Amino Acids
Predicting protein decomposition: the case of aspartic-acid racemization kinetics
M. J. Collins*, E. R. Waite and A. C. T. van Duin
Fossil Fuels and Environmental Geochemistry (Postgraduate Institute), NRG, Drummond Building, University of Newcastle-upon-Tyne,
Newcastle-upon-Tyne NE1 7RU, UK
The increase in proportion of the non-biological (D-) isomer of aspartic acid (Asp) relative to the Lisomer has been widely used in archaeology and geochemistry as a tool for dating. The method has proved controversial, particularly when used for bones. The non-linear kinetics of Asp racemization have prompted a number of suggestions as to the underlying mechanism(s) and have led to the use of mathematical transformations which linearize the increase in D-Asp with respect to time. Using one example, a suggestion that the initial rapid phase of Asp racemization is due to a contribution from asparagine (Asn), we demonstrate how a simple model of the degradation and racemization of Asn can be used to predict the observed kinetics. A more complex model of peptide bound Asx (Asn+Asp) racemization, which occurs via the formation of a cyclic succinimide (Asu), can be used to correctly predict Asx racemization kinetics in proteins at high temperatures (95^1408C). The model fails to predict racemization kinetics in dentine collagen at 378C. The reason for this is that Asu formation is highly conformation dependent and is predicted to occur extremely slowly in triple helical collagen. As conformation strongly in£uences the rate of Asu formation and hence Asx racemization, the use of extrapolation from high temperatures to estimate racemization kinetics of Asx in proteins below their denaturation temperature is called into question.
In the case of archaeological bone, we argue that the D:L ratio of Asx re£ects the proportion of nonhelical to helical collagen, overlain by the e¡ects of leaching of more soluble (and conformationally unconstrained) peptides. Thus, racemization kinetics in bone are potentially unpredictable, and the proposed use of Asx racemization to estimate the extent of DNA depurination in archaeological bones is challenged.
Keywords: aspartic-acid racemization; deamidation; kinetics; dating; bone; collagen
1. INTRODUCTION
How predictable is protein preservation in the fossil record? The question remains largely unanswered despite its pertinence to current research and potential applications. Ancient DNA studies have mirrored research into ancient proteinsöpractitioners initially prospected for the oldest or most spectacular samples rather than systematically investigating the processes of long-term survival. Despite the 25-year head-start that ancient protein research has had over that of DNA, the failure to address the issue of diagenesis means that we still have not successfully moved beyond this prospecting phase.
One area of ancient protein research where much greater investment has been made into predictive models of diagenesis, is that of amino-acid racemization. The gradual increase in non-biological isomeric forms of the constituent amino acids within proteins has been used as both a dating tool and a method of palaeothermometry. Amino acids with one or more chiral carbon centres undergo isomerization (termed racemization); the rate of inter-conversion is governed by time and temperature. Thus, if temperature remains constant, the increase in
| *Author for correspondence (m.collins@ncl.ac.uk). 354, 51^64 |
the D-isomer can be used chronometrically (e.g. Goodfriend 1991); conversely if the age is known information can be obtained on palaeotemperature (e.g. Miller et al. 1997).
The most common approach has been to use kinetic models based upon isomerization of free amino acids at high temperatures to estimate racemization rates at burial temperatures. The problem with this approach (recognized early in the investigations, e.g. Bada 1972; Bender 1974) is that the rate of racemization in fossil proteins is governed not by the simple, predictable kinetics of free solution, but by the local environment of the bound residue, which changes during protein diagenesis and can be in£uenced by the burial environment. The challenge is to identify those key factors which in£uence racemization in the burial environment.
2. ASPARTIC-ACID RACEMIZATION
Of all the amino acids used for racemization analysis, it is undoubtedly aspartic acid (Asp) which has the most chequered history as a dating tool. Measurable increases in D-Asp accrue over years (Goodfriend 1992) to tens of millennia (Goodfriend1991). In the early1970s the method excited interest as a means of dating archaeological bone
51 & 1999 The Royal Society
because it required less material than early radiocarbon dating techniques (e.g. Bada & Protsch 1973). Unfortunately problems were encountered with the method. Upper Pleistocene racemization dates for Palaeoindian remains in California, which supported an early colonization of North America, became a cause ce¨ le© bre (e.g. Pollard & Heron 1996) when AMS14C dates proved that the bones were only 5100+500BP (Bada 1985). In retrospect, part of the error arose from inaccurate radiocarbon dating of the calibration material, but racemization dating in the archaeological community was nearfatally mauled. Slowly, however, the method is re-establishing itself in the `geoarchaeological’ community (e.g. Brooks et al. 1990; Goodfriend 1991 1992; Johnson & Miller 1997), and has found a valuable niche in the forensic sciences for estimating age at death (Ritz & Kaatsch 1996). Furthermore, Asp racemization is becoming more widely used as a screening method for bones prior to DNA analysis (Poinar et al. 1996; Krings et al. 1997).
3. PROBLEMS WITH THE KINETICS OF ASPARTIC ACID
The very strength of `Asp’ racemization, its unusually rapid initial rate (Goodfriend 1992), is also a drawback, in that the kinetics are atypical. The rapid initial rate of `Asp’can be illustrated by plotting relative rates of racemization for di¡erent amino acids measured in the same system (¢gure 1). Where amino acids display comparable kinetics, their rates will correlate. This is not the case for `Asp’ which has both a `fast’ and `slow’ component (¢gure 1). The atypical pattern of `Asp’ racemization has prompted a number of explanations. Smith & Evans (1980) have suggested that the change to a slower rate in heated collagen was due to the increasing proportion of more slowly racemizing free amino acids as a result of hydrolysis of peptide bonds over time. Goodfriend (1991) notes that the signal measured as Asp is actually the combined responses of Asp+asparagine (Asn), the latter decomposing to Asp upon acid hydrolysis; to avoid confusion this combined signal is hereafter referred to as Asx. Goodfriend (1991) suggested that the pattern of Asx racemization may be due to di¡erences in the rate of succinimide formation (and concomitant racemization) from Asn and Asp. Brinton & Bada (1995) argued that the e¡ect was due to di¡erences in the rate of racemization and the direct decomposition of Asn to Asp by hydrolysis of the amide group. Using pure Asn they were able to obtain kinetic patterns for Asx similar to those reported by Goodfriend et al. (1992) (¢gure 2, inset).
The kinetics of Asx racemization have proved more di¤cult to describe mathematically than those of most other amino acids. In order to produce a proportional increase in D-Asx with age, Goodfriend & Hare (1995) and Goodfriend et al. (1996) adopted power-function transformations of the D:L ratio for both mollusc and ostrich shells (e.g. ¢gure 2). Such transformations are useful in calibrated investigations to assess age over a broad range of D:L ratios. Applying the same power transformation to the data from Brinton & Bada (1995) also produces a strong correlation with time (¢gure 2). However, the transformation does not o¡er an explanation
Figure 1. Comparative rates of racemization for di¡erent amino acids from mollusc shells during arti¢cial diagenesis plotted against D:L ratio of alanine (Ala). Note that unlike methionine (Met, solid squares), proline (Pro, solid circles) and phenylalanine (Phe, open circles), the rate of increase of Asx (D) is not linear with respect to Ala, but is distorted by a rapid initial phase (data from Goodfriend & Meyer (1991)).
of the observed kinetics. Di¡erent power-functions are used for di¡erent taxaöin the case of the mollusc Mya, no transformation of the D:L ratio satisfactorily linearizes the data over the full range of values (Goodfriendetal.1996).
In addition to the di¤culties in obtaining a function to describe the complex kinetics of Asx, other problems with the application of the method include the following:
- in free solution serine and threonine racemize more rapidly than Asp (Wonnacott (1979) cited in Smith & Evans (1980)), but in proteins Asx residues have the highest rate;
- an 1800-year age for a living deep-water anemone Gerardia determined by 14C (Dru¡el et al. 1995) is apparently contradicted by an Asx date of only 250 years estimated by extrapolation from high temperature experiments (Goodfriend 1997);
- despite the fact that Asx racemization in dentine is predictable (Ritz & Kaatsch 1996), the results can be signi¢cantly in£uenced by preparation method (Collins & Galley 1998), and Asx in rat dentine appears to racemize ten times faster than human
dentine (Ohtani et al. 1995);
- in historical material the D-Asx concentration in dentine collagen all tends towards a consensus value, independent of age (Gillard et al. 1991; Carolan et al. 1997).
