Functionalization of catalytic nanoparticles (NPs) on semiconductor metal oxide (SMO) sensing layer is an indispensable process to obtain improved sensitivity and selectivity for high performance chemical sensors. It is a critical challenge to achieve homogeneous distribution of nanoscale catalysts on SMO in consideration that gas sensing characteristics of SMO-based sensing layer are significantly influenced by the size and distribution of catalysts. Here, we propose a highly effective functionalization method to achieve well-distributed catalytic NPs onto one dimensional (1D) SMO nanofibers (NFs) using protein cage templates: apoferrtin. By simply replacing precursor in the apoferritin assisted method, not only precious catalyst such as Pt but also non-precious catalysts such as La and Cu were successfully synthesized in nanoscale (i.e., 3-5nm). Furthermore, the apoferritin-encapsulated catalysts exhibited high dispersion property due to repulsive force between protein shells. For this reason, catalytic NPs were homogeneously decorated on ZnO NFs after electrospinning followed by calcination. Catalytic Pt NPs and Cu NPs functionalized ZnO NFs exhibited approximately 6.38-fold (R-air/R-gas = 13.07) and 2.95 fold (R-air/R-gas = 6.04) improved acetone response compared with the response (R-air/R-gas = 2.05) of pristine ZnO NFs. In the case of La NPs functionalized ZnO NFs, 9.31-fold improved nitrogen monoxide response (R-air/R-gas =10.06) was achieved compared with the response of pristine ZnO NFs. The four catalyst-ZnO composite NFs successfully distinguished simulated breath components such as acetone, toluene, nitrogen monoxide, carbon monoxide, and ammonia with well-classified patterns by principal component analysis (PCA). This work demonstrated a robustness of synthetic and functionalization method using bio-inspired protein templates combined with electrospinning technique and a promising potential of using non-precious catalysts to establish diverse sensing material libraries that can be applicable to breath pattern recognition for diagnosis of diseases. (C) 2016 Elsevier B.V. All rights reserved.