The introduction of beta-hydroxy-alpha-amino acids (beta HAAs) into organic molecules has received considerable attention as these molecules have often found widespread applications in bioorganic chemistry, medicinal chemistry and biomaterial science. Despite innovation of asymmetric synthesis of beta HAAs, stereoselective synthesis to control the two chiral centres at C-alpha and C-beta positions is still challenging, with poor atomic economy and multi protection and deprotection steps. These syntheses are often operated under harsh conditions. Therefore, a biotransformation approach using biocatalysts is needed to selectively introduce these two chiral centres into structurally diverse molecules. Yet, there are few ways that enable one-step synthesis of beta HAAs. One is to extend the substrate scope of the existing enzyme inventory. Threonine aldolases have been explored to produce beta HAAs. However, the enzymes have poor controlled installation at C-beta position, often resulting in a mixture of diastereoisomers which are difficult to be separated. In this respect, l-threonine transaldolases (LTTAs) offer an excellent potential as the enzymes often provide controlled stereochemistry at C-alpha and C-beta positions. Another is to mine LTTA homologues and engineer the enzymes using directed evolution with the aim of finding engineered biocatalysts to accept broad substrates with enhanced conversion and stereoselectivity. Here, we review the development of LTTAs that incorporate various aldehyde acceptors to generate structurally diverse beta HAAs and highlight areas for future developments.