'TypeError: expected string or buffer' when performing re.sub on beautiful soup result set iteration - regex

Question

content_a is a beautiful soup result set (ie the type is <class 'bs4.element.ResultSet'>) that is made up of values whose type is <class 'bs4.element.Tag'>.
If i print 'content_a' i get:
[<div class="class1 class2">Here is the first sentence.
<br/> <br/> Here is some text "and some more text."
<br/> <br/> Here is another sentence.
<br/> Text<br/><span class="class3">Text</span></div>, <div class="class1 class2">Here is the first sentence.
<br/> <br/> Here is some text "and some more text."
<br/> <br/> Here is another sentence.
<br/> Text<br/><span class="class3">Text</span></div>, etc
So it seems to me it should be a simple iterable list of divs.
I am wanting to replace <div class="class1 class2"> with <div class="class1 class2"><p> (my eventual goal being to replace all <br />'s with paragraph tags).
In my test where the source content is a string I have:
import re
blablabla = ['<div class="class1 class2">', '<div class="class1 class2">']
for _ in blablabla:
_ = re.sub('(<div class=\"class1 class2\">)', r"\1<p>",_)
print _
which returns, as required:
<div class="class1 class2"><p>
<div class="class1 class2"><p>
I am trying to perform the same process on each iterable in content_a with:
import re
for _ in content_a:
_ = re.sub('(<div class=\"class1 class2\">)', r"\1<p>",_)
print _
but am getting the error:
...in sub
return _compile(pattern, flags).sub(repl, string, count)
TypeError: expected string or buffer
So the only difference that i can tell between the two examples is that one is a beautiful soup result set and one is just a plain list.
Can anyone see why this error could be occuring?
Edit:
Someone has pointed out here that sub requires a string as the third argument, so the third argument that i am passing is the iterable value which is of type <class 'bs4.element.Tag'>. So perhaps this is the problem. But i need to retain the nature of these values for later modification so i am not sure how to proceed at the moment.
Update/Workaround:
Just to save someone spending time on an answer, i figured out a workaround, basically i realised i could adjust the content later in the process and i did this by converting it to a string with read() and could then perform all the re.sub changes on the required elements in the string.
And the little regex i came up with was:
string = re.sub('([^\r]*)\r', r'\1</p>\n<p>', string)

Answer 1

As suggested, I am posting the workaround I used as the solution:
Update/Workaround:
Just to save someone spending time on an answer, I figured out a workaround, basically I realised I could adjust the content later in the process and i did this by converting it to a string with read() and could then perform all the re.sub changes on the required elements in the string.
And the little regex I came up with was:
string = re.sub('([^\r]*)\r', r'\1</p>\n<p>', string)