Homoisoleucine: A translationally-active leucine surrogate of expanded hydrophobic surface area

Non-canonical amino acids (ncAAs) provide useful tools for the investigation and control of protein behavior.^[1] Several laboratories have used ncAAs to explore the role of hydrophobic forces in stabilizing proteins, including prion proteins, T4 lysozyme, chloramphenicol acetyltransferase, green fluorescent protein, and coiled-coil and helix-bundle proteins.^[2] Here we examine the consequences of introducing the leucine surrogate (2S,4S)-2-amino-4-methylhexanoic acid (homoisoleucine, Hil, 2)^[2a] into the coiled-coil peptide A1 (Figure 1).^[3] A1 contains six heptad repeats, designated (abcdefg), that mediate self-association of the peptide in aqueous solutions. Previous studies have shown that replacement of leucine by (2S,4R)-trifluoroleucine (Tfl, 3) or hexafluoroleucine (Hfl, 4) at the d positions of the heptad repeats leads to substantial stabilization of the coiled-coil structure of A1,^{[2e, 2f, 2h]} presumably through enhanced hydrophobic interactions between fluorinated peptide strands. Similar results have been obtained in other fluorinated coiled-coil and helix-bundle systems.^[4]

Figure 1.

A1 peptide sequence and helical wheel representation of A1 homodimers. The amino acids that comprise the putative heptad repeats are highlighted in gray, with additional emphasis on the leucine residues at the d positions.

Whitesides and coworkers have pointed out that hydrocarbons and fluorocarbons exhibit equivalent “intrinsic” hydrophobicities when changes in molecular surface area are taken into account.^[5] Marsh and coworkers have argued that the “efficient packing” of the bulkier fluorinated amino acids in helix bundle cores can be more important than fluorination per se.^[6] Because the molecular surface areas of Hil and Tfl are nearly identical (and larger than that of Leu by 14–19 Ų),^[7] we imagined that replacement of Leu by Hil might stabilize coiled-coil peptides such as A1.

We first focused our attention on the translational activity of Hil in bacterial cells. Schultz and coworkers reported incorporation of Hil into proteins via chemical misacylation of suppressor tRNA and in vitro translation,^[2a] but we are unaware of previous studies of incorporation of Hil into cellular proteins. E. coli strain LAM1000 (a previously reported leucine auxotroph) was co-transformed with expression plasmid pA1EL and repressor plasmid pREP4.^[2f] pA1EL codes for both the protein A1 and a constitutively expressed copy of the E. coli leucyl-tRNA synthetase (LeuRS) gene. Protein expression was induced in minimal medium depleted of Leu and supplemented with Hil (see Supporting Information for details). Electrophoretic analysis of whole-cell lysates indicated high-level protein expression in media containing as little as 0.25 mM L-Hil, a concentration comparable to the concentrations of canonical amino acids in minimal media.

ATP-PP_i exchange assays confirmed that Hil is activated by the E. coli LeuRS, albeit at a rate substantially lower than that characteristic of the natural substrate (Table 1). The reduced rate of activation of Hil is consistent with our observation that Hil supports high-level protein expression only when LeuRS is over-expressed in the bacterial host.

Table 1.

Kinetic parameters for activation of Leu and Hil by LeuRS.

Substrate[a]	kcat [s−1]	Km [μM]	kcat/Km [rel]
Leu[b]	15.1 ± 2.2	3.7 ±1.9	1
Hil	0.4 ± 0.1	77 ± 65	1/690

^[a]Leu was used as the L-isomer; Hil as a mixture of the D- and L- isomers. The concentrations of Hil reported here are those of the L-isomer. Kinetic parameters are reported as averages determined from three independent experiments with errors reported as averages of the 95% confidence intervals.

