dplyr to paste columns together in remote DB

I'd like to paste two columns together in a remote database using dplyr. I've tried something like this:

mydb %>%
    select(A, B) %>%
    mutate(newcol = paste(A, B, collapse = "_"))

But I get an error saying Error in paste(A, B, collapse = "_") : object 'A' not found . What is dplyr's translation of paste?

I tried:

> translate_sql(paste(x, y, sep = "_"))
<SQL> PASTE("x", "y", '_' AS "sep")

So I plugged that syntax in to attempt:

mydb %>%
    select(A, B) %>%
    mutate(newcol = PASTE("A", "B", "_" AS "sep"))

but using that didn't work either, with the error of Named arguments ignored for SQL PASTE. Thoughts?