The relationships could be represented this way in authorities:
100 1 $a Gregorovius, Ferdinand, $d 1821-1891. $t Lucrezia Borgia. $l English $s (Goldscheider)
500 1 $i Revision of (expression): $a Gregorovius, Ferdinand, $d 1821-1891. $t Lucrezia Borgia. $l English $s (Garner) $w r
500 1 $i Translation of: $a Gregorovius, Ferdinand, $d 1821-1891. $t Lucrezia Borgia. $l German $s (Third edition) $w r
100 1 $a Gregorovius, Ferdinand, $d 1821-1891. $t Lucrezia Borgia. $l English $s (Garner)
500 1 $i Translation of: $a Gregorovius, Ferdinand, $d 1821-1891. $t Lucrezia Borgia. $l German $s (Third edition) $w r
100 1 $a Gregorovius, Ferdinand, $d 1821-1891. $t Lucrezia Borgia. $l German $s (Third edition)
500 1 $i Revision of (expression): $a Gregorovius, Ferdinand, $d 1821-1891. $t Lucrezia Borgia. $l German $w r
100 1 $a Gregorovius, Ferdinand, $d 1821-1891. $t Lucrezia Borgia. $l German [the first expression in the original German, for which LC/PCC does not have a consensus practice]
100 1 $a Gregorovius, Ferdinand, $d 1821-1891. $t Lucrezia Borgia [the work]
The relationship that RDA expects to be represented hierarchically in the AAP string is the relationship between the work and an expression. There is no relationship designator for the relationship between a work and an expression (note that "Translation of" does not specify (expression) because it can only be used to relate one expression to another). Relationships between expressions can be represented as 5XX relationships as shown above in probably too much detail--and preferably not by longer and longer AAPs, e.g.
100 1 $a Gregorovius, Ferdinand, $d 1821-1891. $t Lucrezia Borgia. $l German $s (Third edition) $l English $s (Garner) $s (Goldscheider)
So, I'd say "Gregorovius, Ferdinand, $d 1821-1891. $t Lucrezia Borgia. $l English $s (Goldscheider)" would indeed be sufficient; but doing the authority work for that heading on a PCC bib record could be fairly involved.
Stephen