We investigate the environmental effect on the metal enrichment of star-forming galaxies (SFGs) in the farthest spectroscopically confirmed and X-ray-detected cluster, CL J1449+0856 at z = 1.99. We combined Hubble Space Telescope/WFC3 G141 slitless spectroscopic data, our thirteen-band photometry, and a recent Subaru/Multi-object InfraRed Camera and Spectrograph (MOIRCS) near-infrared spectroscopic follow-up to constrain the physical properties of SFGs in CL J1449+0856 and in a mass-matched field sample. After a conservative removal of active galactic nuclei, stacking individual MOIRCS spectra of 6 (31) sources in the cluster (field) in the mass range 10 ≤ log(M/Mo.) ≤ 11, we find a ∼4σlower [N ii]/Hα ratio in the cluster than in the field. Stacking a subsample of 16 field galaxies with Hβ and [O iii] in the observed range, we measure an [O iii]/Hβ ratio fully compatible with the cluster value. Converting these ratios into metallicities, we find that the cluster SFGs are up to 0.25 dex poorer in metals than their field counterparts, depending on the adopted calibration. The low metallicity in cluster sources is confirmed using alternative indicators. Furthermore, we observe a significantly higher Hα luminosity and equivalent width in the average cluster spectrum than in the field. This is likely due to the enhanced specific star formation rate; even if lower dust reddening and/or an uncertain environmental dependence onthe continuum-to-nebular emission differential reddening may play a role. Our findings might be explained by the accretion of pristine gas around galaxies at z = 2 and from cluster-scale reservoirs, possibly connected with a phase of rapid halo mass assembly at z > 2 and of a high galaxy merging rate.