^[b]Parameters determined for activation of Leu are consistent with previous reports; the value of k_cat measured in this work is within the range of reported values, while the value of K_m reported here is lower than literature values by a factor of 2–10.^{[2f, 2h, 8]}

A1 samples containing Leu and Hil were purified from 25-mL cultures in yields of 15.9 ± 2.5 mg L⁻¹ and 10.2 ± 1.1 mg L⁻¹, respectively. Liter-scale expression and purification improved the yields of A1 sequence variants containing Hil two- to threefold. Liquid chromatography/mass spectrometry (LC/MS) indicated replacement of at least 97% of Leu by Hil (see “Determination of Amino Acid Replacement Levels” in the Supporting Information).

Figure 2A shows circular dichroism spectra of 10 mM solutions of the Leu- (Leu-A1) and Hil- (Hil-A1) forms of A1. Strong minima at 208 and 222 nm confirm that both proteins assume α-helical structures; K2D2^[9] analysis indicates helical contents of 60–69%, consistent with the fact that the putative heptad repeats constitute 57% of the peptide sequence. A1 is expected to exist primarily as dimers under the conditions employed here.^[10]

Figure 2.

Circular dichroism spectra of Leu-A1 and Hil-A1. A) Wavelength scans performed at 1°C. B) Ellipticity at 222 nm as a function of temperature. All experiments were performed with 10 mM peptide in PBS, pH 7.4.

Replacement of Leu by Hil increases the denaturation temperature of A1, as expected. Figure 2B shows the molar ellipticities at 222 nm of solutions of Leu-A1 and Hil-A1 as functions of temperature. Fitting the CD data to a model of a two-state transition between folded dimer and unfolded monomer states^[11] yielded melting temperatures of 58.7±0.2°C and 75.8±0.1°C for Leu-A1 and Hil-A1, respectively.

Table 2 compares the extent to which the thermal denaturation temperature of A1 is elevated by replacement of Leu by bulkier hydrocarbon and fluorocarbon surrogates. Replacement of Leu by Hil raises T_m by 17°C, as compared to 10°C for replacement of Leu by (2S,4R)-Tfl and 22°C for replacement by Hfl. Expansion of hydrophobic side chain volume at the d-position of the heptad repeat constitutes an effective strategy for stabilization of coiled-coil peptides, irrespective of the hydrocarbon or fluorocarbon character of the side chain.

Table 2.

Stabilization of A1 by replacement of Leu with non-canonical amino acids.^[a]

Amino acid at d position	Leu	(2S,4R)-Tfl	Hil	Hfl[b]
ΔTm	0	10[2h]	17	22[2f]

^[a]Increase in melting temperature (as compared to Leu-A1) determined from CD spectroscopy of 10 μM solutions of peptide in PBS, pH 7.4.

^[b]74% replacement of Leu.

We do not mean to suggest similar molecular origins for the hydrophobic properties of hydrocarbons and fluorocarbons. Although Hil and Tfl behave similarly with respect to elevation of the melting temperature of A1, other experiments suggest important differences in the behavior of water adjacent to hydrocarbon and fluorocarbon side chains. Recent studies via ultrafast spectroscopy indicate a marked slowing of water motions upon replacement of Leu by Tfl at solvent-exposed sites^[12]. In contrast, replacement of Leu by Hil is accompanied by increased rates of solvent reorganization. Much remains to be done to elucidate the origins of hydrophobic effects in proteins and other molecular systems.

In conclusion, we find that Hil serves as an effective surrogate for Leu with respect to protein translation in bacterial cells, and that replacement of Leu by Hil leads to substantial stabilization of recombinant coiled-coil peptides. The results reported here also highlight the value of amino acid replacement at multiple sites in peptides and proteins; replacement of a single Leu residue by Hil in T4 lysozyme has been reported to cause an increase of just 1.9°C in the melting temperature of the protein.^[2a] In contrast, replacement of six Leu residues in the putative coiled-coil domain of A1 raises T_m by 17°C.

Experimental Section

Detailed experimental protocols can be found in Supporting Information.