Vacuum-ultraviolet circular dichroism of Escherichia coli dihydrofolate reductase: Insight into the contribution of tryptophan residues

Chemical Physics Letters 572 (2013) 111–114

Contents lists available at SciVerse ScienceDirect

Chemical Physics Letters journal homepage: www.elsevier.com/locate/cplett

Vacuum-ultraviolet circular dichroism of Escherichia coli dihydrofolate reductase: Insight into the contribution of tryptophan residues Eiji Ohmae a,⇑, Koichi Matsuo b, Kunihiko Gekko a,1 a b

Department of Mathematical and Life Sciences, Graduate School of Science, Hiroshima University, Higashi-Hiroshima, Hiroshima 739-8526, Japan Hiroshima Synchrotron Radiation Center, Hiroshima University, Higashi-Hiroshima, Hiroshima 739-0046, Japan

a r t i c l e

i n f o

Article history: Received 19 February 2013 In final form 10 April 2013 Available online 18 April 2013

a b s t r a c t To elucidate the contribution of tryptophan side chains to the vacuum-ultraviolet (VUV) circular dichroism (CD) of Escherichia coli dihydrofolate reductase, we measured the VUVCD spectra of eight tryptophan mutants down to 175 nm. The difference spectra between the wild-type and the mutants clearly demonstrated that the contribution of tryptophan side chains extends to the high-energy peptide CD in the VUV region. These results should be useful for a theoretical study on improving protein secondary-structure analysis by VUVCD spectroscopy. Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction Circular dichroism (CD) spectroscopy in the far-UV region has been widely used for secondary-structure analysis of proteins because it is sensitive to local peptide structure and can be applied to any protein [1,2]. However, aromatic side chains make a significant contribution to peptide CD in some proteins, although these contributions have not been explicitly considered in secondarystructure analysis [3]. Sreerama and Woody showed that the difference between experimental and theoretical CD spectra is decreased when incorporating the contribution of aromatic side chains into the theory [4]. However, such theoretical and experimental studies on the contribution of aromatic side chains have not been performed for CD spectra in the vacuum-ultraviolet (VUV) region, although VUVCD spectroscopy using synchrotron radiation can predict secondary structures more accurately than conventional CD spectroscopy and hence is becoming more important in protein structural biology [5–8]. Experimentally, the contribution of an aromatic residue can be inferred by examining the difference in CD spectra between wild-type and mutant proteins in which an aromatic amino acid residue is replaced with a nonaromatic or weaker aromatic one. Among the aromatic amino acids, tryptophan exerts the most significant contribution to peptide CD. We previously found that the far-UV CD spectrum of Escherichia coli dihydrofolate reductase (DHFR), which consists of 159

amino acid residues including 5 tryptophan, 4 tyrosine, and 6 phenylalanine residues, is largely affected by the mutation of tryptophan residues, Trp22, Trp30, Trp47, Trp74, and Trp133 (Figure 1), to leucine, phenylalanine, or valine [9]. In the present Letter, we measured the VUVCD spectra of wild-type and these mutant DHFRs without any ligands down to 175 nm to elucidate how the contribution of tryptophan side chains extends to high-energy transitions of peptide CD.

2. Materials and methods The wild-type and mutant DHFRs were sufficiently purified as described previously [9] to a single band on SDS–PAGE gel (Figure S1). The VUVCD spectra of these proteins were measured from 260 to 175 nm using the VUVCD spectrophotometer constructed at the Hiroshima Synchrotron Radiation Center (HiSOR) and an assembled-type optical cell at 25 °C under high vacuum (10 4 Pa) [10,11]. The light-path length of the cell was adjusted to 10 lm using a Teflon spacer. The solvent used was 10 mM potassium phosphate (pH 7.0) containing 0.1 mM EDTA and 0.1 mM dithiothreitol. The protein concentration was 350 lM. 3. Results and discussion 3.1. VUVCD spectra of DHFRs

Abbreviations: CD, circular dichroism; DHFR, dihydrofolate reductase; VUV, vacuum ultraviolet. ⇑ Corresponding author. Fax: +81 82 424 7389. E-mail address: [email protected] (E. Ohmae). 1 Present address: Institute for Sustainable Sciences and Development, Hiroshima University, Higashi-Hiroshima, Hiroshima 739-8526, Japan. 0009-2614/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cplett.2013.04.019

