Each table has a default query (The one on top with the check on it). When you dragged your tables in to the dataset to create the query, it wrote a SQL statement which it uses to schema your table. Keep that query simple, you might not actually use it in code, and you can always edit that query to update the table schema. In the picture you provided, your default query is actually a stored procedure.
Every time you open the default query it connects to your datasource and allows you to select new columns that weren't in there before. If you want to update your existing columns, delete all the columns out of the table before you attempt to open the query. When you save the query, your updated columns get added back.
Make sure your connection string has permissions to view column information.