Seitan Espetinho with Vinagrete

What it is

Seitan, steamed, cubed and grilled on skewers. The dough starts with vital wheat gluten, a dry powder that's three quarters protein, worked into a paste of blended chickpeas, tomato paste, soy sauce and smoked paprika. Two foil-wrapped logs steam until firm. Once cooled, they're cut into cubes, tossed in a tempero of garlic, lime and more smoked paprika, and threaded onto skewers with onion and red pepper. Over the top goes vinagrete: tomato, onion and green pepper chopped small, with parsley, vinegar and olive oil.

Espetinho means little skewer, and vinagrete belongs at a churrasco: Panelinha even pitches its own version as the thrifty trick that makes the churrasco tastier. I'm Brazilian and lacto-vegetarian, so this is the skewer I'd hand you off my grill. Why chickpeas in the dough? Two reasons. My kitchen hunch is that they make the bite softer and less rubbery. The second reason has numbers behind it, and it's the whole story of this champion.

And this is Protein Champion number twenty, the last one. Twenty recipes, each built around a different plant protein, and wheat gluten closes the series with a huge pile of protein and one very clear weak spot. Think of it as the final exam, open book.

Why you should eat it

Here's the short version, before anyone at the churrasco asks where your protein is. USDA FoodData Central lists 75.2 g of protein per 100 g of vital wheat gluten, the dry powder seitan starts from, and each plate here gets 45 g of it. The chickpeas add theirs, and the card below counts the plate. Fiber is where gluten goes quiet: 0.6 g in that same 100 g. The chickpeas pick it up, with 6.3 g of fiber per 100 g of tinned, drained chickpeas in the same database, and by my arithmetic that one tin brings more than half the fiber in the recipe. Gluten is lean, at 1.85 g of fat per 100 g, so most of the fat here is olive oil, and the saltiest spoon is the soy sauce, 5,490 mg of sodium per 100 g. Gluten has one weak spot, an amino acid called lysine. That's for while the logs steam.

Tips

• Before anything else: the heart study in case four opens by noting that gluten triggers inflammation and intestinal damage in people with celiac disease. Seitan is concentrated wheat gluten, and the dough carries soy sauce. If you have celiac disease or a wheat or soy allergy, or a doctor has told you to avoid gluten, this dish isn't for you.
• Knead until the dough is springy, then stop. My kitchen reading: the longer you work it, the tougher the seitan gets.
• Leave slack in the foil. Seitan swells as it steams and needs the room.
• Cook it in two moves. Steam the logs on day one, chill them, and grill on day two: the Food Standards Agency gives leftovers 48 hours in the fridge, so day two fits inside the window.
• Cool the seitan before you cube it. Warm seitan tears on the skewer.
• Hungry after training? There's no rice or bread on this plate, so either one, or a ladle of feijão (Protein Champion number one), makes a good side. The card doesn't count them, and the feijão brings more lysine to the party.

Substitutions

• Tinned chickpeas: Tinned white beans (cannellini), drained and rinsed Another legume in the dough, so the teammate idea still stands. The card and the lysine sums are built on chickpeas, so neither will match.
• Griddle pan: The grill of your oven, or a charcoal churrasqueira Same skewers. Watch them closely, because times change with the heat source.
• Lime juice: Lemon juice Same job in the tempero. The card is built on lime, so it won't match exactly.
• Soy sauce: Tamari Same salty, savoury job. Don't swap it hoping for a gluten-free plate, though: the seitan itself is wheat gluten. The card is built on soy sauce, so check the tamari label for sodium.

Storage

Seitan, steamed or grilled: cool it, then get it into the fridge within one to two hours, and eat it within 48 hours or freeze it. Keep the fridge between 0 and 5 °C, and once frozen seitan has fully defrosted, use it within 24 hours (UK Food Standards Agency). The vinagrete keeps up to 3 days in the fridge (Panelinha), though it's at its brightest the day you make it.

While it cooks: the champion with one weak spot

Logs in the steamer, vinagrete in the fridge, tempero whisked. You've got 26 minutes and one job, topping up the water, so pull up a chair.

