Skip to main content
NCBI home page
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 2003 Mar 7;100(6):3040–3042. doi: 10.1073/pnas.0030051100

Chiral tubule self-assembly from an achiral diynoic lipid

Serhii Pakhomov 1, Robert P Hammer 1, Bijaya K Mishra 1, Britt N Thomas 1,*
PMCID: PMC152241  PMID: 12626752

Abstract

Tubules possessing μm-scale chiral substructure self-assemble from an achiral isomer of the tubule-forming diynoic phosphatidylcholine, 1,2-bis(10,12-tricosadiynoyl)sn-glycero-3-phosphocholine [DC(8,9)PC], showing that molecular chirality is not essential for tubule formation. CD spectroscopy shows that these structures' helical sense of handedness instead originates in a spontaneous cooperative chiral symmetry-breaking process. We conclude that the chiral symmetry-breaking must originate in the unusual feature common to the chiral and achiral tubule-forming molecules, the diynes centered in their hydrocarbon tails.

The chiral diynoic phosphatidylcholine 1,2-bis(10,12-tricosadiynoyl)sn-glycero-3-phosphocholine [DC(8,9)PC] (compound 1 of Fig. 1) self-assembles in ethanolic solutions to form microscopic (≈0.5 μm × ≈30 μm) hollow cylinders possessing an exterior helical trace similar to that found on a paper drinking straw. This trace is a remnant of the helical winding of a uniform-width phospholipid ribbon that forms tubules, and its helical sense of handedness is determined by the molecule's chirality: R-DC(8,9)PC and S-DC(8,9)PC form tubules possessing right- and left-handed exterior helical traces, respectively.

Figure 1.

Figure 1

Open in a new tab

(Upper) R-DC(8,9)PC. (Lower) β-TFL.

The striking molecular chirality/helical handedness correspondence has led to the idea that tubule formation is driven by molecular chirality. In the so-called “chiral packing” class of tubule formation and structure theories, the tubule-forming molecule's chiral shape causes the directors of neighboring molecules in the close-packed phospholipid bilayer membrane to be offset by a small angle. The cumulative effect of this director tilt is a helical twist along the membrane that results in the winding of the membrane to form closed cylinders. However, it has been recently shown that the Lβ′-phase helical ribbons that grow from enantiopure Lα-phase spherical vesicles are a nearly racemic mix of left- and right-handed helices (1). Minutes after the sphere-to-tubule transition is complete, lipid from the still-cooling DC(8,9)PC-saturated solution completely ensheaths the vesicle-derived helices through the coaxial helical growth of a second, outer cylinder. Paradoxically, whereas the core handedness ratio is nearly racemic, the outer cylinders have a uniform helical sense of handedness that corresponds to DC(8,9)PC chirality. Similar results have been obtained with two enantiomerically pure DC(8,9)PC analogs in which the phosphoryl oxygen linking the phosphatidylcholine headgroup to the chiral glycerol backbone has been removed or replaced by a methylene (-CH2-) group (2, 3). That three molecules possessing different chiral centers each produce nearly racemic helix handedness ratios suggests that the helices are not macroscopic expressions of molecular chirality. Instead, this apparently random membrane chiralization suggests a chiral symmetry-breaking mechanism that is a consequence of the Lα-to-Lβ′ phase transition from which the helices form.

Additional support for chiral symmetry-breaking models comes from studies of tubule formation in DC(8,9)PC racemates. Chiral packing theories predict that reduction of the chiral packing order, e.g., by making making tubules from nonenantiopure DC(8,9)PC preparations, will increase tubule diameter. Further, tubule diameter is predicted to diverge to infinity in the zero net chirality racemate, i.e., only flat sheets are predicted to form. However, tubules of unchanged diameter and both helical senses of handedness were found to form in the DC(8,9)PC racemate (4). These unexpected results can be explained by positing a spontaneous resolution of the zero net chirality mixture into microdomains of opposing chirality wherein chiral packing could occur. The linear dependence of CD spectroscopy peak heights on DC(8,9)PC enantiomeric excess (5) has been interpreted as evidence of nearly complete enantiomeric separation during tubule formation (5, 6). This R-, S-enantiomeric microphase resolution is tantamount to prohibiting the attainment of zero chirality over the lengthscales of tubule nucleation at the Lα (spherical vesicle)-to-Lβ′ (tubule) phase transition temperature.

An alternative interpretation for the persistence of tubule formation in zero net chirality DC(8,9)PC preparations are chiral symmetry-breaking models that do not rely on molecular chirality (6, 7). In these models, attainment of zero chirality over nucleation length scales is not prohibited, and the molecule's chirality is unimportant. However, these models predict equal numbers of left- and right-handed helices, at variance with the well-known DC(8,9)PC chirality/helix handedness correspondence. A qualitative model has been suggested by Spector and coworkers (6) in which DC(8,9)PC helices form through symmetry-breaking that is traceable to the DC(8,9)PC tail diynes, but in which the DC(8,9)PC's chiral headgroup powerfully biases the outcome of the symmetry-breaking process.

