Figure S1. Viruses used in this study and control experiments. (A) Plasmid constructs consisting of the following components were packaged into AAVs: a promoter of interest (black boxes); either Slc39a14, the gene coding for Zip14, or Channelrhodopsin (ChR2) (red and purple boxes); and a histology visible tag (green, blue, and yellow boxes). (B) A control experiment was performed to check whether expression of large transmembrane proteins, like ChR2, was enough to produce MRI signal changes similar to the expression of Zip14. A plasmid construct consisting of the Synapsin promoter, ChR2 gene, and yellow fluorescent protein (YFP) was packaged into AAV9. These viruses were injected into the S1BC. (C) MRI was performed immediately after injection and 2 weeks after injection using the same protocol as used for Zip14 expression imaging. No MRI signal changes were observed at the injection site (yellow arrow) immediately after nor at 2 weeks following injection. Matched immunofluorescence staining for YFP shows strong expression at the injection site (white arrow) and in areas projecting away from the injection site. (D) Quantification of MRI signal at the injection site confirmed no significant differences between the injected S1BC and contralateral control S1BC 2 weeks after injection (n = 5, p = 0.389) (E,F) No heterogeneity in MRI signal intensity was apparent across the cortical depth nor along cortical layer 5. Dotted lines denote $$$ \pm $$$ SD.

Figure S2. Immunohistology checks for AAV–hSyn-Zip14-EGFP injections. (A) Expression of both Zip14 and the histology-visible EGFP tag was confirmed in neurons infected with AAV-hSyn-Zip14-EGFP. (B) High-resolution confocal microscopy further confirmed colocalization of Zip14 and EGFP. Both proteins were localized to the cell membrane and to puncta within the cytoplasm. (C) Expression of Zip14 in the S1BC was confirmed 6 months after injection of AAV-hSyn-Zip14-EGFP. (D) Colocalization of Zip14 and EGFP persisted at least 6 months after injection and confirmed that long-lasting Zip14 expression was caused by the AAV.

Figure S3. Supporting information for MRI protocol development. (A) The MPRAGE protocol was optimized on animals injected with AAV–hSyn-Zip14–EGFP into the S1BC. TIs included 400 ms, 1100 ms, 1500 ms, and 2000 ms. Optimal contrast between normal-appearing gray matter and Zip14 expression-borne signal enhancement was apparent at a TI of 1100 ms. (B) Representative R₁ map generated from variable TI images showed a large magnitude R₁ increase in the Zip14 expressing areas at the injection site. (C) Modeled MRI signal as a function of TI for white matter (corpus callosum), normal-appearing gray matter (cortex, control), and Zip14 expressing areas (cortex, Zip14).

Figure S4. Supporting information for in vivo anterograde tracing from the S1BC with Zip14 expression MRI. (A) Schematic quantification strategy. Dashed lines represent 240 μm-thick line regions of interest (ROIs) drawn for each anatomical area to produce the quantitative data as shown in Figure 2D (cyan lines) and Supporting Information Figure S4B,C (magenta lines). Normal-appearing gray-matter ROI is represented by the black dotted line in the cortex, medial to the Control S1BC. (B) Line intensity analysis from midline through the injected (blue line) and contralateral control (red line) hemispheres displayed focal signal enhancement in the VPM. Some hyperintensity was observed through the caudate putamen (CP), perhaps indicating processes from neurons in the S1BC, although this enhancement was not significant. (C) Line intensity analysis through the injected (blue line) and contralateral control (red line) hemispheres centered on S2 along layer 5 also confirmed hyperintensity in the known projecting site. Dotted lines denote $$$ \pm $$$ SD. (D) High-resolution confocal immunohistology of the S1BC injection site for the histology-visible EGFP tag showed both cell bodies and processes. (E,F) Confocal immunohistology of the anterograde projecting areas VPM and S2. Only the neuronal processes in these areas were EGFP-positive, confirming that the observed MRI signal enhancement was produced by only the neurons projecting from the injection site.

