Molecular hybridization assay, especially involving the use of tissues directly, has been developed as a rapid, simple and important technique for plant pathogen detection and/or gene expression analysis on a large scale. In theory, this method relies on the specific binding of a labeled probe to a target nucleotide sequence. However, occasional false positive reactions can pose a problem in its application and the cause is often not well understood. Here, we show that in tissue-printing hybridization to detect Peach latent mosaic viroid (PLMVd) strong signals could arise by interactions between the viroid probe and plant proteins. Such probe-protein interactions made it difficult to show significant correlations between viroid infection and the level of hybridization signals. These results challenge the traditional view that proteins may hamper PCR reactions but have no influence on molecular hybridization. They further demonstrate that such probe-protein interactions in a plant could compromise the quality of molecular hybridization assays for viroid detection. Our results uncovered an important source of false positive reactions in tissue-printing hybridization and suggest that specificity can be improved by removing proteins.