Start with the good news, because it's big. Three quarters of vital wheat gluten is protein: 75.2 g in every 100 g of the dry powder, says USDA FoodData Central. That's a lot of protein in very little room.

Now the weak spot. Protein is a chain built from amino acids. Nine of them your body can't make, so they have to come from food, and every food brings its own mix. Wheat protein runs short on one of those nine: lysine. USDA's gluten record lists no amino acids at all, which is a little rude of it, so I went to the lab papers. In a Maastricht analysis of wheat protein powders, lysine came to about 1.4 g of every 100 g of protein, and the wheat drink in case one worked out about the same. Tinned chickpeas, by USDA's record: 6.72 g. The reference that case two's authors cite asks for 4.5 g.

Teacher hat on, here's the picture I'd draw on the board. Gluten is a giant box of crayons with hardly any red in it. You can draw almost anything, and lots of it, but your strawberries come out pale. Then the kid next to you, the chickpea, slides a few reds across the table.

How many reds? My own sums say one tin of chickpeas brings about 11.5% of this dough's protein but 38% of its lysine, lifting the share from about 1.4 to about 2.0 g per 100 g of protein. A real top-up, and still well under that 4.5 line. I'd rather tell you that than round it up.

The picture stops in two places. First, a drawing gets finished in one sitting, and your body doesn't keep score one plate at a time. Case one's authors write that a shortfall like this would only hold true with a very limited variety of plant proteins eaten over a long stretch. The way I see it, beans at lunch and seitan at dinner fit inside that. Second, the box is huge. In case one, 30 g of wheat protein built muscle protein as fast as milk protein did in young men, even though their blood lysine didn't rise at all. One trial, but the weak spot didn't sink the champion there.

One honest footnote: every share here is counted in the food, before digestion, and gluten's numbers come from powders, never from a seitan log like yours.

So, in one line: gluten brings a mountain of protein with little lysine, the chickpeas top up the lysine, and the rest of your day covers what's left.

Case one: wheat protein against milk protein

Thirty-six healthy young men in Maastricht, the Netherlands, 23 on average, active but on no structured training plan. Three groups of twelve, and neither the men nor the researchers knew who drank what. After an overnight fast, each drank one shake: 30 g of milk protein, 30 g of wheat protein, or 15 g of each. Then five hours of blood samples and three small thigh muscle samples, to measure how fast new muscle protein was being built.

Muscle building rose after all three shakes, and they came out level: 0.053% an hour for milk, 0.056 for wheat and 0.059 for the blend, with no significant difference. Plot twist: the wheat shake carried 0.4 g of lysine against the milk's 2.0 g, and blood lysine didn't rise at all after it. The muscle got built anyway, at the same pace.

For your plate: each plate here carries about 34 g of protein from the gluten alone, close to the trial's 30 g. In these young men, that much wheat protein built muscle protein as fast as milk protein did.

Where it stops: the wheat was a powder already broken down for drinking (a hydrolysate), taken at rest, never seitan and never after a workout. Twelve per group is small. The authors say the results are unlikely to translate to older or less healthy people, and they describe their own earlier trial in older men, where 35 g of wheat protein didn't measurably raise muscle building while casein, a milk protein, did; nearly double the wheat, they write, could make up for it. A public-private partnership ran it, with money from Tereos, which made the wheat protein and helped choose it, plus Cargill and Kellogg; two authors have had grants or fees from FrieslandCampina, Tereos Syral and PepsiCo. Careful, and narrow.

Case two: the sequel, wheat with a legume on the team

Same lab, same registered trial and the same twelve milk drinkers as case one, so this is part two, not a second team checking the first. Twelve more young men, average age 22, drank a 30 g plant blend: 15 g of wheat protein, 7.5 g of corn protein and 7.5 g of pea protein. Teamwork on paper: wheat and corn run low on lysine, and pea protein's lysine, the authors write, is among the highest of the plant proteins. The blend carried 0.7 g of lysine per 30 g, against the milk's 2.0.

