SCIENTIFIC CORRESPONDENCE

gas in the sediments and rocks beneath the Conclusions from Nyos disaster . If there are zones of gas-saturated SIR-In August 1986, the Lake Nyos dis­ There was general concern that any ( or near saturated) water beneath either aster brought death and devastation to a attempt to remove gas from the lake might of the lakes, then any reduction in pres­ densely populated part of the Northwest itself create an unstable situation and sure might trigger another lethal gas 1 Province of '. The death toll. induce a massive gas release. Before any release''. It was therefore thought prudent which was officially estimated at more attempt at controlled degassing, to recommend that the rate of water than 1,700, was far in excess of that for any therefore, equipment should be installed extraction should not exceed the natural previous similar disaster. In the immedi­ to monitor the stability of the lake, and recharge rate, as lowering of the water ate aftermath many countries not only appropriate plans should be prepared to level would reduce the pressure at depth. offered humanitarian assistance but also deal with any medical emergency which S.J. FREETH* sent teams of scientists to investigate the might arise. It is also recommended that Geological Hazards Research Unit, cause of the disaster. All participants were any system of gas extraction should first be University College of Swansea, invited to attend a conference organized tested on . 100 km southeast Swansea SA12 8PP, UK by the Cameroonian Ministry of Higher of Nyos. This lake was the source of a gas G.W. KLING Education and Scientific Research in release which killed 37 people in 1984 ( ref. Woods Hole, USA Yaounde in March 1987 to "find out as 5), but as it is only half as deep as Lake M. KUSAKABE much as possible about the cause of the Nyos it is potentially far less dangerous. Okayama University " and ''determine the There was general agreement that pipes J. MALEY precautions which should be taken to should he installed to remove the gas-rich ORSTOM, Montpellier avoid a similar tragedy in the future". bottom water from the lakes and that the F.M. TCHOUA Although the scientific discussions in degassed water should be discharged out­ University of Yaounde Yaounde were interesting, it soon became side the lake basins, so as to avoid disturb­ K. TIETZE apparent that there were irreconcilable ing the natural stratification" of the lakes. BGR, Hannover differences between various groups con­ Concern was expressed over the total lack •Convenor and Chairman of the Inaugural Meeting of the cerning almost every aspect of the disas­ of information concerning the amount of International Working Group on Crater Lakes. ter. The final communique, which was widely circulated and subsequently pub­ specialized small G protein in Schwann lished', accurately reflected the scientific Mitogen signal cells and melanocytes, inactivate this G arguments but offered little sound advice Sm-The recent reports by Xu et al.' and protein. and inhibit the stimulation of to the government of Cameroon other Buchberg et al.' that the proteins encoded cyclase. Thus, even partial loss-of-func­ than to suggest that more scientific studies in the human and mouse neurofibroma­ tion mutations in the NF-1 gene would were needed. During the meeting, a group tosis type 1 (NF-1) genes are structurally result in elevation of intracellular cAMP, of us set up an international working related to the mammalian GTPase­ upregulated expression of polypeptide group to act as an information exchange activating (GAP) and yeast IRA 1/2 pro­ growth-factor receptors, and increased and coordinate future research'. teins immediately suggest a biochemical Schwann-cell and melanocyte proliferation. A seminar on the Lake Nyos gas disas­ explanation for the aberrant proliferation The problem with this hypothesis is that ter was convened on 11 September by our of melanocytes and Schwann cells that is the only organism in which an IRA-like group to which all those who had worked one of the hallmarks of NF-1. This explan­ protein is known to regulate the cyclase­ on the disaster were invited. During the ation. not advanced in either report. stimulating activity of a small G protein is discussions it became apparent that a con­ derives directly from the demonstrated S. cerevisiae. That the NF-1 protein might sensus had now arisen concerning the ability of cyclic AMP to trigger Schwann be a negative regulator of cAMP meta­ cause of the disaster and the urgent need cell and melanocyte division. bolism in certain mammalian cells is none­ for remedial measures to prevent a The following features of signal trans­ theless consistent with the autosomal recurrence. Our main objective is to offer duction and Schwann cell and melanocyte dominant genetics of NF-1 transmission, helpful scientific advice to the political biology apply: (1) GAP (in mammalian the known interaction of IRAl/2, Ras and authorities in Cameroon, but we hope that cells) and IRAl and IRA2 (in Saccharo­ cyclase proteins in yeast, the fact that the our four main conclusions. which are set mycetes cerevisiae) bind to small G NF-1 protein is more similar in amino-acid out below, will also be of wider interest. proteins (such as Ras and Ras-related sequence to IRAl and IRA2 than to GAP 1. The 1986 disaster was caused by a proteins). stimulate the low intrinsic itself. and the demonstration that elevation massive release of CO2 from Lake Nyos. GTPase activity of these proteins. and of cAMP strongly stimulates Schwann cell 2. Lake Nyos now contains about 300 mil­ thereby inactivate them'. (2) A critical and melanocyte proliferation. lion cubic metres of and target of both Ras in yeast and the G,0 This hypothesis also makes two predic­ therefore remains very dangerous. proteins of higher eukaryotes is adenylate tions that are relatively easy to test: both 3. Further CO, is being added at a rate of cyclase. When complexed with GTP. intracellular cAMP and the expression of at least 5 million m' per year and the these G proteins activate cyclase, and cAMP-dependent genes' should be danger is therefore increasing. thereby stimulate the synthesis of intra­ elevated in NF-1 Schwann cells and 4. Another gas disaster could occur at any cellular cAMP'. (3) Cyclic AMP is a melanocytes; and expression of the NF-1 time, so the CO, in Lake Nyos should be strong but cell-specific mitogen for phenotype by these cells should be reduced as a matter of urgency. Schwann cells. melanocytes and certain growth-factor-dependent. The members of the working group con­ other mammalian cells'''. A key com­ GREG LEMKE sidered ways in which carbon dioxide ponent of cAMP mitogenesis in these cells The Salk Institute, PO Box 85800, might be extracted from Lake Nyos. is the cAMP-dependent induction of San Diego, California 92186-5800, USA receptors for polypeptide growth factors 1. Freeth. S. J. & Kay, R. L. F. Nature 328. 299-300 (1987). such as platelet-derived growth factor'. 1. Xu. G. eta/. Ce//62. 599 (1990). 2. Kling, G. W. eta/. Science236, 169-175 (1987). These facts suggest that. just as IRAl and 2. Buchberg, A.M. eta/. Nature 347,291 (1990). 3. Sigvaldason, G. E. J. volcan. Geotherm. Res. 39, 97 3. Hall. A. Science 249,635 (1990). (1989). IRA2 are negative regulators of the G­ 4. Toda, T. eta/. Cel/40, 27 (1985). 4. Kusakabe, M. IWGCLNewsletter1, 1 (1989). protein -cAMP pathway in yeast, so NF-1 5. Raff, M.C. eta/. Ce//15, 813 (1978). 5. Sigurdsson, H. et al. J. volcan. Geotherm. Res. 31, 1 6 Dumont. J.E. et al. Trends biochem. Sci. 14, 67 (1989). (1987). might, in normal circumstances, stimulate 7. Weinmaster, G. & Lemke, G. EMB0J. 9. 915 (1990). 6. Tietze, K. BGRReport 1004701-34 (1987). the low intrinsic GTPase activity of a 8. Monuki, E.S. eta/. Science 249, 1300 (1990). NATURE · VOL 348 · 15 NOVEMBER 1990 201