Figure 2 shows the VUVCD spectra of the wild-type and eight mutant DHFRs at pH 7.0 and 25 °C. The spectra of all DHFRs from 260 to 190 nm were consistent with our previous results measured with a conventional CD spectrophotometer at 15 °C [9]. Interestingly, a substitution of only a single tryptophan residue markedly

112

E. Ohmae et al. / Chemical Physics Letters 572 (2013) 111–114

15 15

10 / deg cm dmol

-1

Trp74

2

10

10 -1 2 -3

-5

5

-10 180 200 220 240 260

Wavelength / nm

0

[θ ]

Trp30

0

[θ ]

10 / deg cm dmol

Trp133

5

-3

Trp47

-5

Trp22 Figure 1. Crystal structure of wild-type DHFR (PDB code 1rx2). Folate and NADPH have been removed from the figure because apo-form DHFR was used in the VUVCD measurements. Five tryptophan side chains are indicated by the red stick model. The figure was created in PyMol [http://www.pymol.org/].

affected the VUVCD spectrum of DHFR. Since all mutants retained substantial enzyme activity (>50% of the wild-type activity except for W22F, which had 13%) [9], significant unfolding of the secondary structures is unlikely, although some structural perturbation may have occurred in W30L and W133V mutants, as suggested by their relatively low stability for urea unfolding [9]. The large difference in the spectra below 200 nm clearly indicates that the tryptophan side chains contributed to the chiroptical property of DHFR down to the VUV region. Before considering the difference spectra between wild-type and mutants, we compared the wild type spectrum with the calculated one based on the crystal structure of apo DHFR (PDB code, 5dfr) using the component spectra of the a-helix (21.4%), b-strand (30.8%), turn (6.9%), and unordered structure (40.9%), which were extracted by deconvolution analysis from the VUVCD spectra of 31 reference proteins down to 160 nm [12,13]. In this calculation, undetectable regions in 5dfr were assigned to unordered structures referring to the crystal structure of the ternary complex of DHFR (Figure 1). As shown in the inset of Figure 2, the experimental and calculated spectra present different features: the calculated spectrum generates a negative peak around 208 nm and no shoulder around 185 nm. Since the negative peak around 208 nm corresponds to the p–p⁄ transition of the backbone peptides along the helix axis [4,14,15], the CD of a-helices in DHFR may be susceptible to their environment. The shoulder around 185 nm does not exist in any component spectra of above four secondary structures and cannot be explicitly explained, although such a shoulder was also found for cytochrome c, ribonuclease A, and thioredoxin [12,13]. A possible explanation for this shoulder is that the positive peak around 195 nm is split into two by the contribution of the aromatic side chains because the positive peak around 192 nm involves two components polarized perpendicular to the helix axis [4]. Among the p–p⁄ transitions of aromatic side chains, L bands are observed in near-UV region except for La bands of phenylalanine and tyrosine at about 210 and 230 nm, respectively, whereas B bands are observed at 180–200 nm except for a Bb band of tryptophan around 230 nm [4]. Actually, N-acetyl-L-tyrosine amide and N-acetyl-L-tryptophan amide have large negative CD peaks of less than 104 deg cm2 -

-10 180

200 220 240 Wavelength / nm

260

Figure 2. VUVCD spectra of the wild-type and tryptophan-mutant DHFRs at pH 7.0 and 25 °C. The solvent used was 10 mM potassium phosphate containing 0.1 mM EDTA and 0.1 mM dithiothreitol. The protein concentration was 350 lM. Wild-type (—), W22F ( ), W22L ( ), W30L ( ), W47L ( ), W74F ( ), W74L ( ), W133F ( ), and W133V ( ). (Inset) Experimental (—) and calculated ( ) VUVCD spectra of wild-type DHFR.