4. AN ALTERNATIVE APPROACH TO ASX KINETICS
As problems with the method begin to emerge, and in the absence of a satisfactory mathematical description of the kinetics, we propose an alternative approach. Instead of relying on ever more complex mathematical
Figure 2. The use of a cubic transformation of the D:L ratio for data from Goodfriend (1992) and Brinton & Bada (1995) to linearize the original kinetics (inset) and derive a slope which is proportional to age.
transformations we use simple mathematical models to describe the kinetics of a system, and then estimate unknown rates by optimization. The models describe the system under investigation as accurately as is practicably possible; the number of unknown parameters is kept to a bare minimum to reduce the number of unique solutions. Our approach can be illustrated by reference to the previously cited study of Brinton & Bada (1995).
Most racemization studies assume that the system conforms to simple ¢rst-order reversible kinetics. However, ¢gure 3 showsthat, whenonly the L-Asnto D-Asn reaction istaken into account inthe model, the initial fast racemization of D-Asx cannot be reproduced as the predicted initial reaction rate istoo slow. Asn is unstable andover the course of the experiment will degrade to Asp.When the potential for decomposition of Asn to Asp is added to the model a much better ¢t with the experimental data is obtained, faster racemizing Asn decomposing rapidly to slower racemizing Asp.The rate of Asn decompositionpredicted solely from the racemization kinetics is in reasonable agreement with the measured rate for this experiment, although these data were not used in the optimization (¢gure 3). If these data are used in the optimization, more accurate estimates oftheremainingunknownreactionratesareobtained.
The simple model illustrated in ¢gure 3 was appropriate for degradation of pure Asn in vitro, but is clearly too simple for studies of Asx racemization in proteins. Here, the most probable pathway of racemization of Asx is via an aminosuccinyl residue (Asu; Geiger & Clarke 1987). Ab initio calculations have been used to illustrate that racemization in the Asu residue can be ¢ve orders of magnitude faster than free Asp (Radkiewicz et al. 1996). Aspartic acid and Asn form Asu by nucleophilic attack on the b-carbonyl group by the NH-group of the downstream (C-terminal) peptide bond, resulting in the formation of D-Asp isomers and isoaspartyl (iAsp)
Figure 3. Illustration of the application of the modelling approach to explain the observed kinetics of increase in D-Asx
(open squares) caused by the simultaneous racemization of
Asn, Asp and the decomposition of Asn (closed circles) to Asp (data from Brinton & Bada (1995)). Lines represent output from the model when optimized using di¡erent data sets.
residues. The rate of formation of Asu is dependent upon factors which increase the deprotonation of the peptide nitrogen. These factors include high pH, (Capasso et al. 1992, 1993), high ionic strength (Capasso et al. 1991; TylerCross & Schirch 1991), high dielectric constant (Brennan & Clarke 1993) and conformational £exibility within the residues (Kossiako¡ 1988; Lura & Schirch 1988; Bongers et al. 1992; Stevenson et al. 1993; Tomizawa et al. 1995).
5. A MODEL OF PEPTIDE-BOUND ASX KINETICS
Most investigations of Asx racemization have used collagen rich tissues (bone or dentine; Appendix A).Thus, collagen would seem a suitable protein in which to model Asx racemization and test our approach; the most satisfactorystudy forour purposes is de Sol (1978) whoinvestigated therateof Asx racemizationincollagenatpH 8 between 95 and1408C.WehavedevelopedamodelofAsxracemization based upon data given in Geiger & Clarke (1987). The original model of Geiger & Clarke (1987) was developed to describe Asn decomposition and Asx racemization in a hexapeptide from adrenocorticotropic hormone and requiressomemodi¢cationforusewithcollagen.Themodi¢cations take into account (i) the impact of residues adjacent toAsn and Asp, (ii) additional information onthe rate of deamidation of Asn, and (iii) an observed equilibrium D:Lratioof 1observedbydeSol(1978)incollagen.
(a) Modi¢cation 1öthe in£uence of £anking residues
Studies on the rate of Asn deamidation in synthetic peptides have revealed that the rate of Asu formation (and hence deamidation) can vary by almost three orders
| Table 1. Relative rate of Asu formation in relationship to residues carboxyl and amino to Asn (Comparison of the enthalpy (H k J molÿ1) and entropy (S J molÿ1N K ) of energy-minimized structures in gas phase using molecular mechanics (Burkert & Allinger 1982). Rates of deamidation are from McKerrow (1973; pH 7.4, 0.02 M phosphate, 37 8C). The gas-phase energies are of limited value as they do not take into account any interactions with water, but they reveal that in general the ease of Asu formation decreases with increasing bulkiness of the downstream residue but are much less sensitive to the upstream £anking residue.) sequence H(Asn) S(Asn) H(Asu) S(Asu) k (sÿ1) DH DG310 Gly-Gly-Asn-Gly-Gly Gly-Gly-Asn-Ala-Gly Gly-Ala-Asn-Leu-Gly Gly-Ala-Asn-Ala-Gly Gly-Ile-Asn-Gly-Gly Gly-Ile-Asn-Val-Gly Gly-Ile-Asn-Leu-Gly Gly-Ile-Asn-Ala-Gly Gly-Ile-Asn-Ile-Gly Gly-Leu-Asn-Ala-Gly ÿ330.74 ÿ305.38376.79 ÿ280.02 ÿ322.00 ÿ381.06 ÿ393.86 ÿ299.47 ÿ313.71 ÿ378.39 ÿ 789.36 811.85 910.10 831.72 887.86 961.36 988.69 910.35 989.80 909.86 ÿ112.4164.27 ÿ139.49 ÿ 44.20 ÿ130.01 ÿ160.79 ÿ166.29 ÿ 70.10 ÿ74.82 ÿ154.23 ÿ 714.22 717.22 819.12 735.56 776.72 891.80 881.38 799.09 916.14 802.17 ö 9.208.402.90ö101010ÿÿÿ888 ö ö 1.60ö10ÿ8 3.7010ÿ8 ÿ218.33241.11 ÿ237.29 ÿ235.82 ÿ191.99 ÿ220.26 ÿ227.57 ÿ229.37 ÿ238.89 ÿ224.16 ÿ ÿ241.63270.45 ÿ265.50 ÿ265.62 ÿ226.44 ÿ241.83 ÿ260.84 ÿ263.86 ÿ261.73 ÿ257.54 ÿ Table 2. The distribution of residues carboxyl to Asn and Asp in bovine type I collagen, and the e¡ect of residues carboxyl to Asn on the rate of deamidation relative to Asn-Gly residue X a-1 Asn-X a-2 total a-1 Asp-X a-2 total rate relative to Glya x s.d. n Gly 6 10 22 15 13 43 1 0.8 5 Ala 2 1 5 6 1 13 80.3 86.8 21 Arg 0 1 1 5 1 11 31.2 ö 1 Prob 2 5 9 0 0 0 73.9 13.7 4 Asp 4 1 9 1 0 2 23.8 ö 1 Lys 2 2 6 2 2 6 79.7 0.4 2 Val 2 1 5 0 0 0 117.3 69.9 4 Leu 0 3 3 2 0 4 86.9 102.7 4 Ser 2 0 4 0 2 2 8.9 5.6 3 Tyr 0 0 0 1 0 2 ö ö Thr 0 0 0 1 0 2 23.8 ö 1 Gln 0 0 0 0 3 3 ö ö total 20 24 64 33 22 88 a Data from McKerrow (1973), Brennan & Clarke (1995), Tyler-Cross & Schirch (1991), Patel & Bouchardt (1990a). |
b Includes Hyp.
depending upon the primary sequence (e.g. Cleland et al. 1993). Although the rate is in£uenced by both £anking residues, the greatest in£uence is exerted by the downstream (i.e. carboxyl to Asn) residue which takes part in Asu formation; generally the bulkier the side chain of the downstream residue the less readily Asu formation occurs (Brennan & Clarke 1993; Oliyai & Borchardt 1994). We have observed the potential impact of bulky residues on the rate of Asu formation in molecular mechanics calculations of energy minimized structures (table 1).
