (2004) Phototactic motility in the unicellular cyanobacterium Synechocystis sp. PCC 6803. Photochem. Photobiol. Sci. 3, 512–518. [4] Schmitz, O., Katayama, M., Williams, S.B., Kondo, T. and Golden, S.S. (2000) CikA, a bacteriophytochrome that resets the cyanobacterial circadian clock. Science 289, 765–768. [5] Kaneko, T., Sato, S., Kotani, H., Tanaka, A., Asamizu, E., Nakamura, Y., Miyajima, N., Hirosawa, M., Sugiura, M., Sasamoto, S., Kimura, T., Hosouchi, T., Matsuno, A., Muraki, A., Nakazaki, N., Naruo, K., Okumura, S., Shimpo, S., Takeuchi, C., Wada, T., Watanabe, A., Yamada, M., Yasuda, M. and Tabata, S. (1996) Sequence analysis of the genome of the unicellular cyanobacterium Synechocystis sp. strain PCC6803. II. Sequence determination of the entire genome and assignment of potential protein-coding regions. DNA Res. 30, 109–136. [6] Hughes, J., Lamparter, T., Mittmann, F., Hartmann, E., Ga¨rtner, W., Wilde, A. and Bo¨rner, T. (1997) A prokaryotic phytochrome. Nature 386, 663. [7] Yeh, K.C., Wu, S.H., Murphy, J.T. and Lagarias, J.C. (1997) A cyanobacterial phytochrome two-component light sensory system. Science 277, 1505–1508. [8] Stock, A., Robinson, V.L. and Goudreau, P.N. (2000) Twocomponent signal transduction. Annu. Rev. Biochem. 69, 183–215. [9] Lamparter, T., Esteban, B. and Hughes, J. (2001) Phytochrome Cph1 from the cyanobacterium Synechocystis PCC6803. Purification, assembly, and quaternary structure. Eur. J. Biochem. 268, 4720–4730. [10] Lamparter, T., Mittmann, F., Ga¨rtner, W., Bo¨rner, T., Hartmann, E. and Hughes, J. (1997) Characterization of recombinant phytochrome from the cyanobacterium Synechocystis. Proc. Natl.
H.M. Strauss et al. / FEBS Letters 579 (2005) 3970–3974
[11]
[12]
[13] [14] [15] [16] [17] [18]
Acad. Sci. USA 94, 11792–11797, Erratum in: Proc. Natl. Acad. Sci. USA (1998), 95, 2714. van Thor, J.J., Borucki, B., Crielaard, W., Otto, H., Lamparter, T., Hughes, J., Hellingwerf, K.J. and Heyn, M.P. (2001) Lightinduced proton release and proton uptake reactions in the cyanobacterial phytochrome Cph1. Biochemistry 40, 11460– 11471. Otto, H., Lamparter, T., Borucki, B., Hughes, J. and Heyn, M.P. (2003) Dimerization and inter-chromophore distance of Cph1 phytochrome from Synechocystis, as monitored by fluorescence homo and hetero energy transfer. Biochemistry 42, 5885–5895. van der Woude, W.J. (1985) A dimeric mechanism for the action of phytochrome: evidence from photothermal interactions in lettuce seed germination. Photochem. Photobiol. 42, 655–661. Brockmann, J., Rieble, S., Kazarinova-Fukshansky, N., Seyfried, M. and Scha¨fer, E. (1987) Phytochrome behaves as a dimer in vivo. Plant Cell Environ. 10, 105–111. Gambetta, G.A. and Lagarias, J.C. (2001) Genetic engineering of phytochrome biosynthesis in bacteria. Proc. Natl. Acad. Sci. USA 98, 10566–10571. Landgraf, F.T., Forreiter, C., Hurtado Pico, A., Lamparter, T. and Hughes, J. (2001) Recombinant holophytochrome in Escherichia coli. FEBS Lett. 508, 459–462. Jones, A.M. and Erickson, H.P. (1989) Domain structure of phytochrome from Avena sativa visualized by electron microscopy. Photochem. Photobiol. 49, 479–483. Furuya, M. and Song, P.S. (1994) Assembly and properties of holophytochrome, in: Photomorphogenesis in plants (Kendrick, R.E. and Kronenberg, G.H.M., Eds.), second ed, pp. 105–140, Kluwer Academic Publishers, Dordrecht, The Netherlands.