Feijão Preto with Rice and Couve

What it is

Black beans, soaked overnight and cooked soft in the pressure cooker. Then two ladles of them get mashed into a quick tempero of onion and garlic warmed with paprika and cumin, and that paste is what thickens the pot. Around them: white rice, couve (collard greens) sliced into thin ribbons and flashed in a hot pan, a crumble of fresh white cheese and a few rounds of orange.

The couve and the orange have history: in Brazil's own dietary guidelines they sit next to feijoada on the lunch plate. I've borrowed them for an everyday pot with no meat in it at all. I'm Brazilian and lacto-vegetarian, so this is the plate I'd make you if you came over for lunch.

It's also the first of twenty Protein Champions: twenty recipes, each one built around a different plant protein. Black beans go first because rice and beans sit on almost every lunch plate in the Ministry of Health's own meal examples, and it's the best lesson in plant protein I know.

Why you should eat it

Here's the short version, before anyone tells you beans are just a side dish. USDA FoodData Central lists 8.86 g of protein per 100 g of cooked black beans (call it almost 9), with 8.7 g of fiber in that same 100 g, so every spoonful is doing two jobs for you. That protein is rich in lysine, the amino acid grains run short on, while rice protein leans the other way, with a bigger share of the sulfur pair that nutrition scientists count as very slightly low in beans, so each one tops up the other. The evidence so far says they don't even have to meet on the same plate for that to count, though here they do, and beans come with some real human trials on their side, which I'll tell you about while the pot works. Cook the pot on Sunday and the hard part of lunch is sorted to Wednesday, before you've done Monday's first set.

Tips

• Mash, don't blend. Two ladles of crushed beans thicken the broth by themselves, no flour required.
• Paprika scorches in seconds. Pan off the heat the moment it smells warm.
• Couve wants a very hot pan and a short visit. Pull it while it's still bright green.
• If beans have given you trouble before, you're not being dramatic. Keep the soak: many of the gas-causing carbohydrates leave the bean for the soaking water (Nebraska Extension), and that water goes down the sink. Panelinha goes one better and changes the soaking water once along the way. Then start with a smaller portion and build up. In the cholesterol trials in case four, most of the ones that tracked gas and bloating said it eased as the weeks went on.
• Beans still firm after the squash test? Lock the lid, bring it back up to pressure and give them more time: Nebraska Extension suggests 10 more minutes under pressure. Then let the pressure drop the way your cooker's manual says before you open it. And if very old beans stay hard whatever you do, the beans are to blame and you're off the hook: Panelinha notes that old, very dry grains can stay hard after cooking, and more time won't fix them.
• Forgot to soak? Panelinha's shortcut: cover the rinsed beans with water, bring to the boil, turn off the heat, lid on, and leave them for 1 hour. Then drain and carry on as normal.
• Beans keep 3 to 4 days in the fridge, so a Sunday pot covers lunches to Wednesday. Rice is the fussy one: cook it fresh, it's 20 minutes.

Substitutions

• Dried black beans: About 3 tins (15 oz, 425 g each) of black beans, drained and rinsed. Pick tins that hold only beans, water and salt: Panelinha counts the seasoned ones as ultra-processed and says to skip them. Skip the soak and the pressure cooker. Put the beans in the pot with about 3 cups (720 ml) of water, make the tempero and carry on with the method from the mash, 10-minute simmer included. Tinned beans are softer (Panelinha again), which suits the mash. The nutrition card is built on dried beans, so it won't match; check the tin labels for added salt.
• Pressure cooker: A regular pot: soaked beans covered with plenty of fresh water, lid on, simmered gently for 60 to 90 min until tender, topped up with hot water when needed Same beans, same card. The cook stretches by an hour or so.
• Queijo minas frescal: Any fresh white cheese made with non-animal rennet, or leave it off The card uses USDA's queso fresco record, so a leaner cheese brings fat and energy down, and leaving it off takes some protein with it.
• White rice: Brown or parboiled rice, cooked to the packet's timing Brazil's dietary guidelines prefer brown rice and call parboiled a good alternative. The card is built on white rice, so it won't match.

Storage

Beans: into shallow containers in the fridge, and eat within 3 to 4 days (Nebraska Extension). They also freeze in portions with their broth, and Panelinha gives them 3 months. Rice: chill leftovers as fast as you can, ideally within an hour, and don't leave them cooling in the pan. Keep them in the fridge no more than a day, and reheat them once only, until steaming hot all the way through (UK Food Standards Agency). Couve is at its best the day you cook it.