Bovine type I collagen contains a total of 88 Asx residues. The downstream residues found in collagen are listed in table 2, along with average deamidation (and hence Asu formation) rate relative to an Asn-Gly sequence; almost half of theAsn and Asp residues are adjacent to Gly, (the most reactive combination). The large standard deviations for each of the residues re£ects the e¡ect of other factors, such as inter-study variations in peptide length, bu¡er concentration, pH and adjacent primary structure, (see source references for further details).The data in table 2 are used to derive estimates of the reduction in rate of Asu formation caused by downstream residues. In the model the di¡erence between labile (i.e. Asx-Gly) and bulky residues carboxyl to the Asx residue are accounted for by fast and slow pools (¢gure 4).
(b) Modi¢cation 2öestimate of deamidation rate
Deamidation to form L-Asu is the ¢rst step in the decomposition of Asn residues. We believe that the activation energy (Ea) for deamidation estimated by
Geiger & Clarke (1987; Ea88.7kinetickJmolÿparameter1) is too low.for
We have derived a consensus
kdeamidation using the high temperature data of Sinex (1960; pH 7.35) for collagen and the rates of Bongers et al. (1992; pH 8.0) and Geiger & Clarke (1987; pH 7.4) for Asn-Gly residues in oligo- and polypeptides. The temperature dependence in the three studies is remarkably consistent (¢gure 5) and yields a consensus Ea of 93.6 kJmolÿ1 (cf.
Figure 5. Rate of deamidation of Asn in proteins (triangles) and peptides (squares and circles). Note the very similar rates of deamidation observed in collagen and peptides. In the case of collagen, soy protein and caseinate rates are determined from the release of ammonia and thus will include a contribution from glutamate. Note the in£uence of £anking residues, the rate of deamidation of Asn-Gly is almost two orders of magnitude faster than Asn-Pro in Val-Tyr-Xxx-Asn-Xxx-Ala peptides.
Patel & Borchardt 1990b: 90.8kJmolÿ1; Sendero¡ et al. 1994: 98.9kJmolÿ1). The rate of deamidation in collagen is much higher than reported rates for other proteins, re£ecting the abundance of Asn-Gly residues in collagen sequences (¢gure 5).
(c) Modi¢cation 3öequilibrium D:L ratio
Deviations from the expected 1:1 D:L ratio are anticipated in peptide bound residues if the local environment Table 3. Values for activation energy (Ea) and the preexponential factor (A) used in collagen racemization model, where Xxx is any residue other than Gly
L-AsnImideisopeptide imideisopeptideimide 93.793.690.8 3.961.993.2310101011108 8.014.949.83101010896
Imide peptide 94.1 1.281011 3.18109
within the protein structure favours the D-Asx residue over its L-counterpart. Fujii et al. (1994) have reported the phenomenon of `stereoinversion’, in which certain sites appear to rapidly accumulate D-Asp, such that the D:L ratio at one site in aA-crystallin was 5.7 in iAsp residues of an 11-month-old human lens. In the Val-TyrPro-Asn-Gly-Ala peptide used by Geiger & Clarke (1987) the equilibrium D:L ratio for Asp was estimated to be 0.38 at 378C. It is much more surprising to ¢nd non-unity equilibrium ratios in short peptides than in proteins; we believe that this phenomenon requires further investigation as it is never reported in the geochemical literature. It may be that non-unity equilibrium ratios are common at individual Asx sites but that in a polypeptide they broadly even out. A more likely and probably more signi¢cant cause in geochemical studies is that investigations extend over longer timescales of protein degradation ultimately leading to the release of free amino acids (whose equilibrium D:L ratio is 1). As equilibrium ratios were unity in the modelled data sets we have chosen to use the L-Asx kinetic parameters for both L- and D-residues (table 3). The Ea of reactions involving D-Asx residues (illustrated in ¢g. 4c of Geiger & Clarke (1987)) are markedly higher than
those for L-Asx residues, in the Val-Tyr-Pro-Asn-Gly-Ala peptide (Geiger & Clarke 1987), favouring D-Asx at high temperatures and L-Asx at low temperatures (e.g D:L ratio 2.05 at 1408C but 0.24 at 118C). Our modi¢cation ensures that equilibrium ratios for our model are always unity, but highlights the lack of su¤cient data on the kinetics of D-Asx residues.
6. TESTING THE MODEL
Havingadapted the model of Geiger & Clarke (1987) for collagen, its predictions canbe comparedwiththe observed racemization rates reported by de Sol (1978). Despite (or perhaps because of) the fact that additional information, such as the rate of deamidation or release of free amino acids, is not available and therefore cannot be tested, the shapesofthecurvesareremarkablysimilar(¢gure6).
The observed break in slope, observed by de Sol (1978), for collagen is predicted by the model and derives from the di¡erent rates of Asu formation from Asn and Asp, and also the in£uence of fast versus slow Asx-Xxx residues. Brinton & Bada (1995) are correct in believing that the initial racemization rate of Asx can be explained by the contribution and decomposition of Asn. However this is not due to the accelerated rate of racemization of free Asn, but rather the greater propensity of some residues, notably (i) Asn-Xxx, and (ii) Asx-Gly, to form Asu than Asp-Xxx residues. The model also predicts that the break in slope occurs at lower D-Asx values at lower temperatures; this temperature dependence can be clearly seen if the model is run assuming 100% Asn-Gly at extremes of temperature (¢gure 7).
If the model works for collagen, can it work for other proteins? If the primary structure of a protein is known, the model should be able to make a reasonable prediction of its racemization rate at high temperature. Changes must be made in the model to re£ect the di¡erent proportions of Asx-Gly to Asx-Xxx, Asn to Asp and also the dampening e¡ect of residues downstream of the Asx. In the case of a-crystallin the model successfully predicts a slower racemization rate than for collagen, which is due to the lower concentrations of (i) Asn and (ii) Gly carboxyl to Asx (¢gure 6).
Despite the success of the model in predicting racemization rates from high temperature experiments, it fails to predict rates in mineralized collagen. Figure 8 illustrates data on racemization of Asx in total human dentine and dentine collagen. The rates are transformed to yield straight lines with a slope 2k assuming reversible ¢rst order kinetics. Correlations between the increase in %DAsx with time are high for total dentine. However the observed rate is 350 times slower than the prediction of the model.
The failure of our model to successfully predict successfully the rate of Asx racemization in vivo illustrates the key problem with the racemization method. The model works for short peptides (Geiger & Clarke 1987) and proteins at high temperatures (de Sol 1978; Masters 1982) for one key reasonöin small peptides and denatured proteins the peptide backbone is conformationally unconstrained. The formation of the Asu places considerable strain on the surrounding peptide backbone and thus secondary and higher order structure severely dampen the rate of deamidation (e.g. Kossiako¡ 1988; Stevenson
Figure 7. Predicted racemization kinetics of Asn-Gly rich protein at high and low temperatures. Compare the position of the break in slope, due to the lower activation energy of deamidation. This is a much more signi¢cant process in high temperature experimental diagenesis than it is at geochemically signi¢cant temperatures. (a) 150 8C; (b) 0 8C.
et al. 1993) and racemization. Collagen, although the protein most commonly analysed in archaeological studies of Asx racemization (Appendix A), o¡ers one of the most extreme examples of this e¡ect. Glycine is carboxyl to almost half of the Asn and Asp residues (table 2), thus denatured collagen is anticipated to display high rates of deamidation and racemization. The rates of deamidation of collagen at high temperatures (Sinex 1960), are similar to those for Asn-Gly residues in oligoand polypeptides (¢gure 5). Within the extreme conformational constraints imposed by the collagen triple helix rates are anticipated to be much slower; Sinex (1960) reported that below the shrinkage (i.e. denaturation) temperature of collagen, the deamidation rate was too slow to be estimated accurately.
Figure 8. Illustration of the use of the model to predict the rate of racemization of Asx in total human dentine (Ohtani &
YamamotoThe high levels1991of; kD-AsxDL Asxfor0.65the; kDLCNBr10Asxÿ3, 3digestsR2100.9773)ÿ5of, Rdentine2 0.0037).and dentine collagen (Cloos 1995
collagen are caused by the method of isolation. An estimate based upon the rate of racemization of free Asp (Bada 1971) is more similar to the rate of human dentine than the model, although this is a coincidence.
Using molecular dynamics (MD) we have estimated that the rate of Asn deamidation by Asu formation at 378C would be some 10 000-fold lower in an extended achain than in random coil gelatin (van Duin & Collins 1998). So slow do we predict the rate of Asu formation to be, that deamidation of Asn in native collagen will probably occur by direct hydrolysis. This suggestion is of relevance to the leather industry. It is common practice to lime hides, as the resulting deamidation of Asn to Asp improves the e¤ciency of chrome tanning (used to stabilize the helix). Introduction of iAsp residues into the achain, which results from deamidation via Asu, would lead to local disruption of the triple helix; if deamidation conditions are chosen to prevent signi¢cant gelatinization this problem should not occur.
