The arsenic for phosphorus swap is accidental, rather than a facultative one, and the question whether arsenic is nonessential or toxic is quantitative, not a qualitative one.

Sci Total Environ. 409 (2011) 4889–4890.
Dani SU.
Department of Medicine I and Clinical Chemistry, University of Heidelberg, Germany.
Abstract - Arsenic shares many physicochemical properties with phosphorus, so that arsenic can be taken up inadvertently by cells through the pathways for phosphorus. As a phosphate analog, arsenate competes with phosphate and enters cells via phosphate transporters. In the cell, arsenate can be recognized as a substrate by enzymes that usually use phosphate as a substrate. The phosphate for arsenate swap results in wasteful 'futile cycles' in metabolic pathways, uncoupled oxidative phosphorylation and extreme DNA instability. The disrupting metabolic effects of arsenic have an evolutionary meaning, so that all living organisms-from chemoautotrophic organisms that grow by reducing or oxidizing arsenic to metazoan-carry highly conserved arsenic resistance genes. Arsenic resistance can result from different strategies including selective transport to maximize phosphate uptake and minimize entry of arsenate, active transport to export arsenate, arsenic storage in specialized compartments, enzyme selectivity toward phosphate, and increased efficiency of DNA repair systems. None of these strategies is infallible, though, and susceptibility to arsenic toxicity varies between taxa in many orders of magnitude. Even arsenic-hypertolerant organisms will stop to grow and will eventually die when exposed to arsenic over species-specific resistance limits. The arsenic for phosphorus swap is an accidental one, it does not warrant a conclusion in favor of the essentiality of arsenic to life as we know it.