Muscle building rose after both drinks: 0.053% an hour for milk and 0.064 for the blend. The blend came out a touch higher, but the gap was too small to call (P = 0.08). In the blood, essential amino acids over the five hours came to 151 after milk and 79 after the blend, about half. The muscle didn't seem to mind.

For your plate: the wheat-plus-legume idea worked in a shake, and in this recipe the legume is the chickpea in the dough. Size check, though: pea brought a quarter of that blend's protein, and chickpeas bring about 11.5% of this dough's. Corn pulled its weight too, so the pea doesn't get all the credit.

Where it stops: powders in a shake at rest, in healthy young men only. The authors write that the results might not directly translate to whole plant foods, whose other contents can slow digestion, and may not apply to women, older adults or people who eat less protein per meal. At 30 g, both drinks may have hit the ceiling, so smaller doses might show a gap. Tereos, Cargill and Kellogg, who each supplied one of the proteins, helped fund it. A neat sequel, still in a lab.

Case three: six months with seitan on the plate

Overweight men, and women past menopause, all with raised LDL cholesterol, in their mid to late fifties on average, at a Canadian hospital research centre. After a first month with all their food supplied, 39 of them carried on for six months, cooking their own food on one of two diets, both set below their estimated energy needs. The test diet was a low-carbohydrate vegan plan the authors nicknamed Eco-Atkins: 31% of energy from plant protein, and more than half of that protein from gluten, in a high-protein nut bread and wheat gluten products, 'also called seitan'. The comparison was a higher-carbohydrate lacto-ovo vegetarian diet with whole grains and low-fat dairy.

After six months, the Eco-Atkins group had lost 6.9 kg against 5.8 kg, a 1.1 kg gap that only just cleared the statistics bar (p = 0.047). LDL cholesterol came down 0.49 mmol/L more (about 19 mg/dL), which the authors put at 9%, and triglycerides fell more too. Hunger ratings didn't differ.

For your plate: seitan was a daily protein in a diet that did a little better on LDL and weight, without leaving people hungrier. How much of that belonged to the seitan, nobody can say. The diet changed many things at once, and when the authors explain the LDL drop, they point to soy foods, nuts and viscous fibre. Gluten doesn't make that list.

Where it stops: small, and half the Eco-Atkins group dropped out (41% overall), so only 23 people finished. People ate about a third of the cholesterol-lowering foods they were prescribed. It's from 2014. Solae, a soy protein company, helped fund it alongside Loblaw and the Canada Research Chair Program, and three authors worked for Solae. If you're managing your cholesterol, take this to your doctor. A recipe can't make that call for you.

Case four: 110,017 adults and a gluten myth

US researchers followed 64,714 nurses and 45,303 male health professionals, mostly white and in their fifties, with no heart disease at the start; anyone reporting celiac disease was left out. Every four years from 1986 they filled in a food questionnaire, and the team estimated each person's gluten from wheat, rye and barley foods, then compared heart attacks across fifths of intake. Over 26 years, 2,431 women and 4,098 men developed coronary heart disease.

After adjusting for smoking, weight, activity and the rest of the diet, the top fifth of gluten eaters had about the same risk as the bottom fifth: a hazard ratio of 0.95, in a range (0.88 to 1.02) that includes no difference. The twist: gluten that travelled with whole grains went with 15% lower risk (0.85), and gluten that travelled with refined grains with none (1.00). The authors' verdict: promoting gluten-free diets to people without celiac disease should not be encouraged.

For your plate: the internet's 'gluten is bad for everyone' story finds no support here, at least for heart attacks. But mind the dose: the top fifth averaged 7.5 g of gluten a day for women and 10 g for men, and one plate of this espetinho carries about 34 g of wheat protein. My reading, not the authors': their very highest intakes, 26.7 g a day in women and 38.4 g in men, sit near one plate, out at the far tail. This study says nothing about seitan-sized amounts.

Where it stops: observational, so linked or not linked, never caused. Gluten was estimated from questionnaires, and nobody was asked about gluten-free foods. NIH grants and research awards paid for it, and no relevant ties are declared. And it left out people with celiac disease, so the first tip below is for anyone who has it.