The first order self-energy corrections of the kaon in the symmetric nuclear matter are calculated from kaon-nucleon scattering matrix elements using a chiral Lagrangian within the framework of relativistic mean field approximation. It shows that the effective mass and the potential of K + meson are identical with those of K − meson in the nuclear matter, respectively. The effective mass of the kaon in the nuclear matter decreases with the nuclear density increasing, and is not relevant to the kaon-nucleon Sigma term. The kaon-nucleus potential is positive and increases with the nuclear density. Our results indicate that K − meson can not be bound in the nuclei.PACS numbers: 12.39. Fe, 13.75.Jz, 21.65.+f.The in-medium properties of kaons have caused more attentions of nuclear physicists for many years. It was predicted 20 years ago that Bose-Einstein condensation of K − mesons is possible in the nuclear matter with a density up to several times of the normal nuclear matter density, which is called kaon-condensation phenomenon [1]. This prediction had inspired great interests of nuclear physicists in the following years [2,3,4,5]. It's assumed that there exists kaon-condensation in the high density nuclear matter in the core of neutron stars, and hence it can be utilized to soften the equation of state of beta-stable matter in neutron stars [4,5].In 1999, by fitting the K − atomic data, it was predicted that there can be deeply bound kaonic atoms, i.e., kaonic nuclei [6,7]. Although the experimental data at KEK and of FINUDA collaboration were interpreted as the evidences that deeply bound kaonic atoms exist [8,9,10,11], whether there can be deeply bound states of K − meson in the nuclei is still an issue of great controversy [12,13,14,15].Whether or not there exist deeply bound states of K − depends upon properties of K − in the nuclear medium, and is closely relevant to the depth of potential-well of K − in the nuclear matter. Nevertheless, it can be seen from present results that the interaction between the kaon and the nucleus is mainly model-dependent. By fitting the data of K − atoms, a strong attractive K − -nucleus potential with the depth of 150-200MeV is obtained [16]. However, the calculation based on the chiral coupling channel approach gives the K − -nucleus potential ranging from 85-140MeV [17,18,19], and the chiral unitary theory that starts from the bare K − N interaction predicts an even more shallower K − -nucleus potential in the range of 50 -70 MeV [20,21]. The medium properties of kaons have also been studied from mean field theories, built within the framework of chiral Lagrangians [22,23], based on Walecka model extended to incorporate kaons [24,25], or using explicitly quark degrees of freedom [26]. Moreover, the behavior of the effective masses of kaons at the finite nuclear density has been studied in the SU (3) Nambu-Jona-Lasinio model [27].Although these models give different K − -nucleus potential depths, most of them predict negative values of the K − -nucleus potential in the nuclea...