To probe the role of DC(8,9)PC chirality in tubule formation, the achiral DC(8,9)PC isomer R-DC(8,9)PC (compound 2 of Fig. 1) was synthesized through an unambiguously achiral pathway, the details of which will be described elsewhere. This symmetric β-lecithin, which we denote as β-tubule-forming lecithin (β-TFL), may be regarded as the result of exchanging the DC(8,9)PC phosphatidylcholine headgroup and the adjacent ester-linked hydrocarbon tail. Because β-TFL lacks a chiral center, the possibility of enantiomeric microphase separation creating chiral microdomains, irrelevant in chiral symmetry-breaking models, is eliminated.

β-TFL preparations exhibit the same phase transition behavior as the corresponding DC(8,9)PC solutions, namely, Lβ′-phase helical ribbons form from Lα-phase spherical vesicles upon cooling. But for larger tubule diameters, different solubilities, etc., the achiral β-TFL is found to produce equal numbers of right- and left-handed tubules under the same conditions as the chiral R-DC(8,9)PC (Fig. 2), and with comparable yields. We therefore conclude that molecular chirality is not essential for tubule formation and that a chiral symmetry-breaking process must drive tubule formation.

Figure 2.

Figure 2

Open in a new tab

Scanning electron microscopy images of partially unwound R-DC(8,9)PC and β-TFL tubules, somewhat unusual for both molecules, that reveal the tubules' helical substructures. (Left) Partially unwound R-DC(8,9)PC tubules always have the right-handed helical sense of the R-DC(8,9)PC tubule exterior cylinder. (Right) A partially unwound left-handed β-TFL tubule; no handedness preference is shown by β-TFL. (Magnification: ×10,000.)

CD spectroscopy has been used to probe chirality in evolving tubules (5, 9). Dissolved enantiopure DC(8,9)PC and its spherical vesicles generate comparatively weak CD signals, but as tubules form in enantiopure R-DC(8,9)PC and S-DC(8,9)PC preparations, strong CD signals of opposite sign and equal magnitude appear, with extrema at ≈200 nm. Because phosphatidylcholine and diynes absorb in the vicinity of the tubule CD signal maximum, the strong tubule CD signal was attributed to the placement of these dichroic chromophores into the highly chiral tubule environment. The absence of appreciable CD signal at wavelengths corresponding to the tubule's 500-nm helical pitch eliminated the tubule's gross helical structure as a possible origin of CD signal through the selective reflection of one sense of circularly polarized light, as has been observed in cholesteric liquid crystal systems (8).

CD spectroscopy measurements over 200 nm ≤ λ ≤ 400 nm performed on tubules made of: (i) enantiopure R-DC(8,9)PC; (ii) pure β-TFL; and (iii) β-TFL “doped” to a 6.1% concentration of R-DC(8,9)PC are shown in Fig. 3. All three preparations' spherical vesicles CD signals were small, essentially featureless at the scale drawn, and are omitted for clarity.

Figure 3.

Figure 3

Open in a new tab

CD spectroscopy of tubules made from enantiopure R-DC(8,9)PC (the solid trace), β-TFL (the short-dashed trace), and β-TFL containing 6.1% R-DC(8,9)PC (the long-dashed trace). For clarity, the weak CD signals from spherical vesicle phases are not shown. The spectra were acquired at 20°C.

The flat β-TFL tubule CD spectrum (the dashed line of Fig. 3), in marked contrast to the strong CD signal emanating from R-DC(8,9)PC tubules, can be explained in two ways: either no CD signal is generated by β-TFL tubules at all, e.g., the dichroic chromophores are not in chiral environments, or alternatively, opposing CD signals generated by the left- and right-handed β-TFL tubules cancel to zero.

To differentiate between these alternatives, a heated achiral β-TFL solution was doped with a small amount of R-DC(8,9)PC and cooled through the spherical vesicle phase to the tubule phase. In the framework of the tilt/chiral symmetry-breaking theory, the equivalence of tilt directions that lead to left- and right-handed β-TFL tubules may be lifted if a chiral dopant is present, which should in turn alter the ratio of left- and right-handed tubules. This effect is readily seen in the β-TFL/R-DC(8,9)PC tubule CD spectrum (the long-dashed line of Fig. 3).

