California leafy green lines cool salad with a vacuum. The crop goes into a chamber, the pressure drops, water evaporates off the leaves and takes the field heat with it. It takes minutes and it is good for the cold chain. In 2008 a team at UC Davis measured what the same operation does to E. coli O157:H7 sitting on those leaves. The count inside the tissue went from 198,000 to 2.65 million CFU per gram, and neither a stringent surface sterilization nor four rounds of washing could get those cells out.
Internalization means the bacterium ends up inside the leaf tissue rather than on its surface. The word is clumsy, the idea is plain: tap water cannot reach those cells, and neither can chlorine in a processing flume, because plant tissue stands between the cell and the solution.
The FDA says so in writing
The produce safety guidance puts it directly: antimicrobials do not significantly minimize pathogens that have attached to produce or that have been internalized in the tissue, so a wash water process control is not a kill step. What the tank actually does, and why the chemistry is in there at all, is covered in our guide to wash water sanitizers. This piece is about the second half of that sentence.
Door one: stomata, and they work in daylight
The neatest experiment on this comes from Sela’s group at the Volcani Center in Israel, published in 2009. They incubated iceberg lettuce leaves with tagged Salmonella in the light and in the dark. In the light the bacteria gathered around open stomata and moved into the inner leaf tissue. In the dark they attached in a scattered pattern and hardly got in at all.
Then came the control that makes the paper worth reading. They forced the stomata open in the dark with fusicoccin, and it changed nothing. So the open door is not the point. The bacterium is drawn to compounds the leaf makes while photosynthesizing, and it travels there under its own power: mutants with broken motility and chemotaxis internalized far less.
Where those cells then sit on the leaf was mapped with confocal microscopy back in the late 1990s, and that work is covered in our piece on biofilms.
Not every leaf is equally hospitable
The same group compared seven crops two years later, counting the share of microscope fields with at least one tagged Salmonella beneath the epidermis.
| Crop | Share of fields with internalized Salmonella |
|---|---|
| Arugula | 88% |
| Iceberg lettuce | 81% |
| Basil | 46% |
| Red lettuce | 20% |
| Romaine lettuce | 16% |
| Parsley | 1.9% |
| Tomato leaves | 0.56% |
Arugula and parsley share a shelf in the same fridge, and they differ by more than fortyfold. One more detail from the same paper: over a two-year survey, iceberg ranged from 0 to 100%, and the high end fell mainly in summer. Why summer is its own season for outbreaks has an article of its own.
Door two: roots and irrigation
In 2002 a Rutgers team grew lettuce in manure-amended soil and on contaminated irrigation water. E. coli O157:H7 entered the plant through the root system and traveled through the edible portion. The authors put the consequence plainly: once inside, the cell is protected from sanitizing agents by virtue of being out of reach.
That is soil and field. Water culture behaves differently: three studies spanning eighteen years found soil more conducive to internalization than hydroponics, and they are covered in our piece on countertop hydroponic gardens.
Overhead watering sits next to the root question. In 2014 Erickson and colleagues sprayed plants two ways: once with a high dose, and repeatedly with a low dose to imitate irrigation. The single high-dose spray internalized more in spinach than in lettuce. The repeated low-dose spray did not get into lettuce at all, while it did get into spinach and parsley, and the authors put that down to the film of moisture that builds up over repeated spraying and carries cells to the stomata.
The lesson there belongs to the grower rather than the cook: what matters is the water quality and how long the leaf stays wet. The same holds under glass.
Door three, and the one that counts in practice: cut edges
In 2016 a group at Ghent reproduced industrial processing of cut iceberg lettuce and watched where E. coli O157 and Salmonella actually went. Both attached and internalized through cut edges and wounds, and from there penetrated into the parenchyma. Internalization through stomata into the parenchyma was not seen, although some cells did turn up in the substomatal cavity.
The same paper contains a line worth reading twice: washing the cut edges with sanitizers before packing worked no better than a rinse with precooled tap water. Cut surfaces are the main postharvest risk, and storage conditions or the gas mix in the bag barely move the number.
The limits of these numbers
Erickson’s 2012 review collects the entry routes in one place: natural apertures, meaning stomata, the lateral junctions of roots and flowers, and damaged tissue, meaning wounds and cut surfaces. It then adds two caveats, without which every number above reads wrong.
First, the experiments used very high concentrations, from a million cells per gram of soil or per milliliter of water upward. Field levels like that are rare. Second, the cells that do get in are short-term residents. In Erickson’s own 2014 work, E. coli internalized into spinach could no longer be detected two days later.
So internalization is no argument against washing. It explains why washing carries no guarantee, and why hunting for a stronger product to wash with is a dead end. The investigators who wrote up the 1996 apple juice outbreak put it the same way: washing, brushing and a sanitizer at a state-of-the-art plant did not eliminate O157:H7 from the fruit, and the rule that followed the case demanded a kill step in the juice itself. That story is in our piece on E. coli in apple juice.
What it means in a home kitchen
A whole head beats a bag of cut leaves, and the reason is the cut edges. Every cut is an open wound that bacteria enter readily and that no rinse will clear afterward. A bag of cut salad is convenient, and it arrives with hundreds of those edges.
After that, cold does the work. The number of cells inside is small, but in a warm room it climbs, and the infectious dose for E. coli O157 is measured in single cells. Cut greens left warm add up to a log a day, and that is the one figure a home cook can genuinely change.
Wash the greens anyway: most cells are on the surface and water takes a share of them off. Cut with a clean knife on a clean board, because your own knife puts onto fresh cut edges what was not there before, and cross-contamination has an article of its own.
Key points
- Internalization means bacteria inside the leaf tissue. The FDA states outright that wash water does not remove them.
- Salmonella enters stomata in the light and travels there itself, following the products of photosynthesis. Forcing stomata open in the dark changes nothing.
- Crops differ widely: arugula at 88% and iceberg at 81% against parsley at 1.9%.
- Through roots, E. coli arrives from manure and irrigation water and spreads through the edible portion.
- After harvest the main gateway is cut edges. Washing those edges with a sanitizer worked no better than a cold water rinse.
- Vacuum cooling on the line drove thirteen times more cells into the tissue, and four washes did not remove them.
- The experiments ran at very high doses, and internalized cells persist only briefly. The practical answer stays simple: keep the greens cold and whole.
What else a leaf hides on its surface is covered in our piece on biofilms. How few cells it takes to make someone ill is in our article on the infectious dose. The big picture is in our food safety hub. How contamination gets into the leaf out in the field is covered in our account of the 2006 spinach outbreak.
