key: cord-0033772-ku90rkcu
authors: SILVI, BERNARD; D'ARCO, PHILIPPE; CAUSÀ, MAURO
title: Ionicity in silica
date: 1991
journal: Nature
DOI: 10.1038/353394c0
sha: f6ec585b7a241964e90dc2f4cf07778c991b723f
doc_id: 33772
cord_uid: ku90rkcu

nan

SIR -Hochberg and Waage suggested in their News and Views article 1 that some new genetically modified insect viruses will be acceptable as biological control agents because they have "highly restricted host ranges". There is widespread agreement that specific biological control agents are much to be preferred, on environmental grounds, over chemical pesticides. But the dangers of nonspecific biocontrols are great, and much damage has resulted from their use 2 . How specific are these genetically modified organisms? They are derived from the Autographa californica nuclear polyhedrosis virus (AcNPV), a baculovirus, which has a wide and sporadic host range in the lepidoptera. There are around 2,500 species of lepidoptera in Britain 3 , and of course many times more elsewhere. The records of host range of this virus4--6, based on a small fraction of the known species, show that, of twe lve superfamilies tested . eight apparently contain 'permissive' species (species killed by fewer virus polyhedra than are produced by one dead caterpi llar). With this and other baculoviruses one species in a genus may be permissive, others resistant and the LD50s vary markedly between different permissive species, without apparent regard for taxonomy.

The superfamilies known to have permissive species are the Gelichiodea, Pyraloidea, Papilionoidea, Sphingoidea and Noctuoidea: the Bombycoidea, Geometroidea and Yponometroidea may have them , but this needs confirmation by DNA analysis. It seems that 5 -10 per cent of British lepidoptera are permissive for AcNPV, a non-native virus, putting 125-250 species at risk, including some of great conservation value.

The two new genetically modified organisms 8 · 9 , like others derived from the same virus 10 , may have host ranges slightly different from that of the wild type. But unless they can be further engineered to be absolutely specific for a known set of (pest) species, it is difficult to see that they could be used safely in an uncontrolled way in the field. Would not the risk assessment required under EC Directive 90/220, for instance, indicate that they are undesirable?

Hochberg and Waage 1 also note that "disabling engineered viruses so that they do not persist has some appeal". Removing the polyhedrin gene to produce a non-occluded virus 5 · 10 reduces both persistence and, to a small extent, the host range; neither of the two new genetically modified organisms has apparently been modified in this way.

Research into the molecular and other bases of host specificity is likely to be a sine qua non for the successful, muchdesired , replacement of chemical control agents for insects by viral ones.

SIR -Kramer et al. 1 suggest that a change in ionicity is responsible for the transition from the a: to the ~ phase in silica. Their ab initio force-field method indicates that an increase by about 0.1 atomic units of the net charge on silicon stabilizes the ~ structure of quartz and cristobalite with respect to the a: structures, implying that the former correspond to global minima of the Born-Oppenheimer potential-energy surface. Our recent Hartree-Fock calculations 2 on the quartz structures of SiO2 and GeO2 come to opposite conclusions with respect to both the magnitude and the direction of the ionicity effect. Our calculations were performed at a level of accuracy similar to that of the cluster calculations used to derive the force field of ref. The Born-Oppenheimer energy curve E( o) suggests that the high-temperature 13 phase can be viewed not as a minimum of the potential-energy surface but rather in terms of the expectation values of the atomic coordinates with respect to the nuclear wavefunction. Molecular dynamics calculations 4 provide support for this interpretation. Starting from the a: 1 structure, the a:2 twin appears as the temperature is raised, and at 850 K the system alternates dynamically between Calcium binding to fibrillin?

Dietz et al. 1 described an Arg 239-Pro mutation in one of the 34 epidermal growth factor ~EGF)-like repeats of the filbrillin gene found in two patients with Marfan syndrome. These repeats contain a consensus sequence (see diagram) analogous to that found in the first EGF-like domain of coagulation factor IX which is known to bind calcium 3 . It therefore seems likely that fibrillin will also bind calcium. Moreover, the two-dimensional NMR structure of the EGF-like domain 4 of factor IX suggests that Arg 239 in the analogous fibrillin EGF-like domain is located on the two phases. This corresponds to the appearance of a symmetric double-well potential in the Born-Oppenheimer energy surface (see figure) . The softmode normal coordinate can be approximated by the tilt angle o. For the symmetric double well, the expectation value <0>=O, corresponding to the 13 phase, whereas at low temperature the symmetry is broken and <0> takes negative and positive values for a:1 and a:2, respectively. When the nuclei are considered as quantum particles, the a: -13 transition is displacive, in agreement with experiments 5 , whereas a classical description corresponds to the dynamical picture of the molecular dynamics calculation 4 • As pointed out by Tsuneyuki et al. 4, the thermal energy required to symmetrize the potential increases with the height of the barrier separating the twinned structures. We calculate a higher barrier for GeO2 than for SiO2 (see figure) , which explains why no 13 structure is observed for the germanium analogues of silica polymorphs. the second strand of an antiparallel 13sheet adjacent to the inferred calcium binding site. We propose that local disruption of this calcium-binding site could be responsible for the defect in fibrillin function , although long-range effects of the proline substitution on protein structure cannot be excluded. Supporting the concept of local disruption is the identification of two mutations in the factor IX EGF-like domain, Asp 47-Glu and Asp 64-Asn, present in patients with haemophilia B . These mutations reduce calcium binding appreciably but do not grossly affect the protein's structure 3 . Several groups have proposed that EGF-like domains are involved in protein-protein interactions. We would like to draw attention to a recent Bell's inequality SIR -Maddox's article 1 on non-locality in quantum mechanics reminded me that Bell's inequality has a long history which may be of some interest. A few years ago, I attempted to trace the origins of the probability metric. (The distance between two events is the probability that one of them occurs but not both. Strictly speaking, this is a pseudo-metric from which a metric can be constructed by standard techniques.) My colleague, D. A. Edwards, pointed out that the treatise by Dunford and Schwartz 2 contains both a discussion of a more general metric on measure spaces and a survey of its history. There it is traced back to work of Aronszajn and Nikodym 3 in the late 1920s. Moreover, that metric is actually just a special case of the L 1 metric on integrable functions discussed 10 years earlier by Frechet 4 • (One restricts the L 1 metric to indicator functions of events.)

The fact that the basic inequality was known for so long makes it all the more surprising that, until Bell's independent rediscovery of it in the 1960s, no one seems to have observed that the triangle inequality fails in quantum mechanics and can therefore provide a test of large classes of hidden variable theories. The new paper by Fivel 5 , which Maddox discusses, provides an interesting new insight by comparing the probability metric with the quantum-mechanical Hilbert space metric and thereby isolating the mechanism which gives rise to Bell's inequality in one case but not in the other.

Balliol College, Oxford OX1 3BJ, UK

The Release of Geneticallyengineered Micro-organisms

Final Report on a Feasit:>ility Study on the Estat:>lishment of European Biotechnology Environmental Release Datat:>ase (BAP. CEC DGXII

Linear Operators I General Theory