7. ASX RACEMIZATION IN ARCHAEOLOGICAL BONE AND DENTINE
The mean D:L ratio from an insoluble fraction of a
20 000-year-old bone from Taishaku Konondo Cave Site, Japan (estimated burial temperature 198C) is only 0.084 (Matsu’ura & Ueta 1980). The same amount of racemization which was measured over 20 000 years in the `collagen’ extract is estimated (from our model) to accumulate in gelatin in less than a year. What relevance does our model have to Asx racemization in archaeological bone?
On the evidence of our MD calculations we believe that racemization is unlikely to occur in triple helical collagen below its denaturation temperature (Tm); Tm collagen)(demineralized1508Ccollagen)(M. J.Collins,688C, unpublishedTm (mineralizeddata).
This conclusion is supported by a number of lines of evidence: (i) rates of Asx racemization reported by Ohtani & Yamomoto (1992) for dentine at 1408C are 1000 times slower than gelatinized collagen at the same temperature (de Sol 1978); (ii) Julg et al. (1987) failed to measure racemization below the melt temperature of collagen in the laboratory; (iii) Cloos (1995) failed to observe any increase in %D-Asx in dentine collagen isolated by CNBr digests of human teeth (¢gure 8); (iv) Weiner et al. (1980) observed a correlation between the proportion of gelatin in parchment and the amount of DAsx; some of the D:L ratios in triple helical rich regions of the parchments were of the same order as modern collagen.
If there is no racemization in the collagen triple helix, what does the increase in D-Asx in archaeological bones actually represent? Only 90% of the protein in bone (Tri¤t 1980) and dentine (Linde 1989) is collagen. Some of the so-called non-collagenous proteins (NCPs), notably phosphophoryns in dentine, have sequences rich in Asp-Ser (Ritchie & Wang 1996; Hirst et al. 1997) which should be prone to rapid racemization (Masters 1985). Many studies of dentine proteins have observed a much faster rate of racemization in the soluble protein fraction (e.g. Ohtani & Yamomoto 1991; Ritz et al. 1993), the observed rate being dependent upon the extent of contamination with collagen (Collins & Galley 1998).
These soluble proteins contribute to the increase in DAsx seen in dentine with time, but risk being leached from bones and teeth in the burial environment; no systematic studies have investigated their survival in archaeological bone.
The insoluble fraction of dentine and bone is not composed solely of collagen. Some NCPs are very tightly associated with collagen and are not removed upon extraction (e.g. Stetler-Stephenson & Veis 1986; Fujisawa et al. 1994). Even if pure collagen is isolated we would anticipate racemization in the non-helical (i.e. telopeptide) regionsöiAsp residues have been observed in telopeptides in vivo (Garnero et al. 1997). Complete racemization of the telopeptides would yield a D:L ratio of 0.09, which would be achieved in about 500 years at 118C (i.e. typical UK burial temperatures). We envisage that in the absence of any microbial in£uence (Child et al. 1993) the loss of faster racemizing soluble proteins and the complete racemization of Asp in the telopeptides will result in a trend towards a consensus value for all mineralized collagen over periods of decades to centuries. Both Gillard et al. (1991) and Carolan et al. (1997) observed such trends in historical material.
As collagen degrades over time and the triple helix denatures, the proportion of D-Asx will increase. D-Asx will accumulate at £exible frayed ends of the helix following chain scission (Collins et al. 1995) or in gelatinized regions following wholesale denaturation (e.g. cooking). Further scission will tend to preferentially release these racemized residues if, as evidence suggests, hydrolysis targets the £exible regions of a peptide backbone (e.g. Hensel et al.1992; Mu« ller & Heidemann1993). Thus racemization analysis of the insoluble fraction of bone will indicate the relative proportion of denatured to triple helical collagen, itself a partly time^temperature dependent phenomenon. So called `racemization kinetics’observed in bone in burial environments canbe simpli¢ed to the model illustrated in ¢gure 9. Almost any value of D-Asx can be produced by varying the rate of generation of denatured collagen kdenaturation or the rate of loss of the denatured (and racemized) residues kleaching. The analysis of insoluble (or high molecular weight) extracts, which is being promoted as a re¢nement to the Asx dating in bone, may hold some promise (e.g. Julg et al. 1987; Elster et al. 1991; El Mansouri etal.1996). However the method would be readily compromised by processes (such as condensation reactions; Collins et al. 1992) which retain denatured peptides within the insoluble residue.
8. ASX RACEMIZATION IN BONE AS A PROXY FOR DNA DEPURINATION
Bada et al. (1994, p. 3131) noted that `the rate of (DNA) depurination and aspartic acid racemization at neutral pH are nearly identical’. Subsequently, Poinar et al. (1996) tested the use of Asx racemization to screen a small sample of ancient tissues for ampli¢able DNA. Despite the success reported in their study we believe that there cannot be more than a weak correlation between Asx racemization in bone and DNA depurination. This is because (i) the mechanism of racemization of free Asp in vitro is not equivalent to that of peptide-bound Asx, (ii) the correspondence between the rate of racemization of free Asp and Asx in human dentine is mere coincidence (e.g. the reported rate in rat dentine is tenfold higher; Ohtani et al. 1995), (iii) racemization of Asx is highly conformation dependent but the impact of conformation on DNA depurination is unknown, (iv) the measured extent of racemization of Asx in bone is dependent upon the fraction analysed and the extent of leaching of the soluble fraction (e.g. ¢gure 9). Thus, where two bones of identical age have experienced the same extent of collagen degradation, the more heavily leached sample will have the lower D:L ratio and will appear to be the best for DNA ampli¢cation.
9. ASX RACEMIZATION IN SHELLS
We have illustrated the problem of using Asx racemization in bones, where the presence of £anking residues in gelatin has the e¡ect of accelerating racemization but this is set against the severe conformational constraint of the intact collagen triple helix.What of racemization in other calci¢ed tissues? One problem with application to systems such as mollusc shells or ratite eggshells, is that at present the sequence of most proteins is unknown. Indeed in all cases Asx analyses have been conducted on a mixture of proteins. As can be seen from table 2, di¡erences in
Figure 10. An Arrhenius plot of published kinetics of Asx racemization (data in Appendix A). Note that all experiments yield broadly similar rates and apparent activation energies at high temperatures but that (due to conformational constraints) the rates are slower at lower temperatures (540 8C). This is particularly true of collagen in which Asu formation is severely restricted in the triple helix.
primary sequence can alter the rate of deamidation and hence Asu formation by two orders of magnitude.
If the rate of Asx racemization in proteins is strongly dependent upon conformation, what does this mean for attempts to derive rate constants by extrapolation from high temperatures? Investigations of thermophilic bacteria reveal the importance of conformation in the rate of deamidation and succinimide formation (Daniel et al. 1996). Since the denaturation temperatures of the proteins in biogenic carbonates are unknown, it would seem impossible to use extrapolations to predict low temperature racemization rates. If the rates of initial Asx racemization in high and low temperature studies are compared on an Arrhenius plot, it can be seen that the spread of rates broadens dramatically at lower temperatures, and is slowest for conformationally constrained proteins such as collagen (¢gure 10).
Because conformation will have a greater e¡ect at lower temperatures extrapolations from high temperature studies will tend to overestimate low temperature rates (and thus underestimate ages derived from Asx racemization at low temperatures). The Asx estimated age of 250years for the deep-sea anemone Gerardia made by Goodfriend (1997) using an extrapolation approach is probably of this type as it con£icts with the radiocarbon age of 1800 years (Dru¡el et al. 1995). Correspondingly, if rates of racemization from low temperatures are incorporated in estimates of activation energies they will tend to overestimate the Ea as they do not take account of the in£uence of conformation at lower temperatures. This explains the high Ea calculated by Goodfriend & Meyer (1991) from heating modern shells (121.4kJmolÿ1; cf. Fujii et al. (1996), average value of 109.2kJmolÿ1). The importance of conformational constraints on Asx racemization will diminish as the protein degrades and as the proportion of free amino acids increases. A lower Ea was obtained by Goodfriend & Meyer (1991) when the kinetic parameters were derived from heated fossil material (110.2kJmolÿ1).