While it cooks: the great lunchbox swap

Pressure cooker hissing, rice quiet under its lid. You've got nine minutes with nothing to do, so pull up a chair.

Protein is a long chain built from smaller pieces called amino acids. There are twenty of them, and nine your body can't make, so those have to come from food. Every few months the internet rediscovers the word 'incomplete' and aims it at plant protein. Plot twist: every plant protein carries all twenty, the essential nine included. The mix is what changes from food to food.

So look at the mix in this pot. In USDA FoodData Central's records, lysine makes up about 6.9 g of every 100 g of protein in cooked black beans, and 3.6 g in cooked white rice. The two sulfur amino acids, methionine and cysteine, go the other way: 4.4 g per 100 g of protein in the rice, 2.6 g in the beans. These are shares of the protein sitting in the food, before digestion has its say.

Read that as a trade. Rice runs short on lysine, and the beans have plenty to lend. Rice protein, for its part, carries a bigger share of the sulfur pair, the two amino acids that Mariotti and Gardner (you'll meet them in case one) call very slightly low in legumes.

I come from early-childhood education, so here's the picture I'd draw on the board: the lunchbox swap. One kid has a pile of crackers and a single apple, the other has a pile of apples and a single cracker. They trade a few across the table, crackers one way, apples the other, and now both lunches work. That's rice and beans.

The picture stops in two places. First, the bean kid brought a much bigger lunchbox: gram for gram, cooked black beans carry over three times the protein of cooked rice (8.86 g against 2.69 g per 100 g, USDA), so on this plate the beans do most of the work and the rice fills a corner. Second, kids have to swap during lunch. Your body, as far as the evidence goes, doesn't: a reasonable variety across the day looks fine, so beans at lunch and bread at dinner still count as a team. Eating them together on one plate is a practical safety net when your overall protein runs low. Nice to have, and nobody has shown yet that you need it at every meal.

And Brazil's Ministry of Health put the pairing in the national dietary guidelines back in 2014: combined with beans, cereals are an excellent source of quality protein.

So, in one line: beans bring the lysine rice lacks, rice adds a bigger share of the sulfur pair, and they can do it across your whole day.

Case one: the review that called 'incomplete' misleading

Two nutrition scientists, François Mariotti in Paris and Christopher Gardner at Stanford, pulled together the evidence on protein and amino acids in vegetarian and vegan diets, mostly from western countries: diet surveys, plus the classic studies used to work out how much protein people need.

Their verdict on the internet's favourite word: the claim that plant foods are 'missing' amino acids is false, and 'complete' versus 'incomplete' is a misleading way to talk about them. Grains run low on lysine and legumes very slightly low on the sulfur pair, which only really bites for someone living on rice alone or beans alone. Then the numbers. Using food questionnaires from 392 men aged 30 to 49 in EPIC-Oxford, a big long-running UK study, and an average body weight of 65 kg, they worked out lysine intakes of 58 mg per kg of body weight a day for vegetarians who eat dairy and eggs, and 43 for vegans, against an estimated need of 30. So, in one line: the plant eaters in that data cleared their lysine need with room to spare.

Next, protein needs. The classic way to measure them is nitrogen balance, a test of whether your body takes in at least as much protein as it loses. When a big pool of those studies was split by protein source, animal, plant and mixed diets pointed to about the same protein need (the plant diets there mixed grains with legumes or used soy, never rice or beans alone). Experimental diets built from plant-protein mixes showed no effect on nitrogen balance either. And on timing, they found no evidence yet that plant proteins need a partner in the same meal: a reasonable variety across the day looks fine.

For your plate: rice and beans together is a lovely habit, and you're allowed to relax about it.

Where it stops: the paper describes no systematic search, so the two authors chose and weighed the evidence themselves, and the diet data are self-reported and western. They're fair about the other side, too: when single isolated proteins like wheat or soy were tested, the body's protein building after a meal, in muscle or across the whole body, came in lower than with animal protein. One disclosure: Mariotti leads a research contract with Terres Univia, the French body for plant oils and proteins, and receives no fee for it. Solid, and upfront about its own gaps.

