Heavy water (D2O) is commonly used as a solvent additive in electrospray ionization-mass spectrometry (ESI-MS) to probe biomolecular interactions under the assumption that isotopic substitution does not significantly affect the ionization process. However, D2O differs measurably from H2O in several physicochemical properties, which can influence ionization across biomolecular classes. Here, we systematically investigate how D2O influences native ESI across biomolecular classes and show that D2O introduces a pronounced, polarity-dependent detection bias which can suppress the signal of proteins and peptides in negative ion mode relative to other classes of biomolecules. Protein signal suppression is mitigated by reducing primary ESI droplet size or adding volatile ammonium salts, implicating altered droplet evolution and salt retention in D2O. Its magnitude is strongly dependent on analyte acidity: basic and moderately acidic proteins are most affected, whereas extremely negatively charged proteins, RNA, and polysaccharides remain largely unaffected. Signal loss also correlates with extensive sodium adduction and peak splitting rather than impaired charging. Exploiting this selectivity, we demonstrate controlled switching between peptide and RNA detection in a coacervate-based cellular delivery system. These findings reveal an unexpected solvent-dependent bias in negative-mode native MS measurements which can be exploited for experimental design.