This is usually due to nucleic acid in the supernatant and may be the result of a higher plasmid copy number than you have previously achieved. The addition of Benzonase a broad specificity DNA and RNAse to the lysis buffer should help.

There could be three reasons for this. First there is excess of nucleic acids in your supernatant sample. Try Rnase A (10ug/ml) and Dnase (5ug/ml) after sonication. Keep it for 30mins on ice and then centrifuge. This will reduce the viscosity of the sample due to nucleic acid presence.

The second reason could be too much cell mass and less amount of lysis buffer that lead to viscous supernatant. Try to dilute your cell pellet with excess of lysis buffer.

Third reason could be improper centrifugation. Centrifuge at high speed for 30 mins (atleast) at 4 degree to get clear supernatant.

I hope it will work.

Good Luck! 

Try nuclease treatment and also adjusting your lysis buffer volumes with respect to you cell pellet/cell densities at the time of harvest.

I have no experience with your experimental system.  All contributors are correct in their advice to reduce viscosity due to nucleic acids.  But, since these attempts have failed, I suggest that for some reason your cells have expressed high levels of glycoprotein of some sort.  These molecules are difficult to "shear" so try a glycosylase that is compatible with protease inhibitors.

Please let us learn of your progress Beth!

Thank you everyone for these good suggestions! I suspect it is the nucleic acid problem combined with increased productivity due to the use of Magic Media instead of standard media. I have tried shearing the nucleic acids with a 20g needle and syringe and that helps, although it is time consuming.