The similarities of β-TFL and R-DC(8,9)PC and the structures they form suggests comparisons of the spectra major features should be straightforward. We note that while the pure β-TFL and doped β-TFL tubule CD spectra are indeed very similar, they are not simply multiples of each other: the maxima occur at somewhat different wavelengths, and the R-DC(8,9)PC spectra trace traverses the zero-ellipticity axis whereas the β-TFL trace does not. With these differences in mind, we nevertheless observe that the doped β-TFL tubules generate a CD signal whose extrema has the same sign as tubules made from the dopant, occurs at nearly the same wavelength, and has an intensity ≈1/3 that of the enantiopure R-DC(8,9)PC tubules, roughly 5-fold greater than one might expect from the 6.1% R-DC(8,9)PC concentration. This intensity is all the more impressive given that R-DC(8,9)PC tubule preparations are decidedly more turbid than the pure and doped β-TFL preparations. Because turbidity is caused by tubule light scattering, and all three preparations' tubules are essentially identical, it is apparent there are significantly fewer doped β-TFL tubules than in the enantiopure R-DC(8,9)PC specimen. This behavior is consistent with the generally greater solubilities that β-lecithins have over their α-lecithin counterparts. Because the strong CD signal emanates from tubules and not the lipid per se, the disproportionately strong CD signal originating from the chirally doped β-TFL tubules is therefore under-reported and would be even closer to that of enantiopure R-DC(8,9)PC if the CD signals were normalized to the number of tubules present, rather than the specimen's total lipid content.

The disproportionately large β-TFL/R-DC(8,9)PC tubule CD signal indicates a cooperative “sergeants-and-soldiers” chiralization, similar to that observed in the achiral polyisocyanate polymer systems studied by Green and coworkers (10–14). The achiral polyisocyanates undergo a spontaneous symmetry-breaking to a chiral state, forming helical rodlike structures. Changing a few percent of the polymer's side groups to be chiral sergeants causes all of the achiral side group soldiers to twist with a single handedness. Of particular relevance to tubule-forming systems is the recent observation of sergeants-and-soldiers chiralization of achiral bent-core (“banana” or “bow”-shaped) smectic liquid crystals that spontaneously form chiral layer structures as a result of polar molecular orientational ordering about the layer normal and molecular tilt from the membrane normal (15). The handedness of each layer is determined by the tilt direction selected by the layer, and chiral dopant concentrations as low as 1% strongly bias the relative stability of one chiral state over the other. Spontaneous chiralization of tilted bilayers of achiral molecules has been discussed by Seifert et al. (7).

The strong CD signal emanating from β-TFL tubules at similarly low chiral dopant concentrations and the elimination of molecular chirality as a requirement for tubule formation lead us to conclude that tubule membrane chirality must have its origins in such an orientation/tilt symmetry-breaking mechanism. Because tubule formation is traceable to the presence of the diynes alone, we speculate that the hydrocarbon tail “kinks” resulting from the linear four-carbon diyne enable an orientational ordering of the tubule-forming molecules, just as the the achiral bow-shaped liquid crystal molecules' bent cores do. Although our achiral tubule-forming system eliminates molecular chirality as a requirement for tubule formation, our results do not exclude the possibility that enantiomeric microphase resolution occurs in chiral R-DC(8,9)PC/S-DC(8,9)PC mixtures as tubules form.

At the present time there is no satisfactory theoretical model of tubule structure or formation. The experimental results presented here show that a satisfactory theory must trace the genesis of tubule membrane chirality to the hydrocarbon tail diynes present in the chiral DC(8,9)PC and its achiral isomer β-TFL. Tubules' sergeants-and-soldiers chiralization, similar to that of polyisocyanates and the achiral bow-shaped smectic liquid crystals, suggests that the Ising-based collective chiral symmetry-breaking models developed for the polymer system (16) may be adaptable to tubule chiralization. Because tubule cores and exteriors are never seen to change their sense of handedness along their lengths, the number of handedness changes f per tubule is f = N exp(−Ek/kT) ≪ 1, where Ek is the kink energy. The nearly racemic DC(8,9)PC core handedness ratio corresponds to an Ising regime where EInline graphic, the energy difference between tilt directions selected by a chiral DC(8,9)PC molecule in a tubule core nuclei, is small. However, a comprehensive theoretical description of tubule structure must abandon the general assumption that the helically wound core and exterior ribbons are equilibrium structures (17–23). Tubule cores grow rapidly (≈1 μm/s axial growth) from supercooled spherical Lα-phase vesicles, whereas tubule exteriors' order-of-magnitude slower growth is fed by the cooling DC(8,9)PC-saturated solution (1). Further, the DC(8,9)PC sphere↔tubule phase transition is characterized by a large thermal hysteresis, with slowly cooled Lα-phase spherical vesicles routinely existing ≈3.5°C below the temperature at which slowly heated Lβ′-phase tubules melt. The tubules exterior cylinder is clearly a growth form about a kinetically determined core. It is likely that the “mixed” DC(8,9)PC tubule core handedness is a consequence of tilt fluctuations present in the rapidly evolving sphere-to-tubule transition that permit nucleation and growth of metastable left-hand ribbons; the order-of-magnitude slower tubule exterior growth does not allow such metatstable nuclei to persist, if they form at all.