10. HOW PREDICTABLE IS PROTEIN PRESERVATION?
Asx racemization neatly illustrates the di¤culties of predicting the long-term rates and patterns of diagenesis. Despite the large number of experiments which have been conducted to derive kinetic parameters for Asx racemization, when applied to fossil material the pattern and rates are unpredictable. However greater appreciation of the underlying mechanisms which contribute to the increase in D-Asx help identify many of these `unusual’ features. The use of simple models which describe the key reactions can enable comparison between apparently contradictory rates derived from di¡erent proteins and between high and low temperature experiments. In the light of this study we would conclude the following:
- estimating rates of Asx racemization by extrapolation from high temperatures is to be avoided. At low temperatures, Asx racemization will remain unpredictable due to the overwhelming in£uence of protein conformation of the rates of bound residues;
- rates estimated from (degraded) fossil material may produce more accurate estimates of Ea as the conformational e¡ect will be reduced;
- although they are the most widely used archaeological materials for Asx dating, bones and teeth are probably the worst materials to choose, due to the extreme conformational e¡ect of the collagentriple helix;
- leaching distorts the kinetics by removing the peptides and free amino acids, which racemize more predictably. If the intra-crystalline fraction is used (Sykes et al. 1995), the e¡ect of leaching can be
APPENDIX A
Table A1. Data on racemization rates used in ¢gure 10 avoided;unfortunatelythisapproachisnotapplicableto bone and dentine (see Hare 1980).
11. SUMMARY
- Mathematical models can be used to derive unknown rate constants in a simple system if the chemistry of the reactions is known.
- A mathematical model of Asx (Asn+Asp) racemization via Asu formation of Geiger & Clarke (1987) was modi¢ed for use with collagen. The modi¢cations included the in£uence of £anking adjacent residues, deamidation rate of collagen and an observed D:L equilibrium ratio of 1. The model successfully predicted racemization rates for collagen between 95^1408C but not at 378C.
- Asx racemization in collagen at low temperatures is not predicted by the model due to the in£uence of conformation on rate of Asu formation.
- The D:L ratio in bones re£ects the ratio of nonhelical (high proportion of D-Asx) to triple helical (L-Asx) collagen. Initially the D:L ratio will trend towards a value of 0.09, re£ecting complete racemization of the non-helical telopeptides.
- Over archaeological time the D:L ratio will increase due to the decreasing proportion of triple helical collagen. The D:L ratio itself will re£ect the rate of loss of helix (dependent on time and temperature) and leaching of soluble material (dependent on burial environment).
- Asx racemization is not a reliable proxy for the extent of DNA depurination in bones.
(vii)Asx racemization rates cannot be extrapolated from high temperatures to those below the denaturation temperature of the protein under investigation.
This work was supported by NERC grant GST/02/1017 and a Royal Society award to M.J.C., a University of Newcastle studentship to E.R.W. and TMR grant no. ERBFMBICT971871 to A.v.D. The manuscript was improved by the thoughtful comments of Dr Glenn Goodfriend and an anonymous referee.
(Where the rates are at variance with the published rate, we have recalculated the rate from the available data to estimate the initial rapid phase of Asx racemization. Some temperatures were not given and these have been estimated by us from data given in Meeson et al. (1995), Sellers et al. (1995) (indicated by *) or The National Oceanographic Data Center Oceanographic Pro¢le Database (http://www.nodc.noaa.gov, July 1998, indicated by {).)
sample details 8C rate (s71) environment author
| dentine dentine dentine dentine dentine bone bone bone bone bone bone | human total human total human vertical section human horizontal section rat boiled boiled heated in water heated in water heated in water heated in water | 37 37 37 37 37 100 120 75 85 95 105 | 1.922.06 101010777111111 2.311.757.92 10101077711116 2.22 1075 2.746.935.111.851.44 1010101077777776 | in vivo dentine in vivo dentine in vivo dentine in vivo dentine in vivo dentine in distilled water in distilled water in distilled water in distilled water in distilled water in distilled water | Ohtani & Yamomoto (1991) Shimoyama & Harada (1984) Saleh et al. (1993) Saleh et al. (1993) Ohtani et al. (1995) Skelton (1983) Von Endt (1980) M. J. Collins, unpublished data M. J. Collins, unpublished data M. J. Collins, unpublished data Hare (1980) |
| bone bone bone bone bone bone bone bone collagen `collagen’ `collagen’ `collagen’ `collagen’ `collagen’ `collagen’ `gelatin’ gelatin gelatin gelatin gelatin gelatin gelatin gelatin dentine tissue tissue tissue protein protein protein protein protein polypeptide peptides peptides peptides peptides peptides peptides peptides peptides peptides peptides peptides peptides peptides peptides peptides anemone anemone anemone anemone anemone eggshell gastropod gastropod gastropod gastropod bivalve bivalve coral coral foraminifera foraminifera foraminifera foraminifera foraminifera foraminifera | Wadi Halfa, Sudan Mindano, Phillipines La Jolla, California Olduvai, Kenya Murray Springs, Arizona Nasera Rock Shelter Mt Carmel, Israel Hungarian collections CNBr collagen digest acid insoluble fraction parchment scrolls German insoluble fraction Levant insoluble fraction Egyptian insoluble fraction Japanese insoluble fraction parchment scrolls heated pH 8.0 initial rate heated pH 8.0 slow rate heated pH 8.0 initial rate heated pH 8.0 slow rate heated pH 8.0 initial rate heated pH 8.0 slow rate heated pH 8.0 initial rate human soluble fraction blood cells inter-vertebral disks eye lens aggrecan, articular Cartilage lysozyme, heated egg white lysozyme, heated egg white acrystallin, heated eye lens gcrystallin, heated eye lens [Asp]n heated Leu-Asp-Ala Leu-Asp-Ala Leu-Asp-Ala Leu-Asp-Ala Leu-Asp-Ala Leu-Asp-Ser Leu-Asp-Ser Leu-Asp-Ser Leu-Asp-Ser Leu-Asp-Ser Glu-Asp-Leu Glu-Asp-Leu Glu-Asp-Leu Glu-Asp-Leu Glu-Asp-Leu Gerardia 14C estimate Gerardia kinetic estimate Gerardia heated Gerardia heated Gerardia heated Struthio Triodopsis Trochoidea museum colltn. Trochoidea Entemnotrochus Mya Hiatella Porites slow rate Porites initial rate Bulimina Orbulina Globorotalia Pulleniatina Pulleniatina Pulleniatina | 26 27 16 25 17 25 19 16* 37 37 22.4 13* 19* 22* 19* 22.4 95.5 114.6 114.6 122.5 122.5 140.5 140.5 37 37 37 37 37 100 100 96 96 100 90 80 70 60 50 90 80 70 60 50 90 80 70 60 50 12.5 12.5 60 80 100 80 19 19 106.5 60 100 100 24.3 24.3 7 4 4 1.8{ 1.8{ 2{ | 3.834.69 101010771212 4.59 10771312 1.432.345.74 101010777121313 5.209.511.58 101010777121312 1.583.806.34 101010777131414 6.971.905.39 101010777131412 3.42 1076 6.633.588.471.911.01 10101010107777755555 8.403.473.728.28 101010107777411911 6.845.45 1010711 5.70 1077115 4.941.127.946.31 1010101077776676 1.143.991.354.40 10101010107777767788 6.02 3.01 9.97 101010777778 3.41 8.683.701.04 101010777887 9.102.79 10107777 1.013.245.79 1010107778913 8.245.961.605.671.69 10101010101077777712876711 2.68 1.43 10711 3.001.58 10107768 2.881.273.174.214.125.702.666.092.734.272.95 