Case two: rice alone against rice and beans

Healthy women in Phoenix, Arizona, between 18 and 65: thirteen signed up, and twelve ate the black bean meal. On separate mornings, after an overnight fast, each one ate a test meal carrying the same 50 g of usable carbohydrate: 3/4 cup of white rice on its own, or 1/2 cup of canned black beans with 1/2 cup of the same rice. Then the researchers checked their blood every 30 minutes for two hours.

From the one-hour check through to the two-hour check, blood sugar sat lower after beans and rice than after rice alone. Across the full two hours, the black bean difference came out as a trend (p = 0.058), just short of the usual cut-off for calling it a clear result. Insulin ran a little higher at the 30-minute mark after the beans, and total insulin over the two hours was no different.

For your plate: same carbs, part of them moved from rice to beans, and in these women the rise came in gentler.

Where it stops: it was one meal per test in twelve women, none of them with diabetes, so the authors say it can't be applied to men or to type 2 diabetes. They declare no conflicts of interest. Promising, and small. If you manage your blood sugar, talk to your doctor before you change your plate around it.

Case three: black beans against beef at breakfast

Thirty-five adults aged 60 and up in Guelph, Canada, average age 72. On three mornings, in random order, each ate the same breakfast tortilla bake built on either 135 g of canned beans (black one day, red kidney another) or 80 g of lean beef. The bean versions carried 21.4 g of protein and 11.9 g of fiber; the beef version, 34.9 g of protein and 1.7 g of fiber. For three hours they rated how hungry and full they felt, then sat down to as much pizza as they wanted.

Fullness and hunger followed the same path after all three breakfasts, with no measurable difference between them. At the pizza, they ate 167.5 g after black beans and 159.5 g after beef, a gap too small for the statistics to call (p = 0.07). Across the whole day: 1,877 kcal after black beans, 1,890 after beef. The beans did lose one round. Taste scored 63.7 for black beans against 74.3 for beef on a 0 to 100 liking scale. I'd love a rematch with a proper tempero.

For your plate: with less protein and far more fiber, black beans kept these people about as full as beef did, as far as this test could see.

Where it stops: it was one breakfast in older adults with little ethnic diversity, so it may not map neatly onto a 20-year-old after leg day. Nobody was blinded (you know when you're eating beef), and the rest of the day's food was self-recorded. The Ontario Bean Growers paid for the work. A fair hint, not a verdict.

Case four: pulses and your cholesterol

Canadian researchers pooled 26 randomized trials, 1,037 people in all: mostly middle-aged, men and women in about equal numbers, some with normal blood fats and some with high, at moderate risk of heart disease on average. In each trial one group ate a diet built around pulses (the bean, lentil, chickpea and dried pea family, with beans the most common) and the other ate a diet with the same calories and no pulses. The typical dose was 130 g a day, about one serving, for about six weeks.

LDL cholesterol, one of the main blood-fat numbers doctors target to cut heart risk, fell by 0.17 mmol/L on the pulse diets (about 6.6 mg/dL), roughly 5% from where people started. Two related targets, apoB and non-HDL cholesterol, showed no clear change, and only one trial measured each of them. Tummy trouble came up too: flatulence in 7 of the 11 trials that asked, bloating in 6, and most of them reported that it eased as the weeks went on.

In 2026 an updated version of this review came out, 38 trials and 2,095 people, with a similar answer: LDL down 0.14 mmol/L on average, and its authors say pulses likely lower LDL a little. For beans, the drop grew with the dose up to one cup a day, about 0.25 mmol/L more for each extra half cup, but the uncertainty around that line is wide (anywhere from 0.02 to 0.48), so treat it as loose. Full honesty: that update sits behind a wall I couldn't get through, so I've read only its published summary, and I can't tell you who took part or who paid for it.

For your plate: in these trials, pulses most days nudged LDL down a little. That's the real size of it.

Where it stops: the authors of the 2014 review say most of their trials were low quality and shorter than three months. Their review was funded by a Canadian Institutes of Health Research grant and by Loblaw Companies Limited, and several authors had received support from pulse organisations, Pulse Canada and the Saskatchewan Pulse Growers among them. One plate of this recipe holds about 1.3 cups of cooked beans, past the one cup where the 2026 dose line ends, so don't read 'more beans, more drop' beyond it. If you're managing cholesterol, your doctor is the right person to decide what this means for you.