dmol 1 at 185 and 195 nm, respectively, with a similar negative peak being expected for N-acetyl-L-phenylalanine amide at 180– 190 nm [1]. Therefore, the p–p⁄ transitions (B bands) of aromatic side chains could couple with the p–p⁄ transition of backbone peptides of helices to split the CD peak around 192 nm although the relative conformations between backbone peptides and aromatic side chains in the protein are different from those in the model compounds [16,17]. Such coupling effects of p–p⁄ transitions between backbone peptides and aromatic side chains might be significant for DHFR as judged from a large difference in VUVCD spectra below 200 nm (Figure 2). However, the theoretical CD spectra calculated by Sreerama and Woody for eight proteins including cytochrome c did not show such a shoulder, regardless of excluding or including the contribution of aromatic side chains in the calculation [4]. On the other hand, theoretical calculations by Bulheller et al. revealed that the charge transfer between neighboring main chain peptides significantly contributes to CD spectra in the VUV region [18]. Therefore, it is unlikely that this shoulder originates from only the direct contribution of the aromatic side chains, although these aromatic contributions would be fairly large for DHFR since this shoulder was affected by the mutation of tryptophan residues.

3.2. VUVCD difference spectra between wild-type and mutant DHFRs Figure 3 shows the difference in VUVCD spectra (wild-type VUVCD – mutant VUVCD) for the eight tryptophan mutants. In wild-type DHFR, two tryptophan side chains, W74 and W47, are in close proximity and give rise to strong coupling. Such exciton coupling is absent in the W74L and W47L mutants, so the difference spectra for W74L and W47L are close to each other and distinctly show a couplet with negative and positive peaks centered around 220 and 230 nm, respectively. This finding is consistent

E. Ohmae et al. / Chemical Physics Letters 572 (2013) 111–114

5

-3

2

10 / deg cm dmol

-1

10

Δ[θ ]

0

-5 180

200

220

240

260

Wavelength / nm Figure 3. Difference VUVCD spectra between wild-type and tryptophan-mutant DHFRs at pH 7.0 and 25 °C. W22F ( ), W22L ( ), W30L ( ), W47L ( ), W74F ( ), W74L ( ), W133F ( ), and W133V ( ).

-1 -3

2

10 / deg cm dmol

Δ[θ ]

trum would give a negative ellipticity around 200 nm and positive ellipticity below 190 nm where the phenylalanine spectrum becomes more negative than the tryptophan spectrum (unpublished data). To confirm the contributions of the phenylalanine side chains to the VUVCD spectrum of DHFR, we calculated the difference spectra (W22F–W22L and W74F–W74L) between phenylalanine and leucine mutants at sites 22 and 74 (Figure 4). The obtained spectra showed two positive peaks at 220–230 and 197 nm, and were similar to the spectrum of N-acetyl-phenylalanine amide, although peak wavelengths were slightly different. Thus, it is obvious that the introduced phenylalanine side chains contribute to the VUVCD spectrum of DHFR, although the detailed discussion needs the VUVCD measurements of the phenylalanine-mutant DHFRs now in progress. 4. Conclusion The present Letter clearly demonstrates that the contribution of tryptophan side chains extends to the VUV region down to at least 175 nm. The positive and negative peaks in difference VUVCD spectra below 200 nm are indicative of the complicated contribution of aromatic residues to the higher energy transitions (n–p⁄ and p–p⁄) of the peptide group. The theoretical difference spectra for some DHFR mutants (e.g., W74L and W47L) are in good agreement with the experimental ones in the far-UV region above 200 nm [4]. Such a comparative study between theoretical and experimental spectra would be powerful for elucidating the aromatic contribution to peptide CD. This in turn would lead to improvements in protein secondary-structure analysis by VUVCD spectroscopy because the CD spectra in the VUV region provide notably more information than in the far UV region.