Figure S5. Supporting information for in vivo anterograde tracing from the VPM to S1BC with Zip14 expression MRI. (A) Schematic quantification strategy. Dashed lines represent 240 μm-thick line regions of interest (ROIs) drawn for each anatomical area to produce the quantitative data as shown in Figure 2H (cyan lines) and Supporting Information Figure S5B,C (magenta lines). Normal-appearing gray-matter ROI is represented by the black dotted line in the cortex, medial to the Control S1BC. (B) Line intensity analysis from midline through the injected (blue line) and contralateral control (red line) hemispheres. Hyperintensity was apparent in the injected VPM. The CP ipsilateral to the injected VPM was also hyperintense, likely caused by processes of the neurons in the injected VPM. (C) Line intensity analysis across the cortical depth in the injected (blue line) and contralateral control (red line) S1BC. High MRI signal intensity was apparent in Layer 4 of the injected S1BC. Dotted lines denote $$$ \pm $$$ SD. (D) High-resolution confocal immunohistology of the injected VPM for EGFP showed expression of the AAV–delivered construct in neuronal cell bodies. (E) Confocal immunohistology of the S1BC Layer 4 ipsilateral to the injected VPM revealed EGFP expression in only axons and processes. This finding confirmed that the MRI signal enhancement observed in this area is due to Zip14 expression from the neurons in the VPM projecting to S1BC Layer 4. (F) Manual annotation strategy. (G) An intensity threshold was set to 50% of the whole-brain signal maximum. (H) The combination of manual segmentation and intensity threshold allowed whole-brain segmentation of Zip14 expression-borne signal enhancement. Left: Original image. Right: Segmented image showing the enhanced regions within the manually drawn boundaries. (I) MRI signal heterogeneity within the S1BC was apparent, potentially indicative of individual whisker barrels within the barrel field. The high signal peaks may correspond to the centers of individual whisker barrels, and the minima could correspond to the interbarrel spaces. To test this, line intensity analysis through the S1BC along Layer 4 on MRI and immunohistology for EGFP was performed. To better compare measurements of peak widths and distance between peaks, the immunohistology data were smoothed using Gaussian interpolation. (J) Quantitative comparison of MRI and immunohistology line intensities through S1BC Layer 4. Estimates of barrel widths, defined by the FWHM of each peak, and distance between barrels, defined by interpeak distance, were made. Overall, measurements made on immunohistology trended to be smaller than MRI for both metrics. The cause of this difference is likely partial volume effects of the relatively large MRI voxel compared with the inter-whisker barrel spaces. Acquiring MRI at higher resolution would likely bring the measurement more into agreement with immunohistology. (K) Rendering segmented enhancing brain areas through the MRI volume enabled creation of 3D Zip14 expression maps.

Figure S6. MRI registration workflow. (A) Input image of sagittal slice through the acquired mouse-head MDEFT MRI volume. (B,C) Initial brain mask generation using 3D PCNN. Skull stripped image (D,E) Allen Institute CCFv3 50 μm resolution average autofluorescence target and atlas label volume. (F) ANTs nonlinear registration warped grid to transform the CCFv3 autofluorescence target to the input image space. (G) CCFv3 atlas labels in input image space. (H,I) Coronal slice through the input image volume after skull stripping and CCFv3 atlas labels overlaid.

Figure S7. Immunohistology checks for RetroAAV-hSyn-Zip14-FLAG injections. (A) Immunostaining for Zip14 and the histology visible FLAG tag 2 weeks after injection of the RetroAAV-hSyn-Zip14-FLAG virus showed excellent colocalization. (B) High-resolution confocal immunohistology of the caudate putamen (CP) for Zip14 showed focal expression in neuronal cell bodies and local processes. (C,D) Confocal immunohistology from globus pallidus (GP) and substantia nigra (SN), which house neurons that send their processes to the CP, for Zip14 revealed focal expression in neuronal cell bodies. (E) Line intensity analysis from midline through the CP in the injected hemisphere (red, purple lines) and contralateral control hemisphere (green, blue lines) over time revealed hyperintensity through the injected CP 2 weeks after injection. The contralateral CP did not enhance at any timepoint and was isointense with the injected CP immediately after injection. Shaded region denotes $$$ \pm $$$ SD. (E,F) Line intensity analysis from midline through the areas that send projections to the CP (i.e., the GP and SN) revealed focal signal enhancement in both areas ipsilateral to the injected CP 2 weeks after injection.

Figure S8. Supporting information for AAV–GFAP–Zip14-FLAG. (A) High-resolution immunohistology for Zip14 and GFAP confirmed specific Zip14 expression in GFAP-positive cells. (B) Line intensity analysis centered on the injected (red, purple lines) and contralateral control (green, blue lines) right hemisphere S1 barrel cortex (S1BC) showed hyperintensity in the injected S1BC 2 weeks after injection. The fact that the enhanced MRI signal was more diffuse and lower in peak intensity compared with the AAV-hSyn-Zip14–EGFP injections into the S1BC may be indicative of the relative population sizes of neurons and astrocytes. Shaded region denotes $$$ \pm $$$ SD.