101010101010101010107777777777116141014136131314 | archaeological bone archaeological bone archaeological bone archaeological bone archaeological bone archaeological bone archaeological bone archaeological bone in vivo dentine in vivo dentine Dead Sea Scrolls archaeological bone archaeological bone archaeological bone archaeological bone Dead Sea Scrolls pH 8 0.05M phosphate bu¡er phosphate bu¡er phosphate bu¡er phosphate bu¡er phosphate bu¡er phosphate bu¡er phosphate bu¡er in vivo in vivo in vivo in vivo pH 8 pH 6 unknown unknown | King & Bada (1979) King & Bada (1979) King & Bada (1979) King & Bada (1979) King & Bada (1979) King & Bada (1979) King & Bada (1979) Csapo etal. (1994) Cloos (1995) Gillard etal. (1991) Weiner et al. (1980) Elster et al. (1991) Elster et al. (1991) Kimber & Hare (1992) Matsu’ura & Ueta (1980) Weiner et al. (1980) de Sol (1978) de Sol (1978) de Sol (1978) de Sol (1978) de Sol (1978) de Sol (1978) de Sol (1978) Masters (1985) Brunauer & Clarke (1986) Ritz & Schutz (1993) Garner & Spector (1978) Maroudas etal. (1998) Zhao et al. (1989) Zhao et al. (1989) Masters (1982) Masters (1982) |
| pH 7.7 phosphate bu¡er Steinberg etal. (1984) | |||||
| pH 7 0.1M phosphate bu¡er phosphate bu¡er phosphate bu¡er phosphate bu¡er phosphate bu¡er phosphate bu¡er phosphate bu¡er phosphate bu¡er phosphate bu¡er phosphate bu¡er phosphate bu¡er phosphate bu¡er phosphate bu¡er phosphate bu¡er seawater seawater in distilled water in distilled water in distilled water carbonate skeleton carbonate skeleton carbonate skeleton carbonate skeleton carbonate skeleton carbonate skeleton carbonate skeleton carbonate skeleton carbonate skeleton carbonate skeleton carbonate skeleton carbonate skeleton carbonate skeleton carbonate skeleton carbonate skeleton | Fujii et al. (1996) Fujii et al. (1996) Fujii et al. (1996) Fujii et al. (1996) Fujii et al. (1996) Fujii et al. (1996) Fujii et al. (1996) Fujii et al. (1996) Fujii et al. (1996) Fujii et al. (1996) Fujii et al. (1996) Fujii et al. (1996) Fujii et al. (1996) Fujii et al. (1996) Fujii et al. (1996) Goodfriend (1997) Goodfriend (1997) Goodfriend (1997) Goodfriend (1997) Goodfriend (1997) Goodfriend & Hare (1995) Goodfriend (1992) Goodfriend (1991) Goodfriend (1992) Goodfriend et al. (1995) Goodfriend et al. (1996) Goodfriend et al. (1996) Goodfriend et al. (1992) Goodfriend et al. (1992) Ne¨ methy (1995) Ne¨ methy (1995) Ne¨ methy (1995) Harada etal. (1996) Harada etal. (1996) Harada etal. (1996) | ||||
bone heated in water 105 9.17 1077 in distilled water Hare (1980)
Phil.Trans. R. Soc. Lond. B (1999)
REFERENCES
Bada, J. L. 1971 Kinetics of the non-biological decomposition and racemization of amino acids in natural waters. Adv. Chem. Ser. 106, 309^331.
Bada, J. L. 1972 Kinetics of racemization of amino acids as a function of pH. J. Am. Chem. Soc. 94, 1371^1373. Bada, J. L. 1985 Aspartic acid racemization ages of California Paleoindian skeletons. Am. Antiquity 50, 645^647.
Bada, J. L. & Protsch, R. 1973 Racemization reaction of aspartic acid and its use in dating fossil bones. Proc. Natn. Acad. Sci. USA 70, 1331^1334.
Bada, J. L., Wang, X. S., Poinar, H. N., Pa« a« bo, S. & Poinar, G. O. 1994 Amino acid racemization in amber-entombed insectsöimplications for DNA preservation. Geochim. Cosmochim. Acta 58, 3131^3135.
Bender, M. L. 1974 Reliability of amino acid racemisation dating and palaeotemperature analysis on bones. Nature 252, 378^379.
Bongers, J., Heimer, E. P., Lambros, T., Pan, Y. C. E., Campbell, R. M. & Felix, A. M. 1992 Degradation of aspartic acid and asparagine residues in human growth hormone releasing factor. Int. J. Peptide Prot. Res. 39, 364^374.
Brennan, T. V. & Clarke, S. 1993 Spontaneous degradation of polypeptides at aspartyl and asparaginyl residuesöe¡ects of the solvent dielectric. Prot. Sci. 2, 331^338.
Brennan, T. V. & Clarke, S. 1995 E¡ect of adjacent histidine and cysteine residues on the spontaneous degradation of aspartyl and asparaginyl containing peptides. Int. J. Peptide Prot. Res. 45, 547^553.
Brinton, K. L. & Bada, J. L. 1995 Aspartic acid racemization and protein diagenesis in corals over the last 350 yearsö comment. Geochim. Cosmochim. Acta 59, 415^416.
Brooks, A. S., Hare, P. E., Kokis, J. E., Miller, G. H., Ernst, R. D. & Wendorf, F. 1990 Dating Pleistocene archaeological sites by protein diagenesis in ostrich eggshell. Science 248, 60^64.
Brunauer, L. S. & Clarke, S. 1986 Age-dependent accumulation of protein residues which can be hydrolyzed to Daspartic acid in human erythrocytes. J. Biol. Chem. 261, 12 538^12 543.
Burkert, U. & Allinger, N. L. 1982 Molecular mechanics. Washington, DC: American Chemical Society.
Capasso, S., Mazzarella, L. & Zagari, A. 1991 Deamidation via cyclic imide of asparaginyl peptides: dependence on salts, bu¡ers and organic solvents. Peptide Res. 4, 234^238.
Capasso, S., Mazzarella, L., Sica, F., Zagari, A. & Salvadori, S. 1992 Spontaneous cyclization of the aspartic acid side chain to the succinimide derivative. J. Chem. Soc.öChem. Comm. II, 919^921.
Capasso, S., Mazzarella, L., Sica, F., Zagari, A. & Salvadori, S. 1993 Kinetics and mechanism of succinimide ring formation in the deamidation process of asparagine residues. J. Chem. Soc.öPerkinTrans. II, 679^682.
Carolan, V. A., Gardner, M. L. G., Lucy, D. & Pollard, A. M. 1997 Some considerations regarding the use of amino acid racemization in human dentine as an indicator of age at death. J. Forensic Sci. 42, 10^16.
Child, A. M., Gillard, R. D. & Pollard, A. M. 1993 Microbially induced promotion of amino acid racemization in bone: isolation of the micro-organisms and detection of their enzymes. J. Archaeol. Sci. 20, 159^168.
Cleland, J. L., Powell, M. F. & Shire, S. J. 1993 The development of stable protein formulations: a close look at protein aggregation, deamidation, and oxidation. Crit. Rev. Therapeut. Drug Carrier Syst. 10, 307^377.
Cloos, P. A. C. 1995 Age related changes in proteins in human dentine. MPhil thesis, Det Kongelige Danske Kunstakademi.
Collins, M. J. & Galley, P. 1998 Towards an optimal method of archaeological collagen extraction; the in£uence of pH and grinding. Ancient Biomolecules. (In the press.)
Collins, M. J., Westbroek, P., Muyzer, G. & de Leeuw, J. W. 1992 Experimental evidence for condensation reactions between sugars and proteins in carbonate skeletons. Geochim. Cosmochim. Acta 56, 1539^1544.
Collins, M. J., Riley, M., Child, A. M. & Turner-Walker, G. 1995 A basic mathematical simulation of the chemical degradation of ancient collagen. J. Arch. Sci. 22, 175^183.
Csapo,¨ J., Csapo¨ -Kiss, Ne¨ methy, S., Folestad S., Tivesten, A. & Martin, T. G. 1994 Age determination based on amino acid racemization: a new possibility. Amino Acids 7, 317^325.
Daniel, R. M., Dines, M. & Petach, H. H. 1996 The denaturation and degradation of stable enzymes at high temperatures. Biochem. J. 317, 1^11. de Sol, B. S. 1978 Factors a¡ecting the extent of racemization of peptide and protein bound amino acids: positioning e¡ects and kinetic studies. MS thesis, Utah State University, Logan, UT.
Dru¡el, E. R. M., Gri¤n, S., Witter, A., Nelson, E., Southson, J., Kashgarian, M. & Vogel, J. 1995 Gerardia: bristlecone pine of the deep sea ? Geochim. Cosmochim. Acta 59, 5031^5036.
Elster, H., Gil-Av, E. & Weiner, S. 1991 Amino acid racemization of fossil bone. J. Arch. Sci. 18, 605^617.