4 3 2 1 0 -1 -2 -3

113

180

200

220 240 Wavelength / nm

260

Figure 4. Difference VUVCD spectra between phenylalanine- and leucine-mutant DHFRs at sites 22 and 74. Black and red lines indicate the spectra for W22F–W22L and W74F–W74L, respectively.

with the theoretical prediction [9,19,20] and seems applicable to W74F. It is noteworthy that the difference VUVCD spectra are distinguishable between phenylalanine (W22F, W74F, and W133F) and leucine (W22L, W47L, and W74L) mutants. The three leucine mutants, W22L, W47L, and W74L, showed positive ellipticity below 205 nm with two peaks around 180 and 193 nm. Since leucine and N-acetyl-tryptophan amide show positive and negative ellipticity, respectively around this wavelength region [1,21], subtracting the leucine spectrum from the N-acetyl-tryptophan amide spectrum should give negative ellipticity. Hence the difference VUVCD spectra for these three mutants should also be negative. Therefore, the observed positive ellipticity suggests that the side chain residues induce perturbation of the high energy transition of the peptide group in the vicinity of the mutation sites or via modification of the tertiary structure. Different from these three mutants, large positive peaks around 200 nm for W30L and W133V mutants might be partly due to unknown structural changes in these mutants. On the other hand, three phenylalanine mutants, W22F, W74F, and W133F, had a positive peak around 183 nm and a negative peak at 200 nm, with the latter peak being negligible for W22F. These characteristic features might be dominantly attributed to the contribution of the introduced aromatic phenylalanine side chain, which is different from the leucine side chain, because subtracting the phenylalanine spectrum from the tryptophan spec-

Acknowledgements We thank Professors Hirofumi Namatame and Masaki Taniguchi for the use of the VUVCD spectrophotometer at the Hiroshima Synchrotron Radiation Center. This work was financially supported by Grant-in-Aids for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan (No. 24570186, 22870021, and 20550153 to E.O., K.M., and K.G., respectively). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cplett.2013. 04.019. References [1] A.J. Adler, N.J. Greenfield, G.D. Fasman, Methods Enzymol. 27 (1973) 675. [2] G.D. Fasman (Ed.), Circular Dichroism and the Conformational Analysis of Biomolecules, Plenum Press, New York, 1996. [3] N. Berova, K. Nakanishi, R.W. Woody, Circular Dichroism: Principles and Applications, second ed., Wiley-VCH, New York, 2000. [4] N. Sreerama, R.W. Woody, Methods Enzymol. 383 (2004) 318. [5] B.A. Wallace, R.W. Janes, Modern Techniques for Circular Dichroism and Synchrotron Radiation Circular Dichroism Spectroscopy, IOS, Amsterdam, 2009. [6] K. Matsuo, Y. Sakurada, R. Yonehara, M. Kataoka, K. Gekko, Biophys. J. 92 (2007) 4088. [7] K. Matsuo, H. Watanabe, K. Gekko, Proteins 73 (2008) 104. [8] K. Matsuo, H. Namatame, M. Taniguchi, K. Gekko, Biochemistry 48 (2009) 9103. [9] E. Ohmae, Y. Sasaki, K. Gekko, J. Biochem. 130 (2001) 439. [10] N. Ojima et al., Chem. Lett. (2001) 522. [11] K. Matsuo, K. Sakai, Y. Matsushima, T. Fukuyama, K. Gekko, Anal. Sci. 19 (2003) 129. [12] K. Matsuo, R. Yonehara, K. Gekko, J. Biochem. 135 (2004) 405. [13] K. Matsuo, R. Yonehara, K. Gekko, J. Biochem. 138 (2005) 79.

114 [14] [15] [16] [17] [18]

E. Ohmae et al. / Chemical Physics Letters 572 (2013) 111–114 G. Holzwarth, P. Doty, J. Am. Chem. Soc. 87 (1965) 218. Y.-H. Chen, J.T. Yang, K.H. Chau, Biochemistry 13 (1974) 3350. R.W. Woody, Biopolymers 17 (1978) 1451. R.W. Woody, Eur. Biophys. J. 23 (1994) 253. B.M. Bulheller, A.J. Miles, B.A. Wallace, J.D. Hirst, J. Phys. Chem. 112 (2008) 1866.

[19] K. Kuwajima, E.P. Garvey, B.E. Finn, C.R. Mathews, S. Sugai, Biochemistry 30 (1991) 7693. [20] I.B. Grishina, R.W. Woody, Faraday Discuss. 99 (1994) 245. [21] K. Matsuo, Y. Matsushima, T. Fukuyama, S. Senba, K. Gekko, Chem. Lett. (2002) 826.