Inoculation of soil or seeds with plant growth promoting bacteria ameliorates metal toxicity to plants by changing metal speciation in plant tissues but the exact location of these changes remains unknown. Knowing where the changes occur is a critical first step to establish whether metal speciation changes are driven by microbial metabolism or by plant responses. Since bacteria concentrate in the rhizosphere, we hypothesised steep changes in metal speciation across the rhizosphere. We tested this by comparing speciation of zinc (Zn) in roots of Brassica juncea plants grown in soil contaminated with 600 mg kg−1 of Zn with that of bulk and rhizospheric soil using synchrotron X-ray absorption spectroscopy (XAS). Seeds were either uninoculated or inoculated with Rhizobium leguminosarum bv. trifolii and Zn was supplied in the form of sulfide (ZnS nanoparticles) and sulfate (ZnSO4). Consistent with previous studies, Zn toxicity, as assessed by plant growth parameters, was alleviated in B. juncea inoculated with Rhizobium leguminosarum. XAS results showed that in both ZnS and ZnSO4 treatments, the most significant changes in speciation occurred between the rhizosphere and the root, and involved an increase in the proportion of organic acids and thiol complexes. In ZnS treatments, Zn phytate and Zn citrate were the dominant organic acid complexes, whilst Zn histidine also appeared in roots exposed to ZnSO4. Inoculation with bacteria was associated with the appearance of Zn cysteine and Zn formate in roots, suggesting that these two forms are driven by bacterial metabolism. In contrast, Zn complexation with phytate, citrate and histidine is attributed to plant responses, perhaps in the form of exudates, some with long range influence into the bulk soil, leading to shallower speciation gradients.