Dilute tiny amount in DMSO like 500 micro liters and mix. Solution should be bright pink. Then dilute 1:1000 in PBS because DMSO can damage tissue and DiI needs to be delivered in very small amounts. Put on a PBS soaked gauze on wattman filter paper in a 12 well plate for stability. Add about 200 microliter a  little at a time. Wait 5 minutes. Mount and image right away. 

Hi Tatyana, thanks so much for your answer.

I'll try it :)

Hi Ria,

        To label hippocampal neurons with their dendrite trees (including spines), you need to cut 250um slices from fixed brains.  The reason for using thick slices is to prevent pipette tip from penetrating your slices and keep the dye from unpredictable diffusion.  You can make 2% DiI (Invitrogen-Molecular Probes, Grand Island, NY 14072) in N,N-dimethylformamide (Sigma, St. Louis, MO 63103).  Inject DiI into the cell body layer through a glass micropipette (3-5µm in tip diameter), which was attached to a microinjector using pressure.  You need to optimize the volume and speed of DiI injection. DiI-injected slices were kept in dark at 37 °C for 1 w.

        For your reference, you can check a previous paper: Eur J Neurosci. 2007 Apr;25(7):1987-97.

        Hope this is helpful. Good luck!

Hi Guoxiang. Thank you very much for your kindly reply.

I briefly read the paper that you give me. In the method, you use kicroinjector as tool for inject the DiI Unfortunately, our lab doesnt have that.

But, as you mentioned, you are right...the slice should be thick. Because when i tried 100um, the slice was easily broken by needle.

Anyway, Glass micropipet!!! good suggestion....I will try that.

Thanks