Acknowledgments

We thank Noel Clark of the University of Colorado Physics Department for helpful discussions. B.N.T. was supported by National Science Foundation CAREER Grant CHE-9734266.

Abbreviations

DC(8,9)PC

1,2-bis(10,12-tricosadiynoyl)sn-glycero-3-phosphocholine

β-TFL

β-tubule-forming lecithin

References

  • 1.Thomas B N, Lindemann C M, Clark N A. Phys Rev E Stat Phys Plasmas Fluids Relat Interdiscip Top. 1999;59:3040–3047. [Google Scholar]
  • 2.Thomas B N, Corcoran R C, Cotant C L, Lindemann C M, Kirsch J E, Persichini P J. J Am Chem Soc. 1998;120:12178–12186. doi: 10.1021/ja012137a. [DOI] [PubMed] [Google Scholar]
  • 3.Thomas B N, Corcoran R C, Cotant C L, Lindemann C M, Kirsch J E, Persichini P J. J Am Chem Soc. 2002;124:1227–1233. doi: 10.1021/ja012137a. [DOI] [PubMed] [Google Scholar]
  • 4.Singh A, Burke T G, Calvert J M, Georger J H, Herendeen B, Price R R, Schoen P E, Yager P. Chem Phys Lipids. 1988;47:135–147. [Google Scholar]
  • 5.Spector M S, Selinger J V, Singh A, Rodriguez J M, Price R, Schnur J M. Langmuir. 1998;14:3493–3500. [Google Scholar]
  • 6.Selinger J V, Spector M S, Schnur J M. J Phys Chem B. 2001;105:157–7169. [Google Scholar]
  • 7.Seifert U, Shillcock J, Nelson P. Phys Rev Lett. 1996;77:5237–5240. doi: 10.1103/PhysRevLett.77.5237. [DOI] [PubMed] [Google Scholar]
  • 8.Watanabe J, Goto M, Nagase T. Macromolecules. 1987;20:298–304. [Google Scholar]
  • 9.Schnur J M, Ratna B R, Selinger J V, Singh A, Jyothi G, Easwaran K R K. Science. 1994;264:945–947. doi: 10.1126/science.264.5161.945. [DOI] [PubMed] [Google Scholar]
  • 10.Green M M, Peterson N C, Sato T, Teramoto A, Cook R, Lifson S. Science. 1995;268:1860–1866. doi: 10.1126/science.268.5219.1860. [DOI] [PubMed] [Google Scholar]
  • 11.Lifson S, Andreola C, Peterson N C, Green M M. J Am Chem Soc. 1989;111:8850–8858. [Google Scholar]
  • 12.Green M M, Sato T, Teramoto A, Lifson S. Macromol Symp. 1996;101:363–370. [Google Scholar]
  • 13.Khatri C A, Green M M, Morawetz H. Polym Prepr Am Chem Soc Div Polym Chem. 1994;35:807–808. [Google Scholar]
  • 14.Green M M, Khatri C, Peterson N C. J Am Chem Soc. 1993;115:4941–4942. [Google Scholar]
  • 15.Link D R, Natale G, Shao R, Maclennan J E, Clark N A, Körblova E, Walba D M. Science. 1997;278:1924–1927. doi: 10.1126/science.278.5345.1924. [DOI] [PubMed] [Google Scholar]
  • 16.Selinger J V, Selinger R L B. Phys Rev Lett. 1996;76:58–61. doi: 10.1103/PhysRevLett.76.58. [DOI] [PubMed] [Google Scholar]
  • 17.de Gennes P G. C R Acad Sci Ser II. 1987;304:259–263. [Google Scholar]
  • 18.Helfrich W, Prost J. Phys Rev A. 1988;38:3065–3068. doi: 10.1103/physreva.38.3065. [DOI] [PubMed] [Google Scholar]
  • 19.Helfrich W. J Chem Phys. 1986;85:1085–1087. [Google Scholar]
  • 20.Lubensky T C, Prost J. J Phys II. 1992;2:371–382. [Google Scholar]
  • 21.Ou-Yang Z, Liu J. Phys Rev Lett. 1990;65:1679–1682. [Google Scholar]
  • 22.Selinger J V, Schnur J M. Phys Rev Lett. 1993;71:4091–4094. doi: 10.1103/PhysRevLett.71.4091. [DOI] [PubMed] [Google Scholar]
  • 23.Selinger J V, MacKintosh F C, Schnur J M. Phys Rev E Stat Phys Plasmas Fluids Relat Interdiscip Top. 1996;53:3804–3818. doi: 10.1103/physreve.53.3804. [DOI] [PubMed] [Google Scholar]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES