trends in analytical chemistry, vol. 16, no. 7, 1997 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA 401 Conformation zoning of large molecules using the analytical ultracentrifuge Georges M. Pavlov* Institute of Physics, University burg, 198904, Russia 1. Introduction: Arthur J. Rowe National Cen tre for Macromolecular Hydrodynamics (Leicester Laboratory), University of Leicester, LEl 7RH, England Stephen E. Harding National Cen tre for Macromolecular Hydrodynamics (Nottingham Labora tory), University of Nottingham, LE12 5RD, England A substantial proportion of large molecules made naturally or by artificial means exist as linear chains. In biology this includes DNA, mRNA, many important classes of sugar polymers (polysaccharides) and denatured proteins. In physical science this polyvinylchloride includes polyethylene, and many important polymers used in plastics and also the many new ones being explored for use in drug delivery. Crucial to how many of these large molecules function is their conformation in so/ution( either aqueous or organic), a realm unfortunately outside the grasp of high-resolution techniques such as X-ray crystallography. We have now however devised a quick and accessible method for identifying the conformation type or “Zone” of a molecule: Zone A (extra rigid rod type); Zone B (rigid rod type); Zone C (semi-flexible type), Zone D (completely random coil) and Zone E (compact or highly branched particle). To perform this “Conformation Zoning” requires a few milligrams of material and access to one of the new types of high-speed Centrifuge which are now proliferating in academic and industrial establishments. 0 1997 Elsevier Science B.V. *Corresponding 0165-9936/97/$17.00 PUSO165-9936(97)00038-l author. conformation zones of St. PetersHow large molecules behave in solution is the subject of increasing interest amongst biological and physical scientists, and there have been several exciting developments in methodology devised to study this behaviour. One of the more interesting developments has been the revival of analytical ultracentrifugation as a molecular probe [ 1,2 1, for which there have been outstanding recent advances in instrumentation, optimising the ease and precision of data capture. Whereas most of the attention appears to have been directed towards folded structures such as proteins, the latter represents only a minority of the vast array of large molecules (“macromolecules”) a substantial proportion of which are based on a linear chain (or parallel chains) template which can be linked together by non-covalent interactions. These large linear molecules in solution, along with their folded (e.g. proteins) or branched (e.g. starch amylopectin) counterparts can adopt a variety of overall conformations depending on their chemical structure and the medium in which they are dispersed (Fig. 1A). For example a DNA can exist as a rigid rod shape or “Zone A” structure or a tightly condensed structure (Zone E) depending on the environment in which it is dispersed. As a further example, sugar polymers (polysaccharides) can have a range of conformations from extremely rigid rod structures (Zone A) through to perfectly random flexible coils (Zone D). How do we assign these conformation types? Although a small minority of macromolecules - of which globular proteins are an example - can be characterised precisely by high resolution crystallographic and magnetic resonance techniques, the majority are not accessible, particularly those which cannot be crystallized or are highly polydispersed: in biology this includes DNA and mRNA, polysaccharides - until recently the Cinderella molecules of biology - and a vast array of glycoconjugates. For these types of molecule, dilute solution ultracentrifugation-based characterisation methods are invaluable. We propose now a completely novel but nonetheless simple way to assign a molecular conformation type or “zone” 0 1997 Elsevier Science B.V. All rights reserved. 402 A trends in analytical chemistry, vol. 16, no. 7, 1997 is defined as simply M/ L and is known or can be measured from e.g. electron microscopy [ 3 ] X-ray fibre diffraction studies [ 41. M itself can be measured by a variety of techniques such as sedimentation equilibrium [ 2,5 ] or light scattering analysis [ 61. A further significant parameter (Fig. 1b) is the persistence length [ 71, a, which is a measure of how far a polymer chain “persists” along the same direction as one of its ends starts out in, before Brownian motion etc. drags it away. Its value takes into account complicated behaviour such as excluded volume effects in flexible chain “Zone D” polymers or draining effects in rigid chain “Zone A and B” polymers or both (semi-rigid or semi-flexible “Zone C” chains) [ 8 1. There are other equivalent parameters (such as the Kuhn length I), but to keep things simple we just stick with M, L, ML =MIL and a as our fundamental parameters. B 3. Sedimentation parameters Let us now consider the large amount of data that has been published on the sedimentation properties of macromolecules of known class or zone (A-E). (In general the persistence lengths vary from -200 nm for Zone A macromolecules progressively down to - 1 nm for Zone E): The key experFig. 1. (A) The five conformation zones for large molimental parameters are the sedimentation coeffifor a linear ecules. (B) Length parameters and the concentration cient SO (seconds) macromolecule. L= contour length, a= Persistence dependence parameter of the sedimentation coeffilength (defined as the projection length along the inicient, zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPO k, ( cm3 /g). SOis simply the sedimentation tial direction of a chain of length L and in the limit of rate per unit centrifugal field extrapolated to L + infinity). c + 0, and comes, along with k, from fitting the relation 1 /s = ( 1 /SO)( l+k,c) to the sedimentation coefficient s measured at a variety of molecular based on a single experiment in the analytical ultraconcentrations, c (g /cm3 ). centrifuge, with the appropriate optical system The sedimentation coefficient has to be adjusted (such as the XL-A ultracentrifuge with absorption so as to allow for buoyancy effects of the particular optics [ 5 ] and the XL-I ultracentrifuge with both solvent used during the sedimentation experiment, absorption and interferometric optical recording to give the intrinsic sedimentation coefficient [s] systems [ 161). defined by [s] = {SOQ/( 1-vpa)}, where Q is the viscosity of the solvent, po is the density of the solvent and v a parameter known as the “partial 2. Some basic macromolecular specific volume” (essential the reciprocal density characteristics of the polymer) which is well known for a wide range of polymers [ 8 land can be precisely measLet us first consider some of the basic characterured without undue difficulty (- 0.73 cm3 /g for istics of linear polymer chains in solution. Two of -0.5-0.6 cm3/g for nucleic acids proteins, the most important parameters representing a poly0.50-0.60 cm3 /g for polysaccharides and mer chain are its molecular weight (M), and its - 0.65-0.70 cm3 /g for polystyrene). contour length, zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA L. The mass per unit length (ML) trends in analytical chemistry, vol. 16, no. 7, 7997 A 403 3.5 , . 3.0 - . v4 v= i 2.5 7 r g $ 2.01.5- . Schzophyllan v Xanthan A DNA + Cellubse nitrate o Methyl-cellulose = PVP + Pullubn h Pot#yrene . Globular Proteins l.O- A h 0.5 - f 0.0 E i -0.5 I 0.5 l* I 1.0 l . . . h . ‘. I 1.5 . . I I 2.0 2.5 WWM,) B 3.5 3.0 ZONE C: Sem~Fkmble ZONE 2.5 D: Random Coil ZONE E: Globular or T 2.0 z g. 1.5 1.0 0.5 0.0 -0.5 Fig. 2. Sedimentation conformation zoning. (a) Dependence of log( k,ML) versus log([ s]IML)for 82 macromolecules of known conformation type [9-l 51: 10 Schizophyllans (ML = 215 Daltons. A-‘), 12 Xanthans (ML = 194 Daltons. A-l ) and 7 DNA’s (ML = 195 Daltons.A-I) (Zone A); 14 cellulose nitrates (ML = 55 Daltons.A-‘) (Zone 6); 6 methyl celluloses (ML = 36 Daltons. A-’ ) (Zone C); 10 polyvinylpyrrolidones (M L = 44 Daltons.A-’ ), 10 pullulans (ML = 33.8 Daltons.A-’ ) and 3 polystyrenes (ML = 41 Daltons. A-’ ), (Zone D); 10 Globular proteins (Zone E). All molecules in dilute solution conditions. (b) Corresponding “sedimentation conformation zoning” plot. 4. The conformation zone plot convenience lines - drawn simply as guides - have been used to separate each zone, it should be Now consider a plot of zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA k,M L versus [s]/ML. In stressed that the boundaries between each must be Fig. 2a we have accumulated a large amount of data regarded as a continuum (one zone merges into the [ 9-15 ] ( more than 80 data points) for molecules of next etc.) known conformation class (including our own It is quite clear from Fig. 2 that there is a definite recent data for methyl cellulose [ 111 and for pulempirical relationship between k,M L and [ s]/A4~ lulan [ 13 1) and from this empirically constructed specific for each conformation class or Zone. Measthe conformation zones of Fig. 2b. Although for urement of [S ] with k, along with knowledge of M L 404 trends in analytical chemistry, vol. 16, no. 7, 1997 of (a/b), and since ML is itself not a function of (from chemical structure, electron microscopy or x-ray scattering) is therefore quite sufficient to pin(a/b), it follows at once that the exponent will have the value 3 - ( - 1) = 4. In other words, the log-log point the conformation type of a macromolecule. plot will have a slope of 4. This procedure represents a considerable advantage over earlier methods which either could not By contrast, for perfect spheres the terms k, and ML are both invariant in (S/ML) and hence the predistinguish between flexible coils and compact spheres, or required fractionation of a macromoledicted slope of the log-log plot must be zero. Basically the difference between the two cases (long rod cule into different molecular weight species: the sedimentation conformation zoning plot distinvs. sphere) arises predominantly from the fact that guishes between the 5 zonal types without the the regression coefficient k, is a shape (extension) related function for the former case but not for the need for molecular fractionation and multiple latter. Since it is difficult to conceive of a shape measurements. Further, since modern ultracentrifuges permit the measurement of several solution which would generate friction more rapidly with concentrations in the same run ( multiple hole rotors respect to extension than an infinitely long rod, and appropriate data multiplexing), se and k, can be we may presume that all other shapes would gen[ 16 1. measured in a single experiment erate log-log plots with a slope intermediate between zero and 4. It can be seen also from Fig. 2 that the plots appear to converge to a single point: this is not Inspection of the empirical data of Fig. 2 shows a surprising since in the limit zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA L (polymer) -+ L(moclear confirmation of all these theoretical predictions. nomer unit) the distinction between the various conformation types must vanish. 6. Concluding 5. Theoretical remarks basis An accidental discovery? This empirical finding - although at first sight apparently fortuitous - on close scrutiny is no accident and has a clear theoretical basis: The problem of predicting the form of the relationship between log (k, /ML) and log ([s ]/ML) can be simplified down, for a log-log relationship, to considering simply the exponent of the first bracketed term with respect to its argument (the second bracketed term). This will be the slope of the plot. We consider two extreme cases: infinitely long rigid rods and perfect spheres: For infinitely long rods, treated as prolate ellipsoids, (of semi-axes a > b) it has been known for over 40 years [ 17 ] that the frictional ratio (the ratio of the friction coefficient of the molecule to that of a sphere of the same mass and volume) is, for a given width, directly proportional to the axial ratio (a / b), and that hence the sedimentation coefficient can be treated as a linear function of mass/unit length. From this, it immediately follows that the sizerelated term S/M L is linear in (al b)-‘, and the problem is to determine the exponent of IQ/M L with respect to (u/b)-‘. This falls out simply. From the fact that k, varies for highly elongated particles simply with the cube of the frictional ratio, and that the frictional ratio itself, as noted by Peacocke and Schachman [ 17 ] a linear function The molecular conformation itself - as represented by its Zone - will be influenced by (i) the equilibrium rigidity of the chain (ii) the (cross-sectional) diameter of the chain; (iii) the thermodynamics of polymer-solvent interaction, with the most important being the equilibrium rigidity of the chain. Measurement of simple sedimentation - the ease and precision of measurement of which is getting greater and greater with the several major instrumental developments - should be sufficient to unambiguously specify the conformation class of a macromolecule. Furthermore, monitoring any changes in conformation type of a macromolecule in response to a change in the solvent environment it finds itself in (e.g. the condensation of DNA in the presence of polycations or the effect of changing salt ion concentration on the glycoproteins which dictate the protective properties of mucus in the alimentary, tracheobronchial or reproductive systems [ 18 ] or in response to genetic alteration (e.g. of plant cell wall polysaccharides [ 191) will be considerably facilitated. Acknowledgements GMP is an Underwood Fellow of the BBSRC. We thank Professor R.J.P. Williams, FRS, Profes- 405 trendsinanalyticalchemistry,vol.16, no. 7, 1997 [91 T. Yanaki, T. Norisuye, H. Fujita, Macromolecules 13 (1980) 1462-1466. [lOI G. Pavlov, A. Kozlov, G. Martchenko, V. Tsvetkov, Vysokomol. Soedin. 24B (1982) 284-288. G. Pavlov, N. Michailova, E. Tarabukina, E. Kor[Ill References neeva, Prog. Coll. Polym. Sci. 99 (1995) 109113. [ zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA 1 ] H.K. Schachman, Nature 941 ( 1989) 259-260. [I21 K. Kawahara, K. Ohta, H. Miyamoto, S. Naka[ 21 S.E. Harding, TrAC 13 (1994) 439-446. mura, Carbohydrate Polym. 4 (1982) 335-356. [ 3 ] B.T. Stokke, A. Elgsaeter, Micron 25 ( 1994) [I31 G. Pavlov, E. Korneeva, N. Yevlampieva, Int. J. 469-49 1. Biol. Macromol. 16 (1994) 318-323. [ 41 L.E. Alexander, X-Ray Diffraction Methods in [ 141 J.M. Creeth, C.G. Knight, Biochim. Biophys. Polymer Science, Krieger, Huntington, NY, 1979. Acta 102 (1995) 549-558. [ 5 ] G. Ralston, Introduction to Analytical Ultracen[I51 C. Tanford, Physical Chemistry of Macromoletrifugation, Beckman Instruments, Palo Alto, cules, John Wiley, New York, NY, 1961. CA, 1993. 1161 A. Furst, Eur. Biophys. J. 25 (1997) 307-310. [6] S.E. Harding, D.B. Sattelle, V.A. Bloomfield [I71 A.R. Peacocke, H.K. Schachman, Biochim. Bio(Eds.), Laser Light Scattering in Biochemistry, phys. Acta 15 (1954) 198. Royal Society of Chemistry, Cambridge, 1992. [I81 P.Y. Tam, P. Verdugo, Nature 292 ( 1981) 340[ 7 ] H. Yamakawa, Modern Theory of Polymer Solu342. tions, Harper and Row, New York, NY, 1971. 1191 C.J.S. Smith, C.F. Watson, J.Ray, C.R. Bird,P.C. Morris, W. Schuch, D. Grierson, Nature 334 [ 8 ] E.H. Immergut, J. Brandrup, Polymer Handbook, (1988) 724-726. 3rd ed., Wiley Interscience, New York, NY, 1989. sor M.P. Tombs and Professor J.MG. Cowie FRS for their comments on the manuscript. Determination of sulphonated water and wastewater J. Riu, I. Schiinsee, D. Barcel6* Department of Environmental Chemistry, CID-CSIC, c/ Jordi Girona 18-26,08034 Barcelona, Spain C. Rhfols Department of Analytical Chemistry, University of Barcelona, Av/ Diagonal, 647, 08028 Barcelona, Spain An overview of current analytical methodologies for the determination of sulphonated azo dyes in environmental samples is presented. Conventional analytical methods involving liquid chromatography and mass spectrometry with the thermospray (TSP) interface are reviewed, as well as the newly developed methods using atmospheric pressure ionisation (API) interfaces. The combination of *Corresponding 0165-9936/97/$17.00 PUSO165-9936(97)00034-4 author. azo dyes in capillary electrophoresis with mass spectrometry (CE-MS) as a new alternative for the analysis and confirmation of polar compounds (and particularly the sulphonated azo dyes) in environmental samples is also discussed. Finally, the extraction and preconcentration of sulphonated azo dyes from water samples involving various solid phase extraction cartridges are commented upon. 0 1997 Elsevier Science B.V. 1. Introduction Large quantities of dyes are produced and used in diverse applications including textiles, paint pigments, printing inks and food colouring. According to recent information, Western Europe is responsible for nearly 20% of the world dye production, which rose more than 10% annually to 2.2 billion lb. ( lo9 kg) in 1994 [ 11. The textile industry is the 0 1997 Elsevier Science B.V. All rights reserved.