Fujii, N., Satoh, K., Harada, K. & Ishibashi, Y. 1994 Simultaneous stereoinversion and isomerization at speci¢c aspartic acid residues in a-a-crystallin from human lens. J. Biochem. 116, 663^669.
Fujii, N., Momose, Y. & Harada, K. 1996 Kinetic study of racemization of aspartyl residues in model peptides of a-a-crystallin. Int. J. Peptide Prot. Res. 48, 118^122.
Fujisawa, R., Zhou, H.-Y. & Kuboki, Y. 1994 In vitro and in vivo association of dentin phosphophoryn with a1CB6 peptide of type I collagen. ConnectiveTissue Res. 31, 1^10.
Garner, W. H. & Spector, A. 1978 Racemization in human lens: evidence of rapid insolubilization of speci¢c polypeptides in cataract formation. Proc. Natn. Acad. Sci. USA 75, 3618^3620.
Garnero, P., Fledelius, C., Ginetys, E., Serre, C.-M., Vignot, E. & Delmas, P. D. 1997 Decreased b-isomerization of the Cterminal telopeptide of type I collagen a1 chain in Paget’s disease of bone. J. Bone Mineral Res. 12, 1407^1415.
Geiger, T. & Clarke, S. 1987 Deamidation, isomerization, and racemization at asparaginyl and aspartyl residues in peptides. Succinimide-linked reactions that contribute to protein degradation. J. Biol. Chem. 262, 785^794.
Gillard, R. D., Hardman, S. M., Pollard, A. M., Sutton, P. A. & Whittaker, D. K. 1991 Determinations of age at death in archaeological populations using the D/L ratio of aspartic acid in dental collagen. Archaeometry 1990, 637^644.
Goodfriend, G. A. 1991 Patterns of racemization and epimerization of amino acids in land snail shells over the course of the Holocene. Geochim. Cosmochim. Acta 55, 293^302.
Goodfriend, G. A. 1992 Rapid racemization of aspartic acid in mollusc shells and potential for dating over recent centuries. Nature 357, 399^401.
Goodfriend, G. A. 1997 Aspartic acid racemization and amino acid composition of the organic endoskeleton of the deepwater colonial anemone Gerardia: determination of longevity from kinetic experiments. Geochim. Cosmochim. Acta 61, 1931^1939.
Goodfriend, G. A. & Hare, P. E. 1995 Reply to comment by K. L. F. Brinton and J. L. Bada on `Aspartic acid racemization and protein diagenesis in corals over the last 350 years’. Geochim. Cosmochim. Acta 59, 417^418.
Goodfriend, G. A. & Meyer, V. R. 1991 A comparative-study of the kinetics of amino-acid racemization epimerization in fossil and modern mollusk shells. Geochim. Cosmochim. Acta 55, 3355^3367.
Goodfriend, G. A., Hare, P. E. & Dru¡el, E. R. M. 1992 Aspartic acid racemization and protein diagenesis in corals over the last 350 years. Geochim. Cosmochim. Acta 56, 3847^3850.
Goodfriend, G. A., Kashgarian, M. & Harasewych, M. G. 1995 Aspartic acid racemization and the life history of deep-water slit shells. Geochim. Cosmochim. Acta 59, 1125^1129.
Goodfriend, G. A., Brigham-Grette, J. & Miller, G. H. 1996 Enhanced age resolution of the marine quaternary record in the arctic using aspartic acid racemization dating of bivalve shells. Quaternary Res. 45, 176^187.
Harada, N., Handa, N., Ito, M., Oba, T. & Matsumoto, E. 1996 Chronology of marine sediments by the racemization reaction of aspartic acid in planktonic foraminifera. Org. Geochem. 24, 921^930.
Hare, P. E. 1980 Organic geochemistry of bone and its relation to the survival of bone in the natural environment. In Fossils in the making (ed. A. K. Behrensmeyer & A. Hill), pp. 208^219. University of Chicago Press.
Hensel, R., Jakob, I., Scheer, H. & Lottspeich, F. 1992 Proteins from hyperthermophilic Archaeaöstability towards covalent modi¢cation of the peptide-chain. Biochem. Soc. Symp. 58, 127^133.
Hirst, K. L., Ibaraki-O’Connor, I., Young, M. F. & Dixon, M. J. 1997 Cloning and expression analysis of the bovine dentin matrix acidic phosphoprotein gene. J. Dental Res. 76, 754^760.
Johnson, B. J. & Miller, G. H. 1997 Archaeological applications of amino acid racemization. Archaeometry 39, 265^287.
Julg, A., Lafont, R. & Perinett, G. 1987 Mechanisms of collagen racemizationin fossil bones: application to absolute dating. Quaternary Sci. Rev. 6, 25^28.
Kimber, R. W. L. & Hare, P. E. 1992 Wide range of racemization of amino-acids in peptides from human fossil bone and its implications for amino acid racemization dating. Geochim. Cosmochim. Acta 56, 739^743.
King, K. & Bada, J. L. 1979 E¡ect of in situ leaching on amino acid racemisation rates in fossil bone. Nature 281, 135^137.
Kossiako¡, A. A. 1988 Tertiary structure is a principal determinant to protein deamidation. Science 240, 191^194.
Krings, M., Stone, A., Schmitz, R. W., Krainitzki, H., Stoneking, M. & Pa« a« bo, S. 1997 Neanderthal DNA sequences and the origin of modern humans. Cell 90, 19^30.
Linde, A. 1989 Dentin matrix proteins: composition and possible functions in calci¢cation. Anat.Rec.224, 154^166.
Lura, R. & Schirch, V. 1988 Role of peptide conformation in the rate and mechanism of deamidation of asparaginyl residues. Biochemistry 27, 7671^7677.
McKerrow, J. H. 1973 Non-enzymatic deamidation of asparingyl residues in model peptides and proteins. PhD thesis, University of California, San Diego.
Mansouri, M. El, Fouikar, A. El & Saint-Martin, B. 1996
Correlation between 14-C ages and aspartic acid racemization at the Upper Palaeolithic site of the Abri Pataud (Dordogne, France). J. Arch. Sci. 23, 803^809.
Maroudas, A., Bayliss, M. T., Uchitel-Kaushansky, N., Schneiderman, R. & Gilav, E. 1998 Aggrecan turnover in human articular cartilage: use of aspartic acid racemization as a marker of molecular age. Arch.Biochem.Biophys. 350, 61^71.
Masters, P. M. 1982 Amino acid racemization in structural proteins. In Biological markers of ageing (ed. M. E. Ra¡ & M. D. Schneider), pp. 120^137. Washington, DC: NIH
Publications.
Masters, P. M. 1985 In vivo decomposition of phosphoserine and serine in noncollagenous protein from human dentine.
Calci¢edTissue Int. 37, 236^241.
Matsu’ura, S. & Ueta, N. 1980 Fraction dependent variation of aspartic acid racemization age in fossil bone. Nature 286, 883^884.
Phil.Trans. R. Soc. Lond. B (1999)
Meeson, B. W., Corprew, F. E., McManus, J. M. P., Myers, D. M., Closs, J. W., Sun, K.-J., Sunday, D. J. & Sellers, P. J. 1995 ISLSCP Initiative Iöglobal data sets for land ^ atmosphere models 1987^1988. Volumes 1^5. Published on CD-ROM by NASA (USA___NASA___GDAAC___ISLSCP___001-
USA___NASA___GDAAC___ISLSCP___005).
Metwalli, A. A. M. & van Boekel, M. A. J. S. 1998 On the kinetics of heat-induced deamidation and breakdown of caseinate. Food Chem. 61, 53^61.
Miller, G. H., Magee, J. W. & Jull, A. J. T. 1997 Low-latitude glacial cooling in the Southern Hemisphere from amino-acid racemization in Emu eggshells. Nature 385, 241^244.
Mu« ller, H. T. & Heidemann, E. 1993 Untersuchung der
Gesetzma« ssigkeitein fu« r den Sa« ureabbau von Hautkollagen und Indenti¢zierung der Kollagenspaltstellen biem sauren Gelatineprozess. Das Leder 44, 69^79.
Ne¨ methy, S. 1995 Molecular palaeotological studies of shelled marine organisms and mammal bones. PhD thesis, Go« teborg University, Go« teborg.
Ohtani, S. & Yamamoto, K. 1991 Age estimation using the racemization of amino acid in human dentin. J. Forensic Sci 36, 792^800.
Ohtani, S. & Yamamoto, K. 1992 Estimation of age from a tooth by means of racemization of an amino-acid, especially aspartic-acidöcomparison of enamel and dentin. J. Forensic Sci. 37, 1061^1067.
Ohtani, S., Matsushima, Y., Ohhira, H. & Watanabe, A. 1995 Age-related changes in D-aspartic acid of rat teeth. Growth, Development and Aging 59, 55^61.
Oliyai, C. & Borchardt, R. T. 1994 Chemical pathways of peptide degradation. 6. E¡ect of the primary sequence on the pathways of degradation of aspartyl residues in model hexapeptides. Pharmaceut. Res. 11, 751^758.
Patel, K. & Borchardt, R. T. 1990a Chemical pathways of peptide degradation. 3. E¡ect of primary sequence on the pathways of deamidation of asparaginyl residues in hexapeptides. Pharmaceut. Res. 7, 787^793.
Patel, K. & Borchardt, R. T. 1990b Chemical pathways of peptide degradation. 2. Kinetics of deamidation of an asparaginyl residue in a model hexapeptide. Pharmaceut.Res. 7, 703^711.
Poinar, H. N., Ho« ss, M., Bada, J. L. & Pa« a« bo, S. 1996 Amino acid racemization and the preservation of ancient DNA. Science 272, 864^866.
Pollard, A. M. & Heron, C. 1996 Amino acid stereochemistry and the ¢rst Americans. In Archaeological chemistry (ed. A. M. Pollard & C. Heron), pp. 271^301. London: RSC Paperbacks.
Radkiewicz, J. L., Zipse, H., Clarke, S. & Houk, K. N. 1996 Accelerated racemization of aspartic acid and asparagine residues via succinimide intermediates: an ab initio theoretical exploration of mechanism. J. Am. Chem. Soc. 118, 9148^9155.
Ritchie, H. H. & Wang, L. H. 1996 Sequence determination of an extremely acidic rat dentin phosphoprotein. J. Biol. Chem. 271, 21695^21698.
Ritz, S. & Kaatsch, H.-J. 1996 Methoden der altersbestimmung an lebendon personon: Mobgliichkeiten, Grenzen, Zuliissigkeit & Ethische.Vertretbarkeit. Rerhtsmed. 6, 171^176.
Ritz, S. & Schutz, H. W. 1993 Aspartic acid racemization in intervertebral disks as an aid to postmortem estimation of age at death. J. Forensic Sci. 38, 633^640.
Ritz, S., Schutz, H. W. & Peper, C. 1993 Postmortem estimation of age at death based on aspartic-acid racemization in dentinöits applicability for root dentin. Int. J. Legal Med. 105, 289^293.
Saleh, N., Deutsch, D. & Gil-Av, E. 1993 Racemization of aspartic acid in the extracellular matrix proteins of primary and secondary dentin. Calci¢edTissue Int. 53, 103^110.
Sellers, P. J. (and 14 others) 1995 An overview of the ISLSCP Initiative Iöglobal data sets. On ISLSCP Initiative I global data sets for land ^atmosphere models 1987^1988. Volumes 1^ 5. Published on CD-ROM by NASA. Volume 1:
USA___NASA___GDAAC___ISLSCP___001. OVERVIEW.DOC.
Sendero¡, R. I., Wootton, S. C., Boctor, A. M., Chen, T. M., Giordani, A. B., Julian, T. N. & Radebaugh, G. W. 1994 Aqueous stability of human epidermal growth factor 1^48. Pharmaceut. Res. 11, 1712^1720.
Shimoyama, A. & Harada, K. 1984 An age determination of an ancient burial mound man by apparent racemization reaction of aspartic acid in tooth dentine. Chem. Lett. 10, 1661^1664.
Sinex, F. M. 1960 Ageing and the lability of irreplaceable molecules. II. The amide groups of collagen. J. Gerontol. 15, 15^18.
Skelton, R. R. 1983 Amino acid racemization dating: a test of its reliability for North American archaeology. PhD thesis, University of California, Davis, CA.
Smith, G. G. & Evans, R. C. 1980 The e¡ect of structure and conditions on the rate of racemization of free and bound amino acids. In Biogeochemistry of amino acids (ed. P. E. Hare, T. C Hoering & K. King Jr), pp. 257^282. New York:
Wiley.
Steinberg, S. M., Masters, P. M. & Bada, J. L. 1984 The racemization of free and peptide bound serine and aspartic acid at 1008 C as a function of pHöimplications for in vivo racemization. Bioorganic Chem. 12, 349^355.
Stetler-Stevenson, W. G. & Veis, A. 1986 Type I collagen shows a speci¢c binding a¤nity for bovine dentin phosphoryn.
Calci¢edTissue Int. 38, 135^141.
Stevenson, C. L., Friedman, A. R., Kubiak, T. M., Donlan, M. E. & Borchardt, R. T. 1993 E¡ect of secondary structure on the rate of deamidation of several growth hormone releasing factor analogues. Int. J. Peptide Prot. Res. 42, 497^503.
Sykes, G. A., Collins, M. J. & Walton, D. I. 1995 The signi¢cance of a geochemically isolated (intra-crystalline) organic fraction within biominerals. Org. Geochem. 23, 1059^1066.
Tomizawa, H., Yamada, H., Wada, K. & Imoto, T. 1995 Stabilization of lysozyme against irreversible inactivation by suppression of chemical reactions. J. Biochem. 117, 635^640.
Tri¤t, J. T. 1980 The organic matrix of bone tissue. In Fundamental and clinical bone physiology (ed. M. R. Urist), pp. 45^82. Lippincott, PA: Lippincott.
Tyler-Cross, R. & Schirch, V. 1991 E¡ects of amino acid sequence, bu¡ers, and ionic strength on the rate and mechanism of deamidation of asparagine residues in small peptides. J. Biol. Chem. 266, 22 549^22 556.
Van Duin, A. C. T. & Collins, M. J. 1998 The e¡ects of conformational constraints on aspartic acid racemization. Org.
Geochem. (In the press.)
Von Endt, D. W. 1980 Protein hydrolysis and amino acid racemization in sized bone. In Biogeochemistry of amino acids (ed. P. E. Hare, T. C. Hoering & K. King Jr), pp. 297^304. New York:
Wiley.
Weiner, S., Kustanovich, Z., Gil-Av, E. & Traub, W. 1980 Dead Sea scrolls parchments: unfolding of the collagen molecules and racemization of aspartic acid. Nature 287, 820^823.
Wonnacott, D. M. 1979 The application of gas chromatography to the resolution of enantiomeric mixtures of amino acids. Factors a¡ecting the racemization of amino acids. PhD thesis, Utah State University, Logan, Utah, as cited in Smith G. G. & Evans R. C. 1980 The e¡ect of structure and conditions on the rate of racemization of free and bound amino acids. In Biogeochemistry of amino acids (ed. P. E. Hare, T. C. Hoering & K. King Jr), pp. 257^282. New York: Wiley.
Zhang, J., Lee, T. C. & Ho, C. T. 1993 Comparative study on kinetics of nonenzymatic deamidation of soy protein and egg white lysozyme. J. Agric. Food Chem. 41, 2286^2290.
Zhao, M. X., Bada, J. L. & Ahern, T. J. 1989 Racemization rates of asparagine and aspartic-acid residues in lysozyme at 1008 C as a function of pH. Bioorganic Chem. 17, 36^40.
Discussion
G. Eglinton (University of Bristol, UK ). What precisely do you see is the relationship between conformation and rate of reaction? Also, what controls the conformation? The greater stability you site in the case of dentine is a case in point.
M. J. Collins. In the case of proteins, con¢rmation is critical to function and this over the working temperature of the protein, conformation is governed by interplay of the amino acid residues with each other and the solvent; as the protein is heated and/or degrades residues with each other and the solvent; as the protein is heated and/or degrades conformational freedom is anticipated to increase. Conformation in£uences the rate of reaction by hindering the formation of the reactive intermediates, the impact on rate being an interplay between the extent of the restriction and the structural requirements of the intermediates. Thus, in a highly stretched chain deamidation via the formation of a cyclic Asu requires chain shortening and will be hampered, but this conformation will have little in£uence on deamidation by direct hydrolysis. In the case of aspartic and racemization of the dentine proteins, the observed rate is much slower than our model predicts because many of the aspartyl and asparingly residues are within the triple helix of collagen and will not contribute to the pool of D-isomers which are accumulating in other (degrading) dentine proteins.