Factors affecting minipurification

Buffer type
Initially I used 0.1M diammonium citrate, adjusted to pH 6.5 (suggested by W.G. Langenberg). In no case did it extract less virus than neutral phosphate. All citrates (ammonium, lithium, sodium and potassium) were similarly efficient. The sodium salt is cheap, soluble and compatible with sodium dodecyl sulfate. Citrate buffers are surprisingly universal in ability to extract viruses.

Salt concentration
0.2M and especially 0.4 M citrate yield more soluble and cleaner samples than 0.1M. Of roughly 50 viruses, only Beet Soilborne Mosaic was "lost" (insoluble?) in 0.4M citrate. Example of salt concentration series.

Centrifugation speed and time
Brief ultracentrifugation clarifies samples better than low speed centrifugation (it pellets membranes). Ultracentrifugation pellets virus, extended ultracentrifugation will pellet host proteins. Complete virus recovery is unnecessary.

Amount of sample
I use roughly a gram of tissue and 23ml of buffer (enough to cover blendor blades and fill a 1x3.5" polycarbonate centrifuge tube). Only ~1% of this is loaded on a gel. The sample must yield a pellet large enough so it's not washed away during recovery.

Dropping extraction pH to 4.5 usually eliminates rubisco and other host contaminants. Many viruses survive acidification. Dropping the pH to 5.5 has no effect on sample pattern; host contaminants "begin to disappear" at pH 5 and are usually gone at pH 4.5. Examples of pH series.

Adding trypsin or chymotrypsin to the grinding buffer "shaves" the C-terminus of rubisco L but has no effect on rubisco S. Proteases "shave" some viruses (e.g. potyviruses), "nick" others (MCMV) and have no effect on still others (tymoviruses, tobamoviruses). - example

Thiols activate thiol proteases which, in turn, degrade proteins. Iodoacetamide in the grinding buffer inhibits proteolysis. N-ethylmaleimide (NEM) and other alkylating agents work equally well. Example of thiol stimulated proteolysis.

Including DIECA (diethyldithiocarbamate), a chelator, during extraction inhibits browning of extracts. The purported mechanism is chelating Zn from polyphenol oxidases. Fluoride and cyanide also work (avoid, because they're poisonous), but EDTA does not. I suspect DIECA chelates Fe in a form which cannot produce free radicals.

Oxidizing agents
Adding hydrogen peroxide (small amounts) to the extraction buffer has little effect. In fact samples with hydrogen peroxide look better than those w/o by virtue of peroxide inhibiting thiol proteases!

Heating extracts (before first centrifugation) eliminates (precipitates) host components. Most viruses survive 20min at 45C. Alfalfa mosaic and tobacco streak are the most labile viruses I've tested. Many viruses survive 60C or 70C. Higher temperatures precipitate rubisco. In no case has heating stimulated proteolysis! A convenient heat treatment is to immerse tubes in hot (58C) tap water for ten minutes (before first centrifugation). Extracts within the tubes reach about 45C, which can potentially precipitate only the most labile viruses. Example: heat treatment of PVX

Usually unnecessary, but where sample foams severely and heats, a drop of Antifoam A concentrate (silicone based antifoam) usually helps. PVP increases foaming. Antifoam increases friability of ultracentrifuge pellets.

Polyvinylpyrrolidone (soluble PVP)
Tannins (e.g in tree leaves, rose leaves) precipitate proteinaceous components, including viruses. PVP adsorbs tannins allowing recovery of proteins and viruses. PVP 10K, the smallest commercial PVP, is ideal. Insoluble PVP is unsatisfactory (presumably not enough surface area). See minipurifications from tannin rich tissues.

Polyethyleneglycol (PEG)
PEG at low levels increases solution viscosity decreasing protein recovery nonselectively. Increasing PEG level to the point where it would precipitate virus would substantially increase centrifugation times.

Detergents disrupt membranes in the clarification (brief ultracentrifugation) supernatant. Nonionic (Triton X-100) and zwitterionic (e.g. lauryl sulfobetaine) detergents are mild and work well. Tannins precipitate Triton (bind to PEG). Cationic detergents precipitate in proportion to the amount of added detergent (an inconvenience). SDS, the "strongest" known detergent, denatures proteins precipitating them at low concentration and solubilizing them at high concentration. Bringing the initial supernatant to 0.5% in SDS improves purity of stable viruses (e.g. tymoviruses, tobamoviruses). Sarkosyl, a weaker anionic detergent, is comparable to Triton X-100. Any "mild" and